AMD’s eight-core, 16-thread chips lead the Ryzen charge

AMD's first Ryzen CPUs are finally here. This morning, the company is putting its highest-end chips up for pre-order. The Ryzen 7 1800X, the Ryzen 7 1700X, and the Ryzen 7 1700 will mark AMD's return to the high-performance desktop CPU market, and the company's final internal numbers ahead of launch suggest they'll mark a return to competitiveness, as well.

We're flying home with Ryzen review samples as of this writing. Independent review results are still under NDA, but these are the three CPUs that will mark AMD's return to competitiveness in the marketplace:

Model Cores Threads Base clock Boost clock XFR TDP Price
Ryzen 7 1800X 8 16 3.6 GHz 4.0 GHz Yes 95W $499
Ryzen 7 1700X 3.4 GHz 3.8 GHz Yes 95W $399
Ryzen 7 1700 3.0 GHz 3.7 GHz No 65W $329

One of the most oft-repeated numbers from the lead-up to Ryzen has been AMD's goal of a "40% IPC increase" relative to its older processor generations. CEO Lisa Su claimes the company has not only met, but exceeded that goal with a 52% improvement.

With that greater-than-expected IPC gain, Su says the Ryzen 7 1700X is 4% faster than Intel's $1050 Core i7-6900K, a Broadwell-E chip with eight cores and 16 threads of its own. Su also touted the 1700X's 39% Cinebench all-thread advantage over the similarly priced six-core, 12-thread Core i7-6800K.

The Ryzen 7 1800X ups the ante over the i7-6900K with a 9% Cinebench all-threads performance advantage. Su says this chip is now the fastest eight-core, 16-thread CPU on the market.

Cinebench is just one benchmark, of course, and an especially favorable one for demonstrating a processor's multithreaded performance. I'm eager to produce a more complete picture of Ryzen performance from my own test benches with a full range of single-core tests.

Folks who are already impressed with Ryzen can make like a buffalo and stampede over to their favorite online retailer to place a pre-order today. The more cautious can hold off for benchmark info in tandem with these chips' hard launch on March 2.

Comments closed
    • NeelyCam
    • 3 years ago

    Here’s something – Intel’s CPU prices are coming down now:

    [url<]http://www.digitaltrends.com/computing/intel-cpu-prices-drop-ryzen-launch/[/url<]

      • adamlongwalker
      • 3 years ago

      Yes. One of my concerns I posted recently on what Intel is going to do. And from reports it will look like when their next CPU comes out there will be an even another cut in price.

      • chuckula
      • 3 years ago

      That story appears to be bad journalism by somebody who just learned that Microcenter is a thing and flipped out.

      Otherwise, Intel has been making crazy price cuts to their parts because RyZen for the last 6 years or so.

      Case in point: Apparently somebody forgot to tell Newegg about all this discounting.

      [url<]https://www.newegg.com/Product/Product.aspx?Item=N82E16819117645[/url<] [url<]https://www.newegg.com/Product/Product.aspx?Item=N82E16819117726[/url<]

        • ronch
        • 3 years ago

        True, price cuts have been happening all the time but when Intel cuts prices days before Ryzen, a product that by early indications looks to be a disruptive challenger, comes out, it’s a bit naive to say the cuts don’t have anything to do with Ryzen. Intel hasn’t felt this threatened by AMD in a long time and it’s fun seeing them scramble like this after selling slowly evolving tech to the world at high prices.

    • rizkiyoist
    • 3 years ago

    On the first graph they are comparing i7 6800 which scored 1108 vs 1700X at 1537.
    (1108/1537) * 100% = 72%
    So you could say 6800 is 100% – 72% = [b<]28% slower[/b<] (from AMD's perspective) Although (1537/1108) * 100% = 138% So you could also say 1700X is 138% - 100% = [b<]38% faster[/b<] (from Intel's perspective) And in the second graph when they compared 6900 with 1800X... (1474/1601) * 100% = 92% You could say 6900 is 8% slower (from AMD's perspective) While 1800X is (1601/1474) * 100% = 108.6% You could say 1700X is 8.6% faster (from Intel's perspective) It's evident that the percentage varies depending on which perspective you see it from, just like 100 plus 10% is 110, but 110 minus 10% is not 100 but 99. The first graph should've been either -28% written over Intel's score or +38% over AMD, otherwise you would be mixing the calculation resulting in 937.5 score for the 6800.

    • Thbbft
    • 3 years ago

    [b<]"Ryzen 7 1800X Beats The World’s Fastest Desktop CPU (i7-6950) With A One Click Overclock On Air Cooling, Surpasses 7700K In Single-Thread Performance" - donanimhaber[/b<] Verbally transmitted in a video that didn't break the NDA because - they just HAD to comment on that they were seeing. Donanimhaber is considered one of the world's premiere and most reliable review sites. This is beyond the most wildly optimistic projections of Ryzen's performance I've seen. If this holds the 1700X and 1800X will simultaneously be the highest performance desktop GAMING [b<]AND[/b<] WORKLOAD CPUs in the world. Now extend that to servers and the professional market - Intel is in SO much trouble. [url<]https://www.youtube.com/watch?v=UhKmeCdB914[/url<]

      • Pancake
      • 3 years ago

      “Donanimhaber is considered one of the world’s premiere and most reliable review sites. ”

      o_O

        • LostCat
        • 3 years ago

        good thing the world has ever heard of it….owait

      • tipoo
      • 3 years ago

      Nice try, donanimhaber

    • derFunkenstein
    • 3 years ago

    Looks like I won’t be eating any hats, since cheaper Ryzen CPUs [url=https://www.techpowerup.com/230928/amd-ryzen-5-six-core-processors-to-launch-in-q2-2017<]won't be available until Q2[/url<]

      • Airmantharp
      • 3 years ago

      That’s disappointing- these are the parts that will really put the screws to Intel’s pricing and SKU schema, even more so if they turn out to be good overclockers.

        • freebird
        • 3 years ago

        Yeah, but they don’t put much money into AMD’s revenue/bottom line. Especially if the dies are good, but limited and can be sold as 8-core/six-core; also some 8-cores will be low end server parts. Which plays into availability… if this core is this good AMD might devote more of their manufacturing to server parts if the demand is there. Time will tell.

          • Airmantharp
          • 3 years ago

          Not a bad way to look at it, I agree.

          And really the only true complaint at all is that because AMD hasn’t measurably eclipsed Intel in terms of IPC (they’re nipping at their heals), and they haven’t released parts in the >4.2GHz range, they’re still behind in terms of absolute gaming performance and gaming performance for the dollar. Higher clocked four- and six-core parts would certainly help here!

            • derFunkenstein
            • 3 years ago

            The way the leaks have come out, it looks like for some mystifying reason the hexa-core and quad-core parts won’t have any higher clocks than the octa-core version. That really concerns me that there isn’t much OC headroom in this design. At least, not yet.

    • Ninjitsu
    • 3 years ago

    Jeff and crew, I know you’re busy with stuff, but a quick request. Many in the general Arma community are looking at Zen with interest, so it would be great if you could run Arma 3 benchmarks too, if possible.

    Would also be great to see a Arma shootout between overclocked:

    1) various 4C/4T and 4C/8T Ryzens
    2) i3 7350K
    3) i5/i7 Kaby lake

    I’d be more than happy to send you a PM detailing how exactly which settings affect what component (disk/memory/GPU/CPU), with a list of the three benchmarks that I know of and what their characteristics are.

    Thanks, cheers and good luck with the review!

    • Dazrin
    • 3 years ago

    So, to put this in a performance/value graph:
    [b<]Graph removed, go see TR's review.[/b<] Notes: R5-1600X performance assumes it scales directly with # of cores (so 6/8 of the R7-1800X since they are the same clock speeds), to put it mildly, this is "optimistic". i5-7600K performance assumes R5-1600X/1.69 (per the link chuckula posted, thanks!)

      • Chrispy_
      • 3 years ago

      7 edits?
      Have a cookie for making a graph though.

        • Dazrin
        • 3 years ago

        Couldn’t get the link to work using google docs. Haven’t done this enough.

        Now 8 edits since we now have TR graphs so mine can go away.

    • chuckula
    • 3 years ago

    For those of you interested in headwear haute-cuisine, it looks like those 3 RyZen models are it for the launch next week.

    6 core models launch at some point next quarter and the 4 core models are in the second half of the year.

    You can see the slides from the press event here: [url<]https://www.techpowerup.com/230928/amd-ryzen-5-six-core-processors-to-launch-in-q2-2017[/url<]

    • sparkman
    • 3 years ago

    Wait, Ryzen is an AMD CPU that isn’t slow and hot?

      • K-L-Waster
      • 3 years ago

      Maybe you’d better sit down. Do you need a glass of water?

    • Mat3
    • 3 years ago

    AMD now has everything they need for a killer APU. A 4 core Ryzen CPU with a 480X or something newer (on the same die or joined by interposer) would rock.

      • RAGEPRO
      • 3 years ago

      Polaris 10 is too big, memory bus is too wide. Most of the GPU core would sit idle waiting on RAM, methinks. Polaris 11 (or the newer Polaris 12 chip?) would be a better fit and still plenty fast.

        • ptsant
        • 3 years ago

        512 GCN cores is already a very, very decent iGPU and won’t cannibalize the RX 460.

      • tipoo
      • 3 years ago

      It’s surprising that they still havn’t made an APU with a GPU near the PS4s, let alone a 480. Maybe they were waiting on this die shrink, as PC CPU cores would be larger than the low end Jaguar ones in the PS4.

      • Lord.Blue
      • 3 years ago

      Hmm…having HBM2 dedicated on die as a largish cache for the iGPU and the CPU to use…that might work.

    • SiSiX
    • 3 years ago

    Was reading comments on another site and a thought occurred to me: AMD is positioning these three chips (and there are only three atm) against Intel’s HPDT chips (the 6 and 8 core 68/6900 series i7’s, and to a lesser extent the 10 core monster 6950x), but NOT against Intel’s Xeon chips.

    Curious don’t you think.

    I just did a run down of Dell, HP, and Lenovo and not a single one of these OEM’s offers a computer with an i7 68/6900 series chip, just Xeons. I can get i7 7700 chips and I can get Xeon chips (up to and including dual chip/22 core monster setups), but no Extreme Edition processors to be found by the big three. Basically, outside of the boutique builders, the Extreme Edition processors from Intel, the chip that AMD is marketing against, is strictly an enthusiast chip.
    (I did find a single Alienware machine with a i7 6950x option, but Alienware is effectively a boutique builder, even if they are part of Dell.)

    So, the point here is it’s a little like (and the analogy is a little shaky, but bear with me) Lamborghini making a comeback against Ford by saying their new car is faster than the GTO-350 at half the price, which is great, except only a select few actually buy those.

    So does AMD actually have the resources to build the quantities of chips to support this launch at these prices AND actually get them to OEMs? Or are they effectively building what amounts to a “enthuthiast/boutique” chip and hope that Intel doesn’t do anything to upset the apple cart in the meantime.

    Seriously, I hope AMD has a winner, but in business, only a fool GIVES away a product on price if it’s actually competitive or better than what the competition has, especially when they’re bleeding money, even to grab market share. Remember, the real money is in convincing OEMs to buy this chip, not any of us. (And AMD has been a staple of low end market for years at retail. The question becomes one of breaking into that higher end market with more expensive chips.)

      • ptsant
      • 3 years ago

      AMD will launch the server edition later.

      For what it’s worth, Ryzen chips do support ECC, at least based on the MB documentation. My guess is that almost all the special features of Xeons are already “standard” on Zen and in fact AMD doesn’t need to launch a different model to cover that market. It’s probably a question of giving them first to the OEMs, possibly with a different package/warranty and ensuring appropriate software/MB support.

      Things are not that simple with Intel. You buy an enthusiast chip, you don’t get ECC support. You buy a Xeon, you don’t get overclocking. I’m sure there are other differences (VT-d? TSX-NI?) but I really can’t manage to follow the 50 different Xeon flavors.


      EDIT: Apparently, the MB specifications I could find (ASUS X370 and Gigabyte), specifically state that the ECC functionality is not available, but ECC unbuffered will work as non-ECC (at least for ASUS). Can’t say if this is due to the CPU, MB or simply BIOS.

        • Krogoth
        • 3 years ago

        I’m more curious to see how IOMMU (VT-d) and AMD-V support (VT-x) shapes up on Zen line-up.

          • ptsant
          • 3 years ago

          I’m not an expert, but last time I used a VM, both seemed to work nicely on my FX 8350, so it shouldn’t be very different with Zen I suppose. These are old features.

    • chuckula
    • 3 years ago

    Rza79 and a few others pointed out one thing from AMD’s presentation that I missed the first time: Apparently RyZen has [b< ]4.8 billion[/b<] transistors. (See [url=http://www.anandtech.com/show/11143/amd-launch-ryzen-52-more-ipc-eight-cores-for-under-330-preorder-today-on-sale-march-2nd<]Anand's story for that number[/url<]) Now remember, this chip devotes literally zero transistors to graphics, so that number doesn't include a single GPU transistor. To put that in context a [b<]10 core[/b<] 6950X part with a quad-channel memory controller and more PCIe lanes than RyZen is only about 3.2 Billion transistors (http://www.tomshardware.com/reviews/intel-xeon-e5-2600-v4-broadwell-ep,4514-2.html)

    While nobody expected RyZen to be a small chip, the fact that AMD went out of its way to emphasize the comparatively simple and compact design of its cores raises this question: With two fewer cores, a cache that while large is still smaller than a 10-core Broadwell’s cache, only a 2-channel memory controller, and fewer PCIe lanes, what exactly are those extra 1.6 Billion transistors* doing?

    * For reference, a 2600K Sandy Bridge part (the *ENTIRE CHIP*) is a little [edit: over] 1 Billion transistors.

      • anotherengineer
      • 3 years ago

      “what exactly are those extra 1.6 Billion transistors* doing?”

      When you find out let us know.

      *brace yourselves……………the Ryzen jokes are coming*

      however I thought it might have something to do with integrating chipset on the die or did they go that route??

        • chuckula
        • 3 years ago

        There is some south bridge stuff directly on-die, but it’s comparatively simple and appears to be just enough to let RyZen run without a chipset in some embedded systems.

        If the entire southbridge were on-die then there literally wouldn’t be an X370/B350/etc. motherboard since those chips would not exist.

        The southbridge transistors add some complexity, but I guarantee you that a comparatively simple southbridge implementation does not have a transistor budget that’s more than 60% larger than the entire 2600K chip.

        • jihadjoe
        • 3 years ago

        Somewhere in a quantum state of being there and not. Just like the 800 billion missing transistors in Bulldozer.

      • tipoo
      • 3 years ago

      6950X die shot
      [url<]http://cdn.wccftech.com/wp-content/uploads/2016/05/Intel-Broadwell-E-Die-Chipshot.png[/url<] Ryzen die shot [url<]http://images.anandtech.com/doci/11143/Ryzen%20Die%20Shot.jpg[/url<] You're right, there seems to be a lot of uncore, and this, uh, doesn't answer your question at all They do have core complexes that grant each of 4 cores equally timed access to any part of the shared L3, so that would have some wiring overhead, but that seems like maybe a few percent of the die. I guess also keep in mind that no one is counting 5 billion transistors, everyone measures and reports transistor counts a bit differently.

        • synthtel2
        • 3 years ago

        I’m particularly curious about the stuff above the left CCX in that shot. Most of the uncore either looks familiar from other AMD hardware (memory controllers, I/O) or close enough to make a good guess, but I’ve never seen that structure before.

        I annotated a die shot, FWIW: [url<]https://i.imgur.com/ioF3P3u.jpg[/url<]

          • tipoo
          • 3 years ago

          2x72bit (9 big blocks on each side). Haswell-E has 2 of these WTF things, and they’re the actual memory controllers, while what you highlighted as memory controllers are likely GMI.

            • synthtel2
            • 3 years ago

            [url=https://en.wikichip.org/wiki/intel/microarchitectures/haswell#Die<]Where?[/url<] Haswell-E's memory controllers look more like what I highlighted as such, AFAICT. For that matter, all the obvious I/O drivers (top and bottom) look a lot beefier than the WTF thing. The reason I'm pretty sure the memory controllers are the big things at the corners is that those look so much like the memory controllers on GCN chips (numerous examples [url=https://www.flickr.com/photos/130561288@N04/<]here[/url<]). In the WTF thing, there are the 18 large blocks with what looks like 10 or 12 elements each and 24 small blocks with 4 elements each, plus two blocks with a lot more macro complexity and what looks like 4 more elements like the others each (something clock related?), plus all that space in the middle (which could contain a lot of logic). There's also that differently-colored strip along both the top and bottom, which only otherwise shows up along the top and bottom of the CCXes. In the case of something like a memory controller, I'd think of that as heavy wiring to get to the edge of the chip with little loss (does that even work?), but it doesn't cover the area it would need to to do that and it's got one at the top edge too. Between all that, it doesn't end up looking like any of the usual suspects. I did just have one thought: what if it's for communication between dies when there are four of these sharing a package? That would explain having a boatload of low-power elements, since tons of BW would be needed but the signal paths would be very short.

            • tipoo
            • 3 years ago

            Supposedly they are using RAMBUS PHY, so we shouldn’t necessarily expect them to look the same.
            The ones labelled as memory controllers also make sense as GMI links as their split arrangement makes more sense as you may want to connect another die or what-not in any location rather freely.

            But then the PS4/XBO APUs do have such a split memory controller…
            [url<]http://www.nag.co.za/wp-content/uploads/2016/05/AMD-Xbox-One-PS4-APU-die-shot.jpg[/url<] Anywho, hopefully once the NDA lifts someone can tell us!

          • the
          • 3 years ago

          Remember there are several external coherent busses (still HyperTransport?) that are on-die but not utilized by consumer parts. As pointed out else where in this thread, there are some SATA, USB and of course PCIe controllers on-die.

          For comparison, the 10 core Broadwell-E also has two QPI links that are unused on consumer parts that go toward its transistor count too.

          • the
          • 3 years ago

          Looking at your annotated die shot, you may have the memory controller and PCIe blocks reversed. I’m seeing 18 subsections in your WTF block which would be indicative of eighteen 8 bit lanes (remember the extra chip for ECC). The other give away is that both memory channels are located on the same side of the processor. (For comparison, the two 128 bit memory controllers on Intel’s socket 2011 chips are on opposite sides of the die.)

          There are also 18 of the blocks of what you labeled ‘memory controller’ but they are spread out. There are an additional two near the bottom away from the rest. I would fathom that those two are for the chipset link. The other two block look like it could divided into a 4x/4x/4x/2x/2x host lane division.

          I would have guessed that those two big SRAM arrays could be for the coherenecy links (directory cache) but it would make sense to put those next to the coherent link blocks. I couldn’t find another similar block next to either one that’d be responsible for the link.

          There is supposed to be some ARM security processors on the die too. I think they’re on the bottom of the die.

            • synthtel2
            • 3 years ago

            Yeah, looking at that on another day, I probably did get that the wrong way around. The PCI-E controllers are exactly the same as [url=https://www.flickr.com/photos/130561288@N04/28777482876/in/photostream/lightbox/<]P10's[/url<], they just didn't look like it because I had scaling wrong. The XB1's layout also shows that memory controllers don't need to be right at the edge of the chip, and I was definitely looking at the wrong other chips to see how much space memory controllers ought to take. :embarrassed: Looking around a bit more, [url=https://www.flickr.com/photos/130561288@N04/29111683364/in/photostream/lightbox/<]GP104's[/url<] memory controllers look suspiciously similar, right down to the weird asymmetries and physical size (Nvidia's are a bit beefier, presumably to handle VRAM's ludicrous speeds). I also found [url=https://www.rambus.com/rambus-and-nvidia-sign-patent-license-agreement-2/<]this[/url<], and Zen may also be using Rambus stuff. Trouble is, I count the same number of probable drive elements in each of them (320), but GP104's should be twice as wide. Intel's controllers are a bit different, but it still looks like Zen's takes space more equivalent to HSW/BDW-E's than SKL-S's. Is there something we're not being told here? This also means we have 32/34 PCIe lanes in hardware, rather than the 16-22 (?) that are usable here. This is a lot of uncore beef to be disabled, and I'm now very curious what happened to all of it.

      • Rza79
      • 3 years ago

      My best guess would be that their automatized tools are ‘wasting’ transistors.
      The same story kinda happened when Bulldozer was launched. It had more transistors than Sandy Bridge while SB had a GPU on board. It was a full 99mm² larger.
      I think the same applies here. Intel is still massively finetuning their chips by hand while AMD switched over automatized tools, many of which it inherited from ATI.

      Also remember how Bulldozer supposedly had 2 billion transistors at first and then AMD lowered it by 800 million. Who knows how AMD is counting.

      One thing we should not forget is that Ryzen has basic south bridge functionality built in, being a SOC. Though I doubt it would add a billion transistors. 2 Sata & 4 USB 3.0 if i’m not wrong.

      PS: SB has 1.16b transistors

        • chuckula
        • 3 years ago

        Oh yeah, SB is about 1.16 Billion transistors.

        You might be right that these are auto-formatted transistors that get thrown onto the chip but may or may not even be active to do anything.

      • GrimDanfango
      • 3 years ago

      [quote<]what exactly are those extra 1.6 Billion transistors doing?[/quote<] Transisting?

      • ptsant
      • 3 years ago

      Seems like a lot. The surface of the cores is small, at least according to AMD.

      • BaronMatrix
      • 3 years ago

      SenseMI (Infinity Fabric) and an ARM Secure processor…

      • NTMBK
      • 3 years ago

      There’s more than one way to count transistors. Remember when 800 million transistors went missing from Bulldozer? [url<]http://www.anandtech.com/show/5176/amd-revises-bulldozer-transistor-count-12b-not-2b[/url<]

    • kmieciu
    • 3 years ago

    Yeah, cool but … will it [s<]blend[/s<] support ECC memory?

      • ptsant
      • 3 years ago

      I hate to contradict myself, but despite the early rumors, the ASUS and Gigabyte sites say NO, at least right now. Can’t say whether this is a BIOS limitation, a MB limitation or inherent to the chip.

      Gigabyte specifically states that ECC will work but without error-correction functionality.

        • Krogoth
        • 3 years ago

        I suspect the higher-end Ryzens will probably have it. It is a different matter for motherboards though.

    • HERETIC
    • 3 years ago

    RYZEN CHIPSETS-
    Seems like the case trend(you don’t need more than 2 drive bays) is passing on
    to Ryzen chipset.
    Looks like the mainstream B350 ONLY has TWO SATA ports.
    So a optical and one spinning rust is all you get folks.

    As much as I agree with getting rid of legacy stuff-this is too limiting……
    So many decent SATA SSD’s and limited pcie.

    [url<]http://www.legitreviews.com/gigabyte-amd-x370-b350-board-prices-8-core-cpu-board-424_191633[/url<] EDIT My mistake-only saw the chart at legit-2 off cpu. big relief.............................................

    • marvelous
    • 3 years ago

    Her manly face and broad shoulders are turning me on. Come here Ryzen

    • Klimax
    • 3 years ago

    Either they are idiots and are going for price war with Intel. (End result is similar to land war in Asia) Or this will be hype train high-velocity megacrash.

    Too similar to previous launches like Bulldozer, Fiji and Polaris.

    • oldog
    • 3 years ago

    What’s with the black tee and blue jeans?

    Is Lisa Su trying to channel Steve Jobs?

      • Thbbft
      • 3 years ago

      Guarantee there are oodles of Tech Fems dying to get a piece of that action.

        • oldog
        • 3 years ago

        Say what you will about Marissa Mayer as a tech CEO, but at least she knows how to dress.

      • ronch
      • 3 years ago

      No, just the RDF.

      • JustAnEngineer
      • 3 years ago

      A quick search didn’t turn up many photos of her before she became COO.
      [url<]http://www2.technologyreview.com/tr35/Profile.aspx?TRID=397[/url<]

      • derFunkenstein
      • 3 years ago

      Lots of tech CEOs have tried to channel Steve by wearing his black-and-blues.

    • Thbbft
    • 3 years ago

    [b<]WILL INTEL BUY NVIDIA AND HAND THE REINS TO JHH?[/b<] Nvidia is prospering even without an x86 license. Love him or hate him JHH is proven to be an extremely intelligent and effective CEO. Once investors truly start to grasp the full significance of that 52% IPC increase in desktop, mobile, APU, professional and server market share there can be little doubt the stock will tank and Krzanich will be fired. Seemingly on a turn of a dime Intel is suddenly in serious trouble across the entirety of it's x86 markets. 52% IPC gain is PARADIGM SHATTERING and there is only more pain to come over the next four years. MUCH more pain. But there is an alternate route. At least twice in the past Intel has tried to buy Nvidia, but wouldn't agree to let JHH take the helm. That time may have come. They DESPERATELY NEED high performance GPU architectures to couple with their CPUs and now they as obviously need a truly high performance proven CEO, a Lisa Su class CEO, and JHH fits that criteria AND brings along that desperately needed high performance GPU capability. Imagine what JHH could do with Intel's resources and IP at his command. Granted it is a scary thought. JHH's time may have come.

      • NTMBK
      • 3 years ago

      Erm, no.

      • K-L-Waster
      • 3 years ago

      Wow, one product launch and apparently Intel is doomed — every single OEM will drop all of their chips in laptops and servers and cut over en masse to RyZen…

      Or not.

      Look, one the things you need to grasp is that the 52% increase is over Excavator, not Intel’s cores. They aren’t leaping off into the lead, they are basically just undoing how far behind they were with Excavator. According to AMD RyZen is ahead of Intel, but not by leaps and bounds (looks like 5 – 10%) — and there are some questions there about the methodology, plus it’s clear that they chose the most advantageous benchmarks to show off for the launch.

        • Magic Hate Ball
        • 3 years ago

        If Intel survived the original Athlon they’ll survive this no problem.

        And I’m an AMD fan.

          • BurntMyBacon
          • 3 years ago

          [quote=”Magic Hate Ball”<]If Intel survived the original Athlon they'll survive this no problem.[/quote<] For similar reasons as well. The original Athlon was more impressive relative to Intel's counterparts than Ryzen. IIRC AMD was also in a better position than the AMD of today. They simply didn't have the capacity to keep up with demand. Even if Ryzen doubled the performance of Intel's counterparts, they couldn't kick Intel out of the market without the production capacity to supply Intel's current customers. Then you have to consider that Intel won't be sitting still either. Intel already has deeper customer relations, new chips in the pipe, and plenty of margin with which they can cut prices. If AMD wants to win, then they need to keep the pressure on (and not just with pricing).

      • chuckula
      • 3 years ago

      NO! AMD WILL BUY INTEL AND HAND THE REINS TO HECTOR!

        • hansmuff
        • 3 years ago

        HECTOR SALAMANCA, THAT IS

          • Redocbew
          • 3 years ago

          What’s that Hector? You want to design a chip that’ll get us out of the hole we’re in after Bulldozer(hah, funny joke), name it Ryzen(really?), put it up for pre-order(that’s weird, but ok), then hype it until people go bonkers?

          DING!

        • K-L-Waster
        • 3 years ago

        Now that would definitely get AMD charged on anti-trust…

    • torquer
    • 3 years ago

    Willing to trade countless breathless proclamations of Intel’s impending bankruptcy for a timely unbiased TR review.

    • BaronMatrix
    • 3 years ago

    OMG, they pulled the trigger on those prices… They are out for Intel’s blood… And there’s little Intel can do since Ryzen is equal or faster…

    Intel can either drop prices AND lose share or just lose share.. Tough call… But this GUARANTEES that AMD will have 30% of retail by the end of the year and maybe 50% of mobile…

      • maxxcool
      • 3 years ago

      Lol…

        • freebird
        • 3 years ago

        I’m a big AMD fan, but they are not going to get anywhere near those numbers this year…
        if retail means Dell, HP & Lenovo and their mobile parts will only be at most 4core zen with 8MB L3 and probably vega cores for the APU. AND NOT AVAILABLE until 2nd half of 2017…

      • Firestarter
      • 3 years ago

      this doesn’t guarantee anything, even if we enthusiasts get wildly .. enthusiastic that doesn’t necessarily make a huge impact on the rest of the market

        • GrimDanfango
        • 3 years ago

        We enthusiasts presumably comprise the bulk of the audience that buys Intel LGA 2011 “enthusiast” chips.
        I get the feeling that end of the market could be hit hard by Ryzen.

        …or, it could not. Who knows!

      • Pancake
      • 3 years ago

      Can I have some of what you’re drinking? It seems to be having a much better effect than the Stella Artois I’m currently chugging.

      Man, I’m so sick of ales everywhere. More lager options please.

      • Klimax
      • 3 years ago

      Intel can sustain prolonged price war. AMD cannot.

        • GrimDanfango
        • 3 years ago

        I get the feeling Intel will stubbornly refuse to price-war… I really wish they would though. As excited as I am for Ryzen, I need a platform I can stick 128GB of RAM into… so Intel effecting a knee-jerk price reduction on their 8-core chips would be most welcome 🙂

          • Krogoth
          • 3 years ago

          Intel will mostly like invoke some signficant price cuts on their Socket 2011 platform. The current Socket 1511 stock may see some sight changes but nothing earth-shattering. You are going have to wait until AMD puts out Zen-based “A” series chips for price pressure to happen in that market.

            • GrimDanfango
            • 3 years ago

            It’s the 2011 platform I’m interested in… I certainly hope they do feel compelled to cut prices. It’d help offset the insane price it seems I’ll be forced to pay for RAM at the moment.

            • freebird
            • 3 years ago

            I doubt Intel will drop their base prices at 1st. More likely they will run “marketing specials” such as short-time sales and bundling discounts that they write off as marketing expenses. A wholesale price drop drains them of revenue tied up in PC manufacturing contracts with the likes of Dell, HP & Lenovo, which Ryzen isn’t going to “steal” much business from in the short term.

            A much bigger concern to Intel will be how this plays out in the server market. That’s another reason why AMD is emphasizing the Ryzen 8-core vs (Intel Xeon workstation) 8-core comparisons; if the can convince the industry that they are on equal footing and have some added features (encrypted memory and security co-processor) for the cloud, it could be a big selling point for the Snowy Owl/Naples release supposed to come in Q2, but probably very close to mid-year.

        • maxxcool
        • 3 years ago

        This, this and THIS.

      • BaronMatrix
      • 3 years ago

      You guys are YUGE suckers… We shall see… Did you miss the part of the presentation where Su said 99% of CPUs are purchased for UNDER $500…

      I thought prices that low might be revenue-limiting but R7 1700 may be the best selling chip this year…

        • LoneWolf15
        • 3 years ago

        I admire your enthusiasm but you have no head for business.

          • chuckula
          • 3 years ago

          Baron is, shall we say, “famous” in a way.
          He’s actually cited as one of the reasons that Tom’s Hardware (back when Tom was actually there) imploded.

          If everything he said was true, we wouldn’t be talking about AMD attacking Broadwell because Intel would have gone out of business 12 years ago.

            • BaronMatrix
            • 3 years ago

            It pretty much all was and I never implied they would go out of business…

          • BaronMatrix
          • 3 years ago

          Because I know any move Intel makes will lose them money…? Or because you are still the BROOD…?

      • Kougar
      • 3 years ago

      I think you are forgetting a few things. Intel already makes a 10-core chip, and sells chips ranging up to 24 cores for servers. Intel could easily win 100% of the benchmarks if it simply offered more cores than AMD at similar prices, however Intel loves it’s margins too much that it probably won’t do this. It also doesn’t want to risk undermining its Xeon sales with cheap 10+ consumer parts. AMD doesn’t have that constraint.

      But the point is Intel has that option, and it could squash AMD with core-count if it truly wanted to.

      • BaronMatrix
      • 3 years ago

      SUCKERS…. Almost all pre-orders are sold out and Ryzen is at the top of Amazon’s best selling CPU list…

    • Tristan
    • 3 years ago

    AMD made pretty big mistake. Most users do not need expensive 8C CPU right now. These 80 ready mobos and over 100 shops waiting for pre-order look hilarious. Looks like all of this was prepared for much more popular 4C CPU, bur AMD failed to prepare these CPU-s on time.

      • tacitust
      • 3 years ago

      Not really. This way, they will generate a lot of buzz and headline benchmarks over the next few weeks, and early adopters tend to be the least price sensitive consumers out there, so there will be no shortage of customers.

      • Pancake
      • 3 years ago

      Most people do not NEED an expensive 8C CPU. But I live in a first world country and have a fair bit of disposable income. So, an 8C CPU is what I WANT.

      MOAR COARZ!!!

      • ronch
      • 3 years ago

      If you are out for a 7700K wouldn’t you be very tempted to add just $50 more for a 1700X? I know I would be. Folks who buy $300+ CPUs aren’t strapped for cash.

        • Dazrin
        • 3 years ago

        I’m more likely to drop down to the 1600X and only get 2 more cores and close enough performance but save $70-$90 depending on which price leak is accurate. It all depends on reviews of course.

      • BaronMatrix
      • 3 years ago

      Umm, but if you look at the die shot the quad would be much easier… The base idie is quad so I doubt AMD will be deactivating to make quads – especially since Raven Ridge needs a quad not a deactivated octo…

      I guess they could waste a quad for every quad, but that sounds stupid…

      This also may mean if you wait for the retail launch one my come with water… Just priced at MainGear and CyberPowerPC… Not bad prices… But they can get a real premium on these chips…

      Blood in the water… The sharks are circling…

      • Laykun
      • 3 years ago

      The fact that these cores supposedly compete well clock for clock with intel cores means that the quads themselves are going to be quite competitive, the pricing will obviously be the determining factor.

      Personally I miss having my old 6 core 980X and would love to switch over to an AMD 6 core as opposed to having to step up to intel’s enthusiast platform which is always a generation behind their mainstream platform. I think AMD has a strong range here by having all CPUs on the same socket.

    • Dazrin
    • 3 years ago

    WIki says the 1600X and other R5 and R3 chips are scheduled for Q2 2017, has anyone seen something different? I finally have funds in hand for a new computer and had been aiming at an i5-7600K or even i7-7700K build for a while but since I had some delays I might as well wait till March but I really don’t want to wait till June.

    • adisor19
    • 3 years ago

    And now for the question on everyone’s mind : will Apple replace the CPU in the MacPro trashcan with these ?

    It would certainly encourage AMD with such a big win and would also be signal to Intel to play nice.

    Adi

      • K-L-Waster
      • 3 years ago

      I assume that by “everyone” you mean “all 3 people who want a new MacPro”?

        • davidbowser
        • 3 years ago

        *Raises hand*

        Make that 4.

      • Redocbew
      • 3 years ago

      After being neglected so long I’m not sure why Ryzen would suddenly spur Apple into action on upgrading the MacPro. It’s obviously not an issue of performance, and an issue of cost seems unlikely as well. I’m sure a company the size of Apple would have plenty of tricks available for making Intel “play nice” if they had any real inclination to do so. All indications point to Apple just not caring enough to actually make it happen.

      • maxxcool
      • 3 years ago

      Not even a chance in hell.

      • NTMBK
      • 3 years ago

      I suspect that Apple is more interested in the HPC APU. Zen cores, Vega GPU and HBM on an interposer, with a unified memory space; sounds ideal for an integrated system like the Mac Pro.

    • CScottG
    • 3 years ago

    Hmmm.. Why am I’m reading “Cinebench” and thinking “Ashes of the Singularity”?

      • defaultluser
      • 3 years ago

      Because you’re a fool who can’t read that they provided single and multi-threaded benchmarks. They maintain their performance advantage in both.

      And while Ashes was a purpose-built DX12 benchmark with no redeeming gameplay, Cinebench is a real application used by major movie studios. It’s just popular as a benchmark because that’s all people like you use it for. Their product page claims several recent CGI-heavy titles to their credit.

      [url<]https://www.maxon.net/en/products/cinebench/[/url<] Their 3D render is fairly-optimized mix of vector and scalar code, which is why Piledriver 125w eight-cores are still able to exceed a Core i5 when all threads are loaded, even though the Core i5 has 60-70% higher IPC. You know, MIXED rendering code like most video games? This is why Zen seems perfect: they improved the IPC enough to be close in single-thread, while cutting costs in areas like load/store and AVX throughput (both are half of Haswell, on-paper).

        • CScottG
        • 3 years ago

        ..no, I was thinking more in terms of:

        -promo part deux.

        ..btw, nice people skills.

    • WaltC
    • 3 years ago

    Ryzen is great for two reasons:

    (1) It will reveal to Intel customers just how Intel has been gouging them with Core 2 (since 2006–was it?)…;)

    (2) It’s going to force Intel to drop the bottom out of its current cpu lineups to stay competitive; and we’ll get to see whether Intel has been sitting on any CPU architecture announcements because the lack of competition on the high end from AMD allowed them to milk Core 2 for the last decade or so…;)

      • K-L-Waster
      • 3 years ago

      What do you say we wait for actual test results and a quarter or 2 of sales before declaring total victory for Team Red/Orange, hmm?

        • Redocbew
        • 3 years ago

        Nonsense! Ludicrous! Inconceivable!

        • Thbbft
        • 3 years ago

        Why? – !!!! 52% IPC increase !!!!

        Lisa Su flat out declared the 1800x the fastest 8 core processor in the market today. AMD now rules the IPC roost. Benchmarks will reflect that fact. So will sales.

          • thedosbox
          • 3 years ago

          You do remember that AMD claimed Bulldozer was a 50% improvement too right?

          [edit] citation:

          [url<]http://www.anandtech.com/show/5057/the-bulldozer-aftermath-delving-even-deeper[/url<]

          • Jeff Kampman
          • 3 years ago

          [quote<]AMD now rules the IPC roost[/quote<] lol no

            • tipoo
            • 3 years ago

            “Set expectations to *five* buffalo” – Literally Jeff Kampman

          • ultima_trev
          • 3 years ago

          1800X won’t be faster than 6900K in every benchmark. And there’s a good chance if it is faster over all it is because of the higher base clock, 3.6 versus 3.2. There’s also boost, 4 GHz (4.1 with XFR) boost versus 3.7 (4.0 on a single core) for the 6900K. I wish I knew what the all-core turbo was for Ryzen as the 4+ GHz boost is likely on a single care.

          • K-L-Waster
          • 3 years ago

          I can’t tell if that’s irony or you really believe that hype.

          • Redocbew
          • 3 years ago

          The reference point is Excavator, correct? Excavator was a turd. That makes Ryzen a turd and a half? Just imagine how good it would be if it was worth two turds! Holy crap!

      • Gasaraki
      • 3 years ago

      WTF are you talking about? The ‘Core 2’ series are totally different than the ‘Core iX’ architecture. They haven’t been “gouging people since Core 2” and “milking Core 2 for the last decade”.

      The first i7-920s came out at the end of 08 and they blew everything out of the water including Core 2 series chips (Core 2 Duo and Core 2 Quads). They blew everything AMD had also. So how much faster is the new 7700K vs. the i7-920 in terms of raw speed and IPC? I’m guessing over 50% faster in IPC. The 7700K has also a faster base clock, lower TDP (91W vs. 130W), higher turbo, new instruction sets, built in graphics, etc. Guess what, the i7-7700K is around $300, same as the i7-920 when it came out, so even if you went with inflation the 7700K is not more expensive for one of the top performers.

      Now it took AMD 10 years to make something that matches the 6900K. Nice job. Now all Intel needs to do is drop prices of the 6900K to match.

        • AnotherReader
        • 3 years ago

        You are both wrong: Nehalem and Westmere were evolutions of Core 2. Sandy Bridge was a [url=http://www.realworldtech.com/sandy-bridge/<]completely new microarchitecture[/url<] and even Sky Lake is a descendant of Sandy Bridge.

    • Takeshi7
    • 3 years ago

    AMD tested the Intel platform with dual channel memory instead of quad channel memory. So the Intel performance probably could have been a little higher.

    [url<]http://www.tomshardware.com/news/amd-ryzen-7-1700-1700x-1800x,33702.html[/url<]

      • Waco
      • 3 years ago

      Sigh. It’s always something, isn’t it? AMD shoots themselves in the foot for every single comparison like this when they pull this BS.

        • K-L-Waster
        • 3 years ago

        <Facepalm />

        Why why why why why? And what do you expect will happen when sites start doing actual reviews? Sheesh.

          • Waco
          • 3 years ago

          I was sitting here hoping they’d learned a lesson…but no.

          They’ll claim they wanted it to be an equal platform, but all they had to do that was run 4 damn sticks of memory in both.

          UGH.

        • cynan
        • 3 years ago

        Actually, I don’t have too much of a problem as RyZen is apparently dual channel memory configuration, making it a more apples-apples comparison in some ways. And you do need double the dimms for quad channel (potential making platforms more expensive, etc).

        While it may have been a nice change for AMD to err on the side of penalizing themselves instead of the competition, is this really a reasonable expectation for a PR campaign?

        And, of course, anyone basing performance of a new CPU on PR claims, rather than waiting for 3rd party in depth reviews is either desperate or extremely naive.

          • Waco
          • 3 years ago

          I agree entirely (waiting for reviews) but AMD should know better than to even have the [i<]appearance[/i<] of cheating on the benchmarks. They've been dishonest in their comparisons in the past, and this seemed like a bit of a turning point.

            • freebird
            • 3 years ago

            How is it DISHONEST when the posted the EXACT configurations? That are EXACTLY IDENTICAL except for motherboards??? Not to mention, if there was a huge difference between 6900x scores in any of the benchmarks with dual-channel vs. quad channel configs I would imagine someone would have already brought it up all the comparisons floating around on the web. It probably falls within the margin of error of testing approx. 1-2%

            Sorry, I get tired of all the nit-picking and claims of unfairness…
            pour me some more crocodile tears…

            • Klimax
            • 3 years ago

            Because they pulled once again pure marketing BS. Small print like every other dishonest advertising. And using unusual and worse configuration then usual.

            if they altered things to win here, how other benchmarks were altered to win them? (And any open source bench is prime candidate for some fiddling…)

            • freebird
            • 3 years ago

            Keep dreaming…

            • Waco
            • 3 years ago

            It matters because all they had to do was not pull an AMD marketing trick. They could have come clean with a realistic comparison.

            No, it’s not huge. That’s not the point though.

          • jihadjoe
          • 3 years ago

          They are two different chips to begin with so it won’t ever be apples to apples anyway. A huge reason X99 is so expensive is because of the extra hardware needed to get 40 PCIe lanes and quad channel memory, so denying those while claiming the price advantage seems a bit disingenuous.

        • Lugaidster
        • 3 years ago

        I understand the sentiment, but I don’t really share it. They did send the testing configuration after all, it’s not like they are being dishonest with it about how they achieved those numbers. For what is worth, according to the review of the 6900K at Anandtech, using quad-core memory, it still ends up being slower than the Ryzen CPU and it still has a 40-50W higher TDP. Also, on the floor, there were several quad-channel with 32 GBs configured i7 to test there (You can even see it on the Linus Tech Tips video).

        I understand the scepticism, and I share it, but while I’m not aboard the hype train just yet, I’m also not going to fixate on PR talk.

      • chuckula
      • 3 years ago

      That’s sort of dumb simply because I highly doubt having 4 memory channels active would have had any real positive effect in that benchmark.

      Hell, AMD could have loaded that Intel rig up with RAM, shown basically the same results, and then use it as a marketing point that Intel overcharges for RAM or something along those lines.

        • freebird
        • 3 years ago

        Right and if loading it up with 4 DIMMs meant they had to add a CAS latency to the Intel system then there would’ve been complaints… maybe Intel should’ve just came out with a cheaper dual channel 6 or 8 core processor a year or two ago for $500 and took all of AMD’s joy away…

      • AnotherReader
      • 3 years ago

      They shouldn’t have done that, but the picture doesn’t change much. [url=http://www.anandtech.com/show/10337/the-intel-broadwell-e-review-core-i7-6950x-6900k-6850k-and-6800k-tested-up-to-10-cores/6<]AnandTech's review of Broadwell-E[/url<] shows that the 6900k has a [url=http://images.anandtech.com/graphs/graph10337/81824.png<]Cinebench score of 1547[/url<] with 4 channels populated with DDR4 2400.

        • Waco
        • 3 years ago

        Right, but why leave room for doubt? I’m glad the scores don’t move much; I didn’t think Cinebench was particularly memory-bandwidth heavy, but I wasn’t sure.

          • AnotherReader
          • 3 years ago

          I agree. I get why they hobbled the 6900k, but they could have done by having two 6900k systems: one with two channels populated and the other with all 4 channels populated.

        • derFunkenstein
        • 3 years ago

        which makes the decision to not doing it right kind of odd, don’t you think? Why do we need to theorize about benchmark cheating by AMD if they could have set the system up correctly and still won?

        • tacitust
        • 3 years ago

        No doubt they tested it both ways, just to make sure they wouldn’t be called out for it if the Intel forged ahead, but really, I’m fine with them comparing like with like. It’s their announcement press conference. Let others bluster about unfair comparisons — AMD can always come back with another chart touting their massive cost per gigaflop advantage.

      • ultima_trev
      • 3 years ago

      That explains Intel’s missing IPC.

      • Tristan
      • 3 years ago

      Cinebench for 6900K with quad is 185 / 1578, while on AMD show 162 / 1574.
      [url<]https://www.overclock3d.net/reviews/cpu_mainboard/intel_6950x_6900k_6850k_cpu_review/9[/url<] And 6900K is turbo-ed to max. 3.7GHz, while Ryzen have 4GHz and with XFR even more Despite that Ryzen is very good CPU, AMD can't resist to spread false claims.

        • Lugaidster
        • 3 years ago

        I agree with the sentiment, but disagree with the wording. It’s not false claims if you are sharing the specifics of the test. It’s careless yes, but it’s not false claims, not if you say upfront what the test conditions are. Independent reviews will tell the picture in the end.

          • Klimax
          • 3 years ago

          None of those slides with results have that. It is stuck somewhere irrelevant. It is pure BS action by AMD. Sorry, your defense failed.

        • ultima_trev
        • 3 years ago

        Anand Tech showed the 6900K at 153 / 1547:

        [url<]http://www.anandtech.com/bench/product/1729[/url<] I'd guess that OC3D sample was overclocked well beyond stock...

          • jihadjoe
          • 3 years ago

          Most of the previously posted reviews were from early 2016 though, and a newer version of Cinebench R15 has been released since which gives slightly better multithread scaling.

          [url<]http://www.legitreviews.com/intel-core-i7-6900k-processor-review_191040/7[/url<] This is the most recent review of the 6900k I could find, and uses the latest version of Cinebench R15 (.038). The 6900k does 148/1590. If LR's sample were to do 162 single thread like in AMD's setup I imagine the multi thread result would be close to 1700.

        • Checowsky
        • 3 years ago

        Check the specs slide, it has TBM 3 on, so the Intel CPU is running at 4GHz.

          • chuckula
          • 3 years ago

          That’s not necessarily true at all.
          That spec slide is simply stating that TBM 3 is a feature.

          There is no documented evidence that AMD downloaded the driver, and ran the manual setup (yes, there’s a manual setup, TBM 3 is *not* a plug-n-play feature) to select the core that should receive the 4 GHz boost [b<]AND[/b<] to select that the TBM 3.0 turbo boost should be applied to that particular Cinebench program. You should read up about TBM 3 here: [url<]http://www.anandtech.com/show/10337/the-intel-broadwell-e-review-core-i7-6950x-6900k-6850k-and-6800k-tested-up-to-10-cores/2[/url<] The funny thing here is that I don't even think TBM 3 is a particularly good or useful feature, but all of the sudden it's getting this great appreciation from the same people who think Broadwell is a piece of crap.

          • jihadjoe
          • 3 years ago

          TBM 3 is for single core turbo though, so while it may affect the single threaded score it won’t do much for the multi-threaded result.

          Edit: Assuming near-linear scaling for frequency though, having TBM3 does nicely explain the high single-threaded result:

          Legitreviews single thread 6900k: 148 (assumed 3.7GHz)
          AMD single thread 6900k: 162 (assumed 4GHz)

          [b<]3700[/b<]/148 = 25 25 * 162 = [b<]4050[/b<] So the difference in single threaded score could be very neatly explained by TBM3.

      • ronch
      • 3 years ago

      Yeah, that’s troubling. But I imagine most folks would still be OK with it because Ryzen is also much cheaper and as some have pointed out, having quad channels doesn’t change the numbers much, although that’s really dependent on what you’re doing. And it does make for an apples to apples comparison though, which says quite a bit about Zen’s robust design. In the end I think most folks will still get Zen because they’re more affordable and they can live with the compromise. If anything this just puts a dent on AMD’s honesty/credibility, but I imagine they’re not alone.

      • Kougar
      • 3 years ago

      That figures. Probably did it just so they can claim the 1800X is 4% faster instead of ~10% slower.

      Not like it matters since the 6900K is still never going to be $500 faster. There’s nothing to gain except empty bragging rights, yet AMD’s marketing department once again proves it hasn’t changed and can’t be trusted to provide any sort of reliable numbers.

      • Spunjji
      • 3 years ago

      Show me a benchmark where that matters at all and you will be showing me a benchmark they didn’t run.

        • freebird
        • 3 years ago

        They can’t that is why you hear all the whining and complaining… it is just a feather to stick in their $1000 CPU cap that now is match by a $500 AMD hat…just another straw man argument.

    • Dieter
    • 3 years ago

    I haven’t had an AMD system in years, but I do find this very promising. I think the processor will be good; my concern is the chipset.

    Are there still lots of issues with AMD chipset compatibility like there were in the bad old days? I think the last AMD system I setup was in the early XP days.

    If the chipsets are good, then I think Intel has some competition.

      • DarkMikaru
      • 3 years ago

      Honestly, I’ve found that to be an unfair myth. As someone who’s literally built hundreds of AMD based machines from Socket A up to FM2+. I can vouch for the chipsets working just fine. I think the only chipset related issue I can remember over the years was back on Socket A or 754 where the VIA KM chipsets didn’t play well with my Sil80 raid controller. Wouldn’t even boot if it was plugged in. Short of that, and considering how long ago that was…. yeah… their chipsets are fine.

        • Airmantharp
        • 3 years ago

        There have been plenty of various AMD chipset issues over the years, though none lately- however, this new chipset was not produced by AMD (the old ATi chipset people), so you cannot say ‘yeah… their chipsets are fine’, because you, like everyone else including the TR staff about to review them, simply do not yet know.

        • Dieter
        • 3 years ago

        Yeah, it was a long time ago, and at the time I don’t think AMD made their own chipsets, so the third party vendors (VIA, nVidia, maybe others) were the only option, and that spoiled it for me.

        I’m hopeful that there will be good chipsets available so that they’re simply not a concern.

        Thanks for the feedback.

          • Airmantharp
          • 3 years ago

          When they bought ATi, they got the best of the third-party chipset teams; but they still had occasional problems.

          What makes this worse though is that this chipset was farmed out to a third-party.

    • Thbbft
    • 3 years ago

    !!! 52% IPC increase !!! …..”AND WE’RE JUST GETTING STARTED” …..

    If I’m Intel that is the most terrifying line in the slides.

    [b<]INTEL IS LOOKING AT LEAST A 4 YEAR STRETCH BEING THE PERFORMANCE UNDERDOG.[/b<]

      • chuckula
      • 3 years ago

      [quote<]INTEL IS LOOKING AT LEAST A 4 YEAR STRETCH BEING THE PERFORMANCE UNDERDOG.[/quote<] You pro-Intel fanboy! THERE WON'T BE AN INTEL IN FOUR YEARS!

        • Thbbft
        • 3 years ago

        That made me laugh.

        • ronch
        • 3 years ago

        When you start hearing about rumors about Carlos Ghosn and Intel, yeah…

      • Metonymy
      • 3 years ago

      … and we’re just getting started… 🙂 …. yeah….

        • K-L-Waster
        • 3 years ago

        Well, next year they might bump it to 53%….

      • tacitust
      • 3 years ago

      I have zero problem with Intel getting a massive kick up the backside from AMD. Healthy competition can only benefit us consumers.

        • BurntMyBacon
        • 3 years ago

        Same here. Though this feels more like a bratty little kid giving Intel a swift kick to the shins. The question is what happens next. Does Intel:

        A) Roll on the floor giving AMD a chance for a groin shot.
        B) Pound AMD in the head and smash its face in the dirt.
        C) Stumble a bit grabbing a ball bat as AMD runs away laughing while blissfully unaware.

        Note: The bratty kid remark is given here for visual effect and denotes the obvious size difference between AMD and Intel. It is in no way representative of the management or character associated with either company. Any similarities to reality are purely coincidental and subject to the interpretation of the reader.

    • DarkMikaru
    • 3 years ago

    For the first time in a long long time i’m actually excited about my new AMD build. Going to put my i5-6600k on the auction block. Last I was this excited was building my first AMD machine back in 2003 when I chose the Athlon 1800+ @ 1.53Ghz as the heart of my poor mans build. 512MB DDR Ram, 80GB Maxtor HDD, 300w Antec TruPower PSU living in an Antec SLK2600AMB case. Gosh… why do we remember such things.

    Anyway… bravo AMD…well done….. putting my Ryzen parts list together as we speak!

    • Unknown-Error
    • 3 years ago

    52% increase in IPC? Even if that is very workload specific, AMD really after along time managed to not just deliver but actually go beyond what they promised. This is a huge turn of events when looking at the constant f—-ups the past 5-years in both CPU and GPU.

      • Waco
      • 3 years ago

      I’ll have to wait and see what third-party tests look like, but yes, this is better than AMD was claiming in the past.

      Have they learned to not overhype things?

    • Tristan
    • 3 years ago

    Pretty large die, looks like they wastem much of space. At the same time, they point for small core space as advantage. Pretty interesting.

      • Krogoth
      • 3 years ago

      It is comparable in size to the Broadwell-E chips.

    • Firestarter
    • 3 years ago

    if the benchmark results add up to be as exciting as they are hyped up to be now, I’d feel almost morally obligated to buy one just to reward AMD for sticking it to Intel

    • GrimDanfango
    • 3 years ago

    Has there been any definite information yet regarding the x370 chipset PCIe lanes configuration?
    Speculation seems to be placing it at 16 lanes to the chip, and an extra 8 gen 2 lanes to the chipset, but that seems oddly stingy.
    Also, is there any hints of “enthusiast” motherboards coming with 8 DDR4 slots, or is the chipset limited to 4?

    I’m very much in the market for building a new workstation, but I need it to house a GPU, x8 10GbE adapter, a sound card, and ideally an x4 quad-channel USB 3.0 card for running the Rift. I’d also like to whack in 128GB of RAM.
    It’d be marvelous to get 6900K performance for half the price, but I’m getting the distinct feeling the platform is going to fall short of my needs.

      • chuckula
      • 3 years ago

      This has already been pretty much confirmed.
      You get 16 lanes of regular PCIe to the CPU for things like GPUs or other add-in cards.

      There technically is another 4 lanes that can go to M.2 slots and the like (depending on motherboard).

      The rest of the PCIe connectors go to the southbridge, and that southbridge then provides more PCIe connectivity.

      It’s very similar to a standard desktop motherboard from Intel.

        • GrimDanfango
        • 3 years ago

        And is that southbridge confirmed as being limited to maxium of 8x Gen 2 lanes? That’d still leave me 5 short. Also, are those 16 Gen 3+8 Gen 2 definitely *in addition* to NVMe and Sata connections?

        Looks like best case, I’d have to cannibalize my GPU’s allocation down to 8x – I’m hoping to whack a 1080 ti in there at some point, and I’d guess 8x Gen 3 might actually be a bottleneck for such a card…?

      • RAGEPRO
      • 3 years ago

      X370 offers 20 lanes from the CPU and 12 lanes from the PCH, for 32 total.

      Summit Ridge’s memory controller is dual-channel.

        • GrimDanfango
        • 3 years ago

        Oh, well that sounds a little more reasonable on the lanes front.

        So, RAM would effectively have half the bandwidth compared to Intel quad channel, and limited to a max of… 32? 64?

          • Krogoth
          • 3 years ago

          64GiB of unbuffered DDR4 memory

            • GrimDanfango
            • 3 years ago

            Hmm…

            So in other news, is there any word of a separate Ryzen workstation-targetted platform, to compete with LGA 2011-3?

            Overall Ryzen’s looking like a hell of a deal for most… I’m just sad it won’t *quite* be enough for heavy duty VFX sim.

            • freebird
            • 3 years ago

            Yeah, it is the server version called Naples…they’ll probably have a workstation version i.e. 8 or 16 cores…but you’d have to use server memory…

            • Waco
            • 3 years ago

            You can buy 32 GB modules and run them in quad-channel, no?

            • Krogoth
            • 3 years ago

            It is dual-channel with four DIMM slots.

            4x16GiB = 64GiB of memory just like current Skylake and Kaby Lake chips.

            32GiB DIMMs and larger are registered and load-reduced only a.k.a you need special Socket 2011 boards for them.

            • Waco
            • 3 years ago

            Ah, gotcha. It wouldn’t be uncharacteristic for AMD to allow registered/buffered DIMMs on a desktop platform though.

            Time will tell. 🙂

            • the
            • 3 years ago

            32 GB DDR4 DIMMs [i<]currently[/i<] shipping are either registered or load reduced as you correctly point out. 32 GB unbuffered DDR4 DIMMs are part of the JEDEC spec, just that there are no companies shipping such modules (yet). Considering that 16 GB is a main stream amount now with 32 GB easily obtainable and 64 GB for those systems with four slots, there is little demand increase DIMM capacity. The markets that can utilize more memory can afford to move to socket 2011(-3) where they can get more memory slots and/or support for registered DIMMs. It'll be a few years before the 32 GB modules do ship.

            • GrimDanfango
            • 3 years ago

            I’m a market that can utilize more than 64GB… and going for LGA 2011 is going to clean me out! Kitting out a home VFX studio requires some serious enterprise-market bargain hunting.

            I just sorta hoped maybe AMD would come to my rescue 🙂 Entirely understandable that they wouldn’t be aiming quite for my particular niche though.

            • RAGEPRO
            • 3 years ago

            On a Broadwell-E chip, yeah. Summit Ridge’s memory controller only has 2 channels. As far as I know unbuffered DDR4 only goes to 16GB/module, so 4 sticks = 64GB.

    • Pancake
    • 3 years ago

    She looks like a female version of Jen Hsun. Hopefully with the same kickarse attitude. You go girl! So Su me.

      • tipoo
      • 3 years ago

      She seems to have executed very well. Lots of people involved of course, but it’s nice to have a seemingly competent CEO.

        • TwoEars
        • 3 years ago

        Agreed. And she’s an EE (MIT), always nice to have a real engineer running a tech company rather than one of the bean counters or deal makers.

        [url<]https://en.wikipedia.org/wiki/Lisa_Su[/url<]

          • tipoo
          • 3 years ago

          Yup, that’s one parallel to Jen Hsun as well, Nvidia is very much an engineer led company.

          • ronch
          • 3 years ago

          Deal makers? You mean guys like Donald Trump?

            • Anonymous Coward
            • 3 years ago

            Funny thought, if Trump ran AMD, then TR would come under attack.

            • tipoo
            • 3 years ago

            FAKE Reviews! SAD! Bigly bigly! Incoherent word vomit!

            • chuckula
            • 3 years ago

            You know, that does sound exactly like a certain contingent of people when the don’t like how TR’s reviews of AMD products turn out…

            • K-L-Waster
            • 3 years ago

            “So you know what? We’re going to build a wall. A really big paywall, with barbed wire, and it’s gonna be great. It’ll keep all those fake unfair reviews out with all those nasty impartial sites. And we’ll make Intel and NVidia pay for it. You won’t pay a dime. They, they’ll be happy to pay for it.”

            • derFunkenstein
            • 3 years ago

            Kinda like Roy whats-his-nuts when the Fury [s<]X[/s<] [edit: it was the Nano] got very limited review samples. [url<]https://techreport.com/news/29011/updated-amd-vp-explains-nano-exclusion-apologizes[/url<]

      • ronch
      • 3 years ago

      Nice biceps too! 😀

        • Pancake
        • 3 years ago

        Fantasy fight night: Su vs JHH in a cage.

        • Redocbew
        • 3 years ago

        If you’re ogling the AMD CEOs biceps, then you’ve just gone from being excited to weird.

          • ronch
          • 3 years ago

          Oh sure.. I’m ogling them here on my tablet. Heh. Chill, dude. It was just a reference to Jensen’s always flashing his CEO muscles.

            • tipoo
            • 3 years ago

            He can bench 500 using CEO Math®

    • ronch
    • 3 years ago

    The forgotten CPU company is forgotten no longer.

    #MakeAMDgreatagain

      • chuckula
      • 3 years ago

      MAGA! MAGA!

      We’re going to build a processor with YUGE performance and we’re going to make Intel pay for the R&D!!

        • Mr Bill
        • 3 years ago

        Que?

      • tipoo
      • 3 years ago

      I can’t wait to finally be able to buy the 28nm VIA processor!

    • ronch
    • 3 years ago

    Let’s not forget about Rory and Jim, guys. And Paperlauncher too. Those guys have a lot to do with what we are seeing today. And of course, every engineer and architect that toiled tirelessly throughout Zen development.

    Amidst the fanfare let’s remember the guys who tweaked and evolved the Bulldozer architecture too. I’m sure AMD learned a ton of things from their efforts there, heavily tweaking a core that’s difficult to float, all those high density libraries, all those tweaks to improve efficiency. Intel learned a lot from the doomed Pentium 4, and AMD learned a lot from the doomed Bulldozer core too, probably much more than they ever learned from any other core in their history.

    Despite the failure of the Bulldozer architecture I still would like to give kudos to the folks who worked on it. Designing a CPU core is grueling, and everyone who’s worked on one of those things deserves a pat on the back. This is rocket science, folks. Those guys ROCK.

      • Jeff Kampman
      • 3 years ago

      Dude, calm down. It’s not the second coming.

        • chuckula
        • 3 years ago

        WE REFUSE TO BE CALM!

          • davidbowser
          • 3 years ago

          Did you borrow SSK’s keyboard?

            • drfish
            • 3 years ago

            Did you get the memo?

            • davidbowser
            • 3 years ago

            Ahh, yes. I meant SSK’s [b<]OLD[/b<] keyboard.

            • meerkt
            • 3 years ago

            SSK’s using dual keyboards.

            • Neutronbeam
            • 3 years ago

            ICYMI–the keyboard he won in a drawing here at TR had no CAPS capability–general ironic hilarity ensued.

            • the
            • 3 years ago

            Speak for yourself, I thought that that was a sign of the apocalypse.

        • ronch
        • 3 years ago

        It sure is for AMD.

        • NTMBK
        • 3 years ago

        Nah, it’s much more important than that.

        • f0d
        • 3 years ago

        it kinda is
        amd were considered dead like jesus yet here they are making a “second coming”

          • tacitust
          • 3 years ago

          The sign of an impending CPU apocalypse?

          • freebird
          • 3 years ago

          That’s why I posted in a previous “Ryzen” article comment, that AMD probably had an “Easter” release scheduled, hence the “Ryzen/Risen” name, but the chip proved too good to keep in the bag, so this year Easter came early…

          Next, stop Vega and then onto Naples… all aboard the “prophet” train… 😀

        • Neutronbeam
        • 3 years ago

        All is as was foretold…all hail the Wassoning of AMD!

        • flip-mode
        • 3 years ago

        Third?

        • tipoo
        • 3 years ago

        “Set expectations to 6 buffalo”
        -Literally Jeff Kampman

        • A_Pickle
        • 3 years ago

        It… kinda [i<]IS[/i<] though, for AMD. Besides, they're not just coming into this thing like "please oh god i hope we do well," they're like, drunkenly swaggering over to their competition's bar table and saying "u wann go m8?" Eight cores and sixteen threads at $500. 'CUZ INTEL BE HOLDIN' OUT ON US. THIS is the A.M.D. I remember.

          • K-L-Waster
          • 3 years ago

          Thing is, they’ve done that before… then found themselves flat on their backs saying “wha happen?”

          It’s a promising looking product, but it’s wayyyy early to declare it the champ. We haven’t even seen independent testing results yet.

        • maxxcool
        • 3 years ago

        ouch… dead Buffalo ?

      • flip-mode
      • 3 years ago

      I just counted, and you’ve made NINE top-level comments to this article. Dude, you need to calm the F down. It’s getting creepy.

      Edit: to put it in context, post whore chuckula has only made THREE top-level comments. You’ve shamed yourself.

        • DancinJack
        • 3 years ago

        He always does that. He gets mad when you make it known too. It’s really weird.

          • ronch
          • 3 years ago

          Hey you! Wanna step outside?

            • RAGEPRO
            • 3 years ago

            CASH ME OUSSIDE HOWBOWDAH?

        • ronch
        • 3 years ago

        What can I say? Guess I have a lot of things to say about it. I should’ve maybe consolidated them though, like I sometimes do.

      • Metonymy
      • 3 years ago

      Hmmph…. if it weren’t for the Pentium 4 and my Dell with the Prescott, I’ve have had a few much colder winters when the landlord was cheaping out on the heat…

        • Platedslicer
        • 3 years ago

        Try a tropical summer with the landlord cheaping out on the air conditioning, AND a space heater chip on top of that.

      • the
      • 3 years ago

      The main lesson of Bulldozer is that you don’t screw up your cache hierarchy which AMD did thoroughly. Second lesson is that resource sharing is OK as long as you don’t have a bottleneck. Steamroller and Excavator fixed this issue by incorporating two independent decoder blocks. Third lesson is to go full width on vector calculations: having 128 bit wide AVX units really hurt Bulldozer’s performance. Intel went full 256 bit wide in hardware and had double the AVX resources per core (I’m counting AMD’s module design as two cores for this comparison here).

      • onlysublime
      • 3 years ago

      Jim Keller is a major force and AMD was so lucky to get him back in 2012 which set up development of Zen. The architect behind AMD’s last beast of a processor, the Athlon 64, and the author of the X86-64 spec.

        • tipoo
        • 3 years ago

        When he works on a chip and then leaves it’s usually a good sign 😛

        It’s like he goes “Ok, this looks like it’s in good shape, therefore I’m bored crapless”

    • ronch
    • 3 years ago

    I’m just completely amazed at how AMD pulled this off. Plucky little AMD, mustering what resources it had left from 2012 onwards, paying for a huge debt dating back to the Hector era, practically selling about 10 FX chips a month, living on food stamps (razor thin console revenue). And now look at that die! That’s practically what you’d expect from Intel! Excuse me while I pick up my jaw from the floor.

      • f0d
      • 3 years ago

      yeah i wasnt expecting them to get as near intel as they have, if someone told me last year that ryzen has the performance it has i would of straight out called BS

      its performance has convinced me to switch from intel

        • srg86
        • 3 years ago

        I think these days the only thing that would make me switch is if they started using Intel chipsets.

      • TEAMSWITCHER
      • 3 years ago

      Intel has been shipping an IGP on their mainstream desktop processors. Ryzen doesn’t do that, and instead uses the die-space to add cores. The IPC and power efficiency gains are amazing, but we need to see more Benches – other than Cinebench R15 – to see how the processor stacks up on non-rendering workloads.

      I’m really interested in what Intel will do to respond. AMD has never held the performance crown for long. Can Intel drop the hammer again?

        • chuckula
        • 3 years ago

        [quote<]Can Intel drop the hammer again?[/quote<] I thought it already dropped the hammer on its own foot?

        • ronch
        • 3 years ago

        And I reckon AMD has never been this aggressive on pricing when they had a competitive product. More like they tried to stay close to Intel’s prices back then (K8 days). Yes, it would be interesting to see how Intel responds. Will they hold prices steady or will this be an all-out war for market share? You can bet Intel is gonna have a few dozen of these Ryzen chips in their labs, profiling it, looking for flaws, etc.

        • ronch
        • 3 years ago

        I think the real question is: Can AMD drop a Sledgehammer again?

          • Mr Bill
          • 3 years ago

          Can AMD pick up the hammer?
          [url<]https://www.youtube.com/watch?v=teRMsxMLVkY[/url<]

          • freebird
          • 3 years ago

          That would be fine, just keep Bulldozer in the shed…and speaking of CPU code names:

          Ryzen code name should’ve been Mr. Mojo, so we could call it Mr. Mojo Ryzen…

    • Krogoth
    • 3 years ago

    It is smelling more like Phenom II launch redux.

      • ronch
      • 3 years ago

      Oh puhleeeeezzz!!!

      • ultima_trev
      • 3 years ago

      Phenom II wasn’t that bad. It mostly lacked SMT and an aggressive turbo. Would have been closer to Nehalem/Lynnfield/Westmere otherwise. It did hold its own against Conroe and Penryn pretty well however.

        • Klimax
        • 3 years ago

        Aka previous Intel line-up. Sounds almost like damning with faint praise…

        • ptsant
        • 3 years ago

        Phenom II was a perfectly decent processor. Phenom I was bad.

      • davidbowser
      • 3 years ago

      Did you just preemptively “Krogoth” this launch?

      You have to wait at least until TR published benchmarks before you can be unimpressed.

      • derFunkenstein
      • 3 years ago

      You smell like Phenom II launch redux.

        • lycium
        • 3 years ago

        I can understand, a joke’s a joke etc, but doesn’t one gain additional professional obligations (like say, not making jokes about your users/readers) once you’re at the reins?

        Probably an unpopular opinion. *braces for downvotes*

          • derFunkenstein
          • 3 years ago

          Not sure what you’re saying, but I’m open to an explanation. I don’t work here.

            • lycium
            • 3 years ago

            Don’t you write articles for TR? Whoops, my bad.

            • derFunkenstein
            • 3 years ago

            Not for the last 15 months

            • DancinJack
            • 3 years ago

            Man, if someone can’t make a joke like that because they write for TR, I definitely don’t want to write for TR. Ever.

            • lycium
            • 3 years ago

            I probably seem prudish for what I said, but I’m kinda sharp-tongued myself, and sometimes find it difficult holding it in professional context. It sucks, but you do lose the freedom to wield it as you like, at least to a degree.

            Having said that, saying someone smells like a Phenom II launch is a harsh barb if ever there was one 😛

            In the same category: “You’re noisier than a GeForce FX cooler”, “You’re more full of hot air than a Pentium D”, “Your argument makes as much sense as AMD’s exclusive cache policy”, “You’re about as useful as Matrox’s EMBM”, “You’re about as honest as a Geforce 4 MX”, … but I guess one has to be of a certain age for these to burn.

            • derFunkenstein
            • 3 years ago

            Hey I like all those, too.

            • Waco
            • 3 years ago

            All of these made me chuckle. I can’t imagine anyone in the review industry getting raked over the coals for any of them.

            If they did, good riddance, that employer clearly has no sense of humor. 🙂

            • derFunkenstein
            • 3 years ago

            It was the laziest joke I’ve ever made and it wasn’t funny. It should be called out. TR writers should do better.

        • chuckula
        • 3 years ago

        Smells like… [url=https://www.youtube.com/watch?v=hj5b9_iwx-I<]Teen Shovel[/url<].

      • tipoo
      • 3 years ago

      It’s always good practice to keep expectations in check of course, but I think this is going to put them in a better market position than they have been in over a decade regardless. We know the core counts, price, and clock speeds, and they said 50% IPC improvement over Piledriver rather than their targeted 40%. If it just reaches Haswell or Ivy Bridge like IPC it should be pretty competitive, lose 10-20% per-core performance in exchange for twice as many cores and threads and all for a lower than Intel price. Much different prospect than Bulldozer, which was just flat stupid.

        • Krogoth
        • 3 years ago

        Phenom II was competitive in HEDT market for almost a year until Intel followed up with Nehaelem family which completely destroyed it and Sandy Bridge further cementing it. The same thing is likely going to happen with Ryzen. Intel has Skylake-E chips coming down the pipe which most likely retake the HEDT market again.

        AMD still has no straight answer to Intel’s normal desktop and laptop-tier chips at this time. The laptop and normal desktop-tier chips is what is selling these days via sheer volume. By the time they get A series parts based on Zen architecture. Intel will have Coffee Lake coming down the pipe and Cannonlake will be looming on the horizon. They will probably retake the power efficiency crown back which is a kinda a big deal for the laptop market.

          • flip-mode
          • 3 years ago

          Shoot, AMD may as well cancel Ryzen, it sounds like. Good lookin out Krogoth.

          • tipoo
          • 3 years ago

          Do you recon Skylake E would be had for 329 dollars?

          I think there’s a lot of space between AMD razor margins and Intel ~35% margins to play in here that’s still a win win for AMD and consumers. Bulldozer was making it such that even with such a large gap, there was little way to make it appealing.

      • DragonDaddyBear
      • 3 years ago

      So, what you are saying is you’re not impressed?

    • puppetworx
    • 3 years ago

    [quote<]Coming Soon: CPU Review Worth Reading[/quote<] Noice.

    • USAFTW
    • 3 years ago

    Intel apparently decided to put out a statement in response:
    [quote<]We take any competition seriously but as we’ve learned, consumers usually take a ‘wait and see’ approach on performance claims for untested products. 7th Gen Intel® Core™ delivers the best experiences, and with 8th Gen Intel Core and new technologies like Intel® Optane™ memory coming soon, Intel will not stop raising the bar.[/quote<]

      • chuckula
      • 3 years ago

      [quote<]but as we’ve learned, consumers usually take a ‘wait and see’ approach [/quote<] Intel confirms that we are all waiting for RyZen: CONFIRMED!

      • ronch
      • 3 years ago

      Even if Intel starts doing something desperate I’d probably still choose Ryzen out of principle.

        • f0d
        • 3 years ago

        im really liking ryzen for sure but theres no way i have any sort of allegiance to any computer part manufacturer

        if intel gets dirt cheap like ryzen they can have my money next year

        • USAFTW
        • 3 years ago

        Pretty much what I thought.
        Intel has been the dominant force for so long and has been so status quo that I would take Zen, even at the same price, just for a change.
        I don’t know what Intel will want to do. Cut prices? By how much? Cut the price of a 6900K in half would be very unlikely, but even provided that, I’d take Zen, just to support AMD and acknowledge the fact that they quirky brilliance of AMD’s engineers who don’t own an in-house fab goes toe-to-toe with Intel’s best.

          • Stonebender
          • 3 years ago

          AMD’s engineers? Uh, they hired themselves a mercenary to get the work done. We’ve seen what AMD engineers were able to do for a decade now.

          • K-L-Waster
          • 3 years ago

          I think you mean the *known* brilliance of Jim Keller — who left….

        • K-L-Waster
        • 3 years ago

        Oh, for… you *do* realize that “brand loyalty” only benefits the *brand*, not the consumer who falls for it, right?

          • ronch
          • 3 years ago

          Of course, but in an oligopoly such as this we need both companies and guess who is in danger of shutting down.

        • cybot_x1024
        • 3 years ago

        I would buy AMD because, by supporting them, I would be encouraging a price war with intel, which is good for consumers

      • derFunkenstein
      • 3 years ago

      Kinda weird that they felt a need to release a statement. I’m taking that as good news for AMD.

        • chuckula
        • 3 years ago

        It’s true.
        The last time somebody asked them about AMD their response was: You misspelled ARM there.

          • tipoo
          • 3 years ago

          “AM…Wait, the 8086 clone makers are still around?!”

            • derFunkenstein
            • 3 years ago

            Darn you, IBM and your “second supplier” requirement!

    • SiSiX
    • 3 years ago

    I really hope that AMD has a winner in this one, because while I’m vendor neutral (as ALL of you SHOULD be) when it comes to CPU’s (really, it should come down to performance, power consumption, and price), Intel really could use some competition to both keep them honest (really, their prices on the 6C/8C/10C i7’s are nuts atm) AND to keep them paranoid when it comes to moving forward. That said, there are a few things that bother me:

    1) AMD’s obsession with Cinebench and lack of pretty much anything else useful atm.
    2) Offering ‘pre-orders’ on a product that is supposed to be a “hard launch” with no reviews prior to launch.
    3) No indication of what their “95W TDP” actually means for any of their chips given the total lack of reviews. I’m running a i7-5820K which has a “140W TDP” and 6 cores. I can’t even get close to 140 watts until the chip is heavily overclocked. So is that AMD chip hitting its TDP at close to stock speeds or does it have any headroom?
    4) I really appreciate that AMD has put the effort into an 8 core chip. But on a day in and day out basis, except for rendering and video work, there just isn’t anything that even really takes advantage of even 4 core chips atm. And I mean REALLY takes advantage of them. The games that will push a 4 thread system are few and far between. So who are they really building this for given their benchmarking against is effectively a 5+ generation chip (Broadwell-E) against their new chip. Personally, I’d rather see benchmarks against i7-7700 processors in things that aren’t Cinebench or POV-Ray. I seriously hope they’re competitive, but having seen AMD’s past track record, have a feeling they may not be.
    5) Finally, that’s an awful big undercut for a company that’s been struggling financially for as long as they have for something that’s brand new. I’m all for swinging for the fences, but when you offer a product that is “faster, better, and consumes less power” for half the price…well…Grandma didn’t raise any fools here…AND you offer a ‘pre-order’? Yeah, I’ve got a copy of No Man’s Sky sitting on my SSD that I pre-ordered based on teaser videos too. 😉

      • DarkMikaru
      • 3 years ago

      All valid concerns sir but I think this time around it’s on the up and up. And why not? How do you think Honda was killing all the American car manufactures so many years ago. Cars were the same or cheaper and infinitely more reliable and less expensive to maintain. If you can produce a better product at a lower cost..that’s game over. Is this not how AMD beat Intel the first time?

        • Thbbft
        • 3 years ago

        They’re not remotely all valid points.

        If AMD officially says it’s achieved a 52% IPC increase over Excavator, then they have achieved a 52% IPC increase over Excavator. It would be legal suicide to lie about this.

          • K-L-Waster
          • 3 years ago

          Uhh, he had 5 concerns, none of which were IPC related.

          52% improvement over an underachieving predecessor is not the one and only answer to everything.

        • Stonebender
        • 3 years ago

        When did AMD ever really beat Intel? Yeah, they had a better processor for small time, but even then they still had a marginal share of the market.

          • Klimax
          • 3 years ago

          And they priced accordingly. Higher then Intel. (Black Edition Athlons)

      • Rza79
      • 3 years ago

      About point 5-
      I really don’t understand the big undercut. Why are they positioning their 8-core chips against Intel’s 4-core chips? Why not position their 6-core chips against Intel’s 4-core chips? Like you said, they sure can use the money.
      Ryzen is a huge chip. 4.8b transistors according to AMD while Skylake has only 1.75b and Broadwell-E 3.4b.
      And while on this subject, why is Ryzen so big?

        • ronch
        • 3 years ago

        They probably looked back at the RX480 launch and thought, “That was good. Let’s do it again!”

        • freebird
        • 3 years ago

        8MB of L3 cache per 4 cores and 512K L2 cache per core totals 20MB of L2 & L3 cache on the Ryzen die; that is a “large” portion of it.

          • Rza79
          • 3 years ago

          Broadwell-E has 10 cores, 27.56MB of L2 & L3 cache and 40 PCI-E lanes. Your point being?

        • Klimax
        • 3 years ago

        Because performance won’t be there. (Otherwise the only other outcome si price war and that IS suicide for AMD)

      • lycium
      • 3 years ago

      As a rendering and programming guy, these affordable 8-cores really couldn’t have come sooner. Stuff like building large codebases and rendering complex 3D scenes will hopefully do well on Ryzen, making it super attractive to us folks (and there are a fair number of us).

        • Klimax
        • 3 years ago

        It won’t. AVX won’t run anywhere close to Intel cores. (Only 128 bit execution resources and AVX instructions take up double space)

        And most of renders if not done on GPU already uses AVX. (IIRC recently Blender landed large patch for AVX) Same goes for video.

        And i don’t think it will be big win for compiling either.

      • ptsant
      • 3 years ago

      Points 1-3 essentially boil down to “wait for the independent reviews”. In a bit more than a week, we’ll know how it measures in other benchmarks, how much power it uses and whether it is available.

      Point 4 is the usual “don’t need more than 4 cores for gaming” argument. Actually, you can also do fine with 2c/4t on an i3 7350K, for example. Tests of the 7350K against the 7600K show quite clearly that they are practically equivalent for most games (see anandtech). On the other hand, one could argue that if you do have some other use (and it’s not just rendering/video but also scientific applications, development, content creation etc), your gaming experience is not really penalized from a multi-core processor. Your 5820K games just fine, even though the 2c/4t 7350K is probably faster. Finally, people have been saying that better use of multicore is coming for games and will be developed in new engines. Nobody can predict when and if it will happen, but if just 2 big studios update their engines (say Unreal and Frostbite), this will automatically mean dozens of AAA games run better with more than 4 cores.

      Point 5 is a business decision. We know that the Zen cores are comparatively small and much cheaper to make than the SOI Bulldozer cores. The price is also much higher than any recent AMD launch, even since the Phenom era. Just because Intel has extremely fat profit margins, doesn’t mean that AMD is not making a profit at these price points. What I’m saying is that the “low” (not so low to me, anyway) price of $499 is not an indication of a bad performance.

      All that being said, I would only preorder if I had a lot of money to spend and very little patience. The only thing I preorder in the last decade was Witcher 2 and it wasn’t that good. I wouldn’t pre-order a $500 product…

    • chuckula
    • 3 years ago

    Based just on the Cinebench numbers and ignoring any other factors, RyZen clearly scales pretty well out to 8 cores. On single-threading, it’s still not beating Broadwell clock-for-clock [in single thread that 6900K is at 3.7GHz max and the 1800X is at 4 GHz… or some indeterminate number above 4 GHz based on how XFR behaves] but it’s close, and AMD deserves to be commended for that since Broadwell is a world-class core.

      • derFunkenstein
      • 3 years ago

      I’ll certainly take it, though I’ve never bought a PC just to run rendering benchmarks on before. I hope I can use it for other things.

        • chuckula
        • 3 years ago

        I’ve found that buying a heavy case means it can also prevent papers from blowing away.
        So that’s two uses right there.

          • derFunkenstein
          • 3 years ago

          That’s a really solid second-purpose. RYZEN UPGRADE WAF OBTAINED

      • f0d
      • 3 years ago

      i dont mind trading in that very small performance difference (3.7ghz compared to ~4ghz similar performance) for the MASSIVE difference in price

        • chuckula
        • 3 years ago

        Assuming Intel can’t do anything with engineering, they can do things with price.
        That’s not to say that I expect the 6900K itself to experience a real price drop, but I would expect something later this year that’s more price competitive.

        • K-L-Waster
        • 3 years ago

        This. ^^ Getting in a dogfight on performance with a $1000 chip when your product costs half that will be a pretty compelling story. Even more so when you factor in Broadwell also requiring a 2011 motherboard, which usually is also pretty pricey — if RyZen motherboards are closer in price to the 270 series boards, the price difference will be even more pronounced.

          • ptsant
          • 3 years ago

          Completely agree about the MB!

          I could stomach a $500 CPU, but when you add the quad-channel RAM and the $350 MB, the price premium is just crazy for the 2011-v3 platform.

          Ryzen MB seem competitively priced. The ultra premium Crosshair ROG sits at $250-300, which is a lot, but the Asus Rampage V (2011-v3) sits at $500 and the Asus Maximus (1150) at $350.

      • ermo
      • 3 years ago

      Uh, won’t the 6900K [url=http://ark.intel.com/products/94196/Intel-Core-i7-6900K-Processor-20M-Cache-up-to-3_70-GHz<]turbo to 4GHz[/url<] on its best core in ST loads courtesy of the Intel® Turbo Boost Max Technology 3.0? That's pretty much as apples-to-apples as it gets IPC wise...

        • chuckula
        • 3 years ago

        Tell ya what, when you show the documentation from AMD that they properly configured that feature (it’s not always active, and you have to go out of your way to download the driver for it, such as going here: [url<]https://downloadcenter.intel.com/download/26103/Intel-Turbo-Boost-Max-Technology-3-0)[/url<] then I'll change my analysis. Bearing in mind that we've already seen that AMD left out 2 channels of RAM from the Broadwell system, so I'll be happy to believe that they installed and properly configured the TurboBoost 3.0 software [i<]after[/i<] there's written documentation to that effect. Additionally, "XFR" is supposed to make the 4.0GHz paper number be a lower threshold, so saying that the 1800X was "only" running at 4.0GHz means that XFR is also not really a thing.

          • Waco
          • 3 years ago

          I would highly doubt AMD jumped through the hoops to enable that and ensure it was working on a system they’d already hobbled from the start.

          • Demetri
          • 3 years ago

          The slide describing the way the systems were set up seems to suggest ITBM was enabled on the 6900K, but nothing concrete to say for certain. Could just be misleading. Says nothing about whether XFR is enabled, but even if so, that apparently only lets it boost to 4.1ghz as cooling allows.

          [url<]http://images.anandtech.com/galleries/5485/AMD%20Ryzen%20Tech%20Day%20-%20Lisa%20Su%20Keynote-33.jpg?_ga=1.75947477.179056659.1487826274[/url<]

            • Checowsky
            • 3 years ago

            Be a bit odd for them to gain 52% IPC over Piledriver and then sod up enabling ITBM.

            I mean come on, this is a bit petty at this point isn’t it? Just wait for the reviews and see. Stop equating an entire company down to a fumbling person with the name AMD.

      • Klimax
      • 3 years ago

      As long as you don’t care about AVX… (And I don’t think it has IPC close to Broadwell – too few decode resources among other things)

        • raddude9
        • 3 years ago

        There’s nothing wrong with Ryzen’s AVX implementation. All indicators thus far point to Ryzen having a full implementation of AVX.

        Maybe you were thinking about the newer AVX2? (This is also in the Ryzen core, but the 256bit instructions will be running at half speed, i.e. broken up into two 128bit instructions…. as I understand it, please correct me if I have the wrong idea).

        Also, yes, the various leaks seem to indicate that the raw IPC of Ryzen is closer to Haswell, but even if the IPC is behind Haswell (very possible), there are lots of factors that could bring the actual single-thread performance numbers above Haswell depending on the benchmark (boost clocks, cache size, faster memory).

      • Gadoran
      • 3 years ago

      Ryzen scales well with threads only because it can turbo all 8 cores at 4ghz..an more. With cinebench, that don t stress much the cores, with the right cooler this is an easy thing to obtain. Intel real sin was to never optimize Broadwell E for gamers, leaving the cautelative turbo of server cpus with low power in mind. Imo this will change with Skylake E.

    • ronch
    • 3 years ago

    Just an observation. It does seem like AMD’s multi core scaling is now better than Intel’s. Most of the attention has been focused on the Zen core itself but not much or nothing has been said about the Uncore parts. And looking at that die shot (the one over at Ars Technica is better) it seems to me AMD did a lot of work to bring their Uncore up to speed because the Uncore there looks every bit as sophisticated as Intel’s recent Uncores particularly in their higher end chips. The Uncores of past AMD CPUs from K8 all the way to Vishera all seemed lazy and simple in comparison to Intel. I can’t wait to know more about Ryzen’s Uncore.

    • USAFTW
    • 3 years ago

    Queue a particular Ron Paul meme…
    Damn, I so like to buy me some Ryzen but the darned prices are through the freaking roof over here since 2011.
    If they released this back then I could go for a 1800x/x370 combo but now the prices are up at least 3.8-fold.
    Sad!

    • 1sh
    • 3 years ago

    I am assuming the 1800x can hit 4.5ghz without much effort with the wraith cooler.
    Even though overclocking a octo core is double the trouble of overlocking a quad core.

      • chuckula
      • 3 years ago

      RyZen is clearly an overclocker’s dream.

        • ptsant
        • 3 years ago

        I really don’t think so. That would be very, very surprising.

          • chuckula
          • 3 years ago

          I think you missed my [url=http://www.pcworld.com/article/2936630/amd-reveals-hbm-powered-radeon-fury-graphics-cards-new-r300-series-gpus.html<]subtle reference[/url<].

        • Metonymy
        • 3 years ago

        Clearly

      • Airmantharp
      • 3 years ago

      If it wasn’t hard to do, you’d figure they would have shipped with such clockspeeds.

      RyZen certainly looks competitive, but they’re going to have to hit higher clockspeeds to capture the gaming enthusiast market, especially when Intel maintains a slight advantage in IPC. Hopefully their smaller parts will do just that.

        • Waco
        • 3 years ago

        Power usage gates this for the most part, and they wanted to win the performance/watt comparisons.

        I’d bet they launch a 140 watt part within the next 6 months.

          • Airmantharp
          • 3 years ago

          Not that I’d protest, but given the pricing, it looks like the ‘cake and eat it too’ part, say a 4.2GHz+ octo-core, would venture into say >US$700 territory…

          Granted, if I needed the performance, I’d certainly be up to the challenge!

      • Airmantharp
      • 3 years ago

      OC’ing is looking to be grounded more toward 4GHz. So faster than stock Intel eight-core parts, but slower than their overclocks (~4.5GHz range). Still very good performance, but obviously going to be slower than Intel quad-core parts for thread limited tasks at stock and appreciably slower than overclocked Intel quads.

    • ronch
    • 3 years ago

    So what do you think, Krogoth?

      • Krogoth
      • 3 years ago

      Somewhat impressed.

      I would be actually be impressed if Zen is as energy efficient as Broadwell/Skylake.

        • Firestarter
        • 3 years ago

        [quote<]Somewhat impressed[/quote<] [url<]http://gfycat.com/UncommonAgitatedAdmiralbutterfly[/url<]

    • Kretschmer
    • 3 years ago

    I’m scared. If AMD’s internal benchmarking is 9% faster, wouldn’t that make real-world performance more like 50% slower?

      • chuckula
      • 3 years ago

      That crazy talk!

      If AMD’s marketing department says it’s 9% faster, then we all know they screwed up the math and it’s really 29% faster.

    • blastdoor
    • 3 years ago

    I am excited.

      • chuckula
      • 3 years ago

      Blastdoor excited: AMD RYZENS IN MACBOOKS CONFIRMED!

        • blastdoor
        • 3 years ago

        In the unlikely event they were to show up in an Apple product, I presume it would be an iMac or Mac Pro. But I really don’t expect that to happen — I have revised my expectations downward a fair bit.

        I’m excited because this means that if Apple continues to neglect the Mac Pro, then I can combine an iMac with an inexpensive headless Linux box and have 8 cores sitting under my desk to run simulations all day. I’d control it from the iMac. It’s not ideal, but it’ll do.

          • chuckula
          • 3 years ago

          Don’t worry, once you get hooked on Linux, pretty soon that Linux box will be controlling your iMac… MUHAHAHAH!

          • adisor19
          • 3 years ago

          It is plausible for Apple to shove these in the MacPro trashcan. That thing hasn’t been updated in forever mainly due to its low volume, expensive price and Intel’s slow and cautious Xeon pace.

          I honestly see this as a perfect match.

          Adi

            • davidbowser
            • 3 years ago

            I have to admit that i find this idea most intriguing. The Mac Pro is all about the SMT capabilities, so that is not a bad concept.

            • the
            • 3 years ago

            The lack of update is solely Apple’s fault. They skipped over two generations of Xeons (Haswell-E and Broadwell-E) and that’s not even counting the possibility of them using Xeon D. Apple also skipped over two full generations of AMD and nVidia GPUs too.

            • blastdoor
            • 3 years ago

            Agreed.

            I also agree with davidbowser that Ryzen would be a good fit for the Mac Pro. I’m just lowered my expectations quite a bit when it comes to Apple doing things that make sense for the Mac, particularly the Mac Pro.

            • K-L-Waster
            • 3 years ago

            I seem to recall Tim Cook saying a while back something to the effect of they weren’t developing desktop Macs and MacPros because there wasn’t much money in those products.

            People who like them obviously like them alot, but there aren’t enough such people to make it a profitable product line. They make their money on the laptops and the iGadgets.

            • the
            • 3 years ago

            I don’t disagree with the idea that Ryzen would be a good fit for the Mac Pro, especially if AMD permits ECC memory. The one downside is that it would be [i<]another[/i<] lateral move in performance as replacing a 12 core Ivy Bridge-EP with a Ryzen shouldn't move performance much. Apple might be better served with the Naples processor though which is essentially four Ryzen dies in a single package.

            • blastdoor
            • 3 years ago

            Yeah, that makes sense. I could see a range for the Mac Pro from 8 cores up to 32.

            It’s a great fit.

            But Apple would have to lift a finger to switch from Intel to AMD. While Apple is willing to lift a finger for the Mac, it’s not the right finger.

        • Redocbew
        • 3 years ago

        CONFIRMED IS CONFIRMED!

    • ronch
    • 3 years ago

    I’ve said it before and I’ll say it again: it doesn’t matter how a CPU core achieves its performance, whether it’s through more IPC or more clock cycles. If the core matches or outperforms a competing product while using less power and selling for a lower price and occupying a smaller die real estate despite being built on a node that probably isn’t as dense as the competition’s and not as tuned for clock speed and efficiency as the competition’s, it’s a win-win in my book. And here the 1800X absolutely makes the 6900K’s value proposition look frickin ridiculous. Even the 6700K and 7700K, both of which occupy the ‘mainstream Golden Standard’ of Intel’ s lineup, seem gravely threatened. Even the 1700X makes a great show despite presumably not matching the 6900K’s 1T figure. Given how the 1700X can probably turn into a 1800X at the flick of a switch, I don’t see how anyone with the financial capacity can resist it. $399. Good grief. I can imagine Intel fuming right now as these articles are flooding the Internet. Star Wars Episode IV has got nothing on Ryzen. Amazing. I’m nothing short of ecstatic.

      • llisandro
      • 3 years ago

      Yup, $400 is just about what I had imagined my “sure, why not” level to be, given how long hardware retains relevance these days.

      Shame they waited too long and I already bought a 7700K 🙂

    • chuckula
    • 3 years ago

    BTW, this is a rough estimate of the die size that is made based on AMD’s public statements and the [i<]assumption[/i<] that the die shot from this event was not photoshopped to distort the relative dimensions of the chip... so take with a grain of salt. First: Here's a cleaner shot of the picture you see above: [url<]https://www.servethehome.com/wp-content/uploads/2017/02/AMD-Ryzen-Die-Shot.jpg[/url<] Second, I'm going with the numbers that AMD provided for the size of their 4-core "complex" of "44 mm^2" (example here: [url<]http://www.eteknix.com/amd-zen-smaller-core-size-intel-skylake/)[/url<] So, doing the math based on an estimate of what that 4-core complex actually covers in the chip, I'm getting an estimate of about 190 mm^2, which is pretty small (a 10 core Broadwell-EP die is about 245 mm^2). Edit: So extrapolating that out to Naples, you are looking at about 760 mm^2 of total silicon for a 4-die package. Incidentally, as of right now the biggest 24-core Broadwell parts are 456 mm^2 [using an unrealistic over-estimate of scaling up by 33% for adding 8 more cores, a 32-core Broadwell EP would be about 607 mm^2]. You'll notice how the ratio changes going from the 8-core part to the larger parts because adding more cores to the chip but leaving the uncore alone means the overall size of the chip doesn't go up quite as much when you just add more cores. We'll see how big Skylake EP parts with 32 cores get.

      • Goty
      • 3 years ago

      What structures are you considering the be part of the “complex?” Are you just splitting it in half or are you just bounding the rectangular areas that are surrounded by what are likely the IMC and various interconnects? I’d be interested to know what numbers you come up with as best- and worst-case scenarios for die size.

        • chuckula
        • 3 years ago

        Bounding rectangle around what looks like the core complex. There are two of them on display and the same bounding rectangle covered both of them pretty precisely.

        Once again, manufacturers can and do distort these die shots all the time, so this is an estimate. Having said that, I’m not expecting the chip to be substantially smaller than roughly 190 mm^2 or so (no, it’s not smaller than a 4-core Skylake part).

      • ronch
      • 3 years ago

      I imagine a grain of salt hitting a silicon die is like a huge meteor the size of Texas crashing into a city.

        • tipoo
        • 3 years ago

        Sometimes it’s amazing they get appreciable yield rates at all, eh? That would be why those are some of the best controlled rooms out there.

          • ronch
          • 3 years ago

          Yeah. You probably wouldn’t even know how it feels like to have a cold if you spent your whole life living in a fab. Heh.

      • denstieg
      • 3 years ago

      Thanks for the picture link, its a thing of beauty.

    • ronch
    • 3 years ago

    Look at that pretty die shot!! [swoons]

      • Neutronbeam
      • 3 years ago

      Yep, it’s to die for!

    • ronch
    • 3 years ago

    I’d be very interested in one of those 6-cores OCed to 4.0GHz and see how it stacks up against a 7700K in terms of single thread, multi thread, and power draw. I would gladly give up sheer core count for single thread performance.

    • Umbral
    • 3 years ago

    Fewer threads, higher clocks, sub $300.

    • kamikaziechameleon
    • 3 years ago

    I had one AMD cpu, a 1060T x6
    [url<]https://www.amazon.com/AMD-Phenom-1090T-Processor-HDT90ZFBGRBOX/dp/B003FVI2KQ[/url<] The CPU was great but the mobo chipset was super finicky I kept it for 1 year before upgrading to an intel again. Shortest window I've ever kept a CPU (normally I go for 3-5 years on a CPU) Guys if I'm running a sandy bridge consumer CPU on an SSD what kinda difference are we talking about?

      • NTMBK
      • 3 years ago

      Depends on what you’re running. In plenty of games it won’t be a whole lot faster than a Sandy Bridge quad, but if you run something that responds well to core count (like video editing) then it will be about twice as fast.

      • ptsant
      • 3 years ago

      AMD motherboards are usually cheaper. I believe the low-end ones are not as well-made. Whatever the brand, AMD or Intel, I think it’s worth buying at least a mid-end motherboard. For Ryzen, $150 should do the trick easily. Below $100 is probably dangerous territory.

    • ronch
    • 3 years ago

    Oh wow. I can’t believe it!! Here we go!!

    • derFunkenstein
    • 3 years ago

    Random thought: why would I pre-order a CPU? Are they going to be that rare at launch?

      • colinstu12
      • 3 years ago

      probably run out of stock for months, or there will be a long lead time. just like every other over-hyped gadget or piece of hardware is.

      • tipoo
      • 3 years ago

      Because it’s 2017 and there were pre-orders for a wifi router, hah.

      • K-L-Waster
      • 3 years ago

      Caters to the impulse buyers who want it now now NOW!

      • [TR]
      • 3 years ago

      I pre-ordered mine right after participating in the closed alpha. It was tough to choose where to pre-order because different outlets had different bonuses, but the handy spreadsheets people made helped a lot.

      A bit more seriously, though, opening pre-orders so close to launch can’t be more than a marketing stunt. Even if it could help them tailor manufacturing to demand, it’s too late for that, surely.
      If it’s out of stock for a while after launch, no amount of pre-ordering is going to help. And if it isn’t, no need to pre-order blind. We’ve waited a few years, what’s a few more weeks/months?

      • morphine
      • 3 years ago

      Well… it builds up even more hype, at the very least.

      • ImSpartacus
      • 3 years ago

      Anymore, preorders are for fanboys. That goes for preorders of all kinds for all companies.

      I seriously doubt that there will be stock limitations. It’s 2017, ffs.

      • Tirk
      • 3 years ago

      That could be asked of anything that is pre-ordered.

      The Nintendo switch can be pre-ordered, maybe it means Nvidia has a hard time producing enough last generation chips, not likely. Many many software gaming titles are sold with pre-orders through Steam and that has absolutely nothing to do with availability on release.

      • ultima_trev
      • 3 years ago

      If Polaris was a foretelling, then yes.

      • Stargazer
      • 3 years ago

      Obviously to get the pre-order DLC without having to wait for the POTY-edition.
      I’m trying to decide if I should get mine from NewEgg for the extra core, or from Amazon for the higher turbo boost.

        • Waco
        • 3 years ago

        I laughed, then cried inside, when I realized this is the BS we deal with when buying games.

        • meerkt
        • 3 years ago

        TigerDirect. For the RGB LED heat spreader.

    • odizzido
    • 3 years ago

    Samples? I think I might be most interested in a six core CPU so I hope you have one of those.

    • derFunkenstein
    • 3 years ago

    $330 would be more than I’ve ever spent on a CPU. Even in the “bad old days” of $500 Pentium 2s before the K7 dropped, I opted for a K6-2 for around $170. So I won’t be buying an eight-core CPU. Still looking forward to that 4C8T processor [url=https://techreport.com/news/31266/in-the-lab-intel-core-i3-7350k-cpu?post=1016827<]I said I'd buy[/url<].

      • RAGEPRO
      • 3 years ago

      Hey man, it wasn’t so bad. I overclocked my K6-III to over 500 MHz back when my buddies were all running 350-450Mhz Pentium IIs that didn’t overclock for beans thanks to the L2 cache.

      … of course, those chips still smeared my K6 in games. But it was damn fast in Windows!

        • derFunkenstein
        • 3 years ago

        That was pretty much my experience. Of course since I was a broke college student at the time, my video card was a Savage 4 and it didn’t run Quake III for crap anyway.

          • RAGEPRO
          • 3 years ago

          Heh, I rocked a [url=http://www.anandtech.com/show/643<]64MB GeForce 2 Ultra[/url<] for a hell of a long time in that period. I bought it the day it came out and went through like three CPUs on that card. I bought a [url=https://techreport.com/review/2608/hercules-3d-prophet-4500-graphics-card<]Hercules 3D Prophet II 4500[/url<] after I'd had the GeForce 2 Ultra for more than a year and ended up giving it to auxy and going back to my GF2 because it was actually way faster. Haha.

            • ptsant
            • 3 years ago

            The Ultra was a very expensive card at the time (also, I was very poor). I bought a 2 Ti that served me really well. Great product. I also had a great experience with the Geforce 3 and the 6600GT.

            • RAGEPRO
            • 3 years ago

            Yeah, it was expensive. That was back when my buddy and I had just started working and we were in the middle of a bitter fanboy rivalry. I ended up blowing a whole month’s pay on that card. I guess I got my money’s worth out of it, though.

            • DoomGuy64
            • 3 years ago

            I completely skipped the Geforce 1-2 cards, as there were no games that actually needed that level of hardware. Dx8 cards were worth buying somewhat, but pc graphics still didn’t advance that far until dx9 and ATI’s 9700.

            I remember buying a 9800 pro with an Athlon XP, while everyone else I knew was buying the Geforce FX and pentium4. I didn’t try to correct them, as I felt if they were really that stupid, they deserved it. lol, classic dustbuster.

            • ptsant
            • 3 years ago

            I also bought the 9800 Pro, my first ATI card. The difference vs the Geforce 3 was staggering. Great card.

        • f0d
        • 3 years ago

        the smart ones had mendocino celerons – i overclocked 2 of them on a abit bp6 to 550mhz
        those mendocinos were dirt cheap and overclocked like mad

        edit: $123 dirtcheap.! [url<]http://www.cpu-world.com/CPUs/Celeron/Intel-Celeron%20366%20-%20FV80524RX366128%20(FV524RX366%20128).html[/url<]

          • Bauxite
          • 3 years ago

          The icing on the cake was the celerons were faster at the same clocks for some programs back then, anything that mostly lived within the 128K of full speed cache versus the half-speed (or worse) 512K on a PII. Games seemed to fall into that category a lot 🙂

      • tipoo
      • 3 years ago

      Same, but these values also seem to go through the product line, i.e get a 4 core instead of a Pentium, a 6 core instead of an i3-i5, etc.

        • derFunkenstein
        • 3 years ago

        Yeah, if the leaked pricing continues to be correct down the line, you’ll have something like:

        4C4T Ryzen vs 2C4T i3
        4C8T Ryzen vs 4C4T i5
        6C8T Ryzen vs 4C8T i7

        And if you never use the onboard graphics, it all boils down to per-thread performance.

      • ozzuneoj
      • 3 years ago

      Yeah, I’ve been saying all along that the 4c8t models are going to be the best bang for the buck for most gamers. I’m sure having a couple more “real” cores will help in plenty of situations, but the thought of getting 8 threads and competitive performance on a (hopefully) up to date platform for $175 is hard to pass up right now. The option to drop in a chip with twice the cores later on for under $350 is nice too. For the same price with Intel you’re getting either a 2c4t i3 or a base model 4c4t i5, with an upper limit of a 4c8t i7 (for $350) as a future upgrade.

      I think the battle between the $175-$250 CPUs is going to be the most interesting one we’ve seen in a long time.

      • ptsant
      • 3 years ago

      Six core is probably the best perf/value. Thena gain, if you currently have a 2c/4t or 4c/4t, the new one is probably quite nice.

        • derFunkenstein
        • 3 years ago

        After a lot of swapping around, the PC that will wind up getting replaced has a Core i5-3470. And now that I think about it, I might give that one to my dad who is still rocking an Athlon II X2 640. Whatever ends up happening, it’ll be worthwhile. 😆

          • ptsant
          • 3 years ago

          1600X at $259 is probably a decent improvement over your current CPU. So is the Kaby Lake 7600K, for more gaming-oriented builds. We still don’t have benchmarks, of course, but many of the leaks have proven to be quite reliable until now.

      • jihadjoe
      • 3 years ago

      So why didn’t you go for the Celeron 300A @ $150?

        • derFunkenstein
        • 3 years ago

        The CPU I bought was compatible with my current motherboard/RAM (64MB of EDO and some Intel Socket 7 chipset)

        • Voldenuit
        • 3 years ago

        Celeron 333 @ 504 MHz for me. That thing was a beast in its day, trouncing most Pentium IIs.

          • jihadjoe
          • 3 years ago

          AFAIK the Pentium IIs topped out at 450, and weren’t really overclockable because of that off-chip cache so your cheap easily beats all of them. Clock for clock the Celerons with their smaller but faster on-die cache proved to be faster than an equivalently clocked Pentium II.

      • Geonerd
      • 3 years ago

      Yea, $300 is just a little too much to justify. I’m hoping that a mild price war and standard post-release price drops bring an overclockable 8/16 down around $200. At that point, my trusty Thuban may finally get ‘retired.’

        • derFunkenstein
        • 3 years ago

        I don’t think that’ll happen for a very long time. AMD needs to make money on these things for a while. You might see six-core CPUs dip down into that range (the low end of the line) when they eventually come out.

        I’m going to wait for March 2, regardless.

    • K-L-Waster
    • 3 years ago

    Best part of the article?

    [quote<]We're flying home with Ryzen review samples as of this writing.[/quote<]

      • tipoo
      • 3 years ago

      The buffalo has flown the coop, I repeat, the buffalo has flown the coop!

      • maxxcool
      • 3 years ago

      That is good to hear the AMD hates TR bs is done, but 6 days to NDA lift does not leave a lot of time for deep dives 🙁 ..

        • chuckula
        • 3 years ago

        At least TR is in the same boat as everybody else.

          • cygnus1
          • 3 years ago

          I still honestly wouldn’t be surprised if some other sites had samples prior to this event.

            • Meadows
            • 3 years ago

            Trust in Scott.

        • derFunkenstein
        • 3 years ago

        I’m sure based on Jeff’s Twittering that all the other CPUs that go into a Ryzen review are already tested and plotted and his spreadsheets are just waiting on Ryzen figures.

    • PBCrunch
    • 3 years ago

    Weird that the first slide compares the 1700X against Intel, but the second slide uses 1800X. To me, this signals that maybe 1800X isn’t much faster than 1700X when every thread is loaded down and the chip is limited by thermal or power delivery constraints. If this is true, the 1700X seems like the clear winner for content creators and other users of heavily-threaded applications based on its lower price.

    Putting a little math to the actual numbers in the graphs, the 1800X is only 4% faster than 1700X with 16 threads. That probably isn’t worth the extra money AMD is asking for the top-dog parts.

      • derFunkenstein
      • 3 years ago

      Your first clue should have been the ~5% gap in clock speeds.

      Your next clue should have been the absolute values. 1537 vs 1601. That makes the 1700X 96% of the 1800X for $100 less.

        • ptsant
        • 3 years ago

        Only reason to buy the halo product is if it overclocks better (for those interested) or you value having the best more than $100. Comparison between 1700 and 1700X probably depends on the added value of the XFR and the cooler. Will probably buy the 1700X if real benchmarks confirm its performance.

        • raddude9
        • 3 years ago

        Looks like the 1800x is going to be the traditional “halo” product with prices to match. But what about the plain-jane 1700 though, is the 0.4Ghz base clock and 0.1Ghz boost clock improvement with the addition of XFR going to be worth $70? A thorough review is needed….

          • ptsant
          • 3 years ago

          Forgot to mention the “Wraith” cooler that comes with the 1700X. It’s probably worth $10-20 over the standard cooler, if you don’t have a an aftermarket cooler that works with AM4.

          The XFR sounds very nice in theory. I suppose it’s probably more useful for games and single-threaded. Margin for XFR is probably slimmer for multi-threaded.

            • Gasaraki
            • 3 years ago

            “X” Processors won’t come with coolers.

            • ptsant
            • 3 years ago

            Of course they do. They come with the special “Wraith” cooler.

            • Bauxite
            • 3 years ago

            You’re confusing intel’s K with amd’s X, its ok they are both characters with 4 segments.

            • JustAnEngineer
            • 3 years ago

            “K” technically has just three segments, but I can understand the confusion.

            • chµck
            • 3 years ago

            they purportedly have SKUs with included heatsinks
            [url<]https://techreport.com/news/31470/rumor-ryzen-stock-coolers-and-retail-packaging-pictured[/url<]

          • Tirk
          • 3 years ago

          XFR does seem to be the lingering question on how the X chips value propositions compare to their non X counterparts. I’m very excited to see the testing results on how well XFR boosts the chip.

    • chuckula
    • 3 years ago

    Ladies and Gentlemen… The Buffalo has left the building!

      • derFunkenstein
      • 3 years ago

      Apparently didn’t even need to have the buffalo, considering Jeff could tell us he had review samples. All this for nothing, it kind of makes me sad.

        • chuckula
        • 3 years ago

        Darn you AMD for stymieing my Machiavellian NDA-buffalo schemes!

          • derFunkenstein
          • 3 years ago

          It was fun all the same.

      • morphine
      • 3 years ago

      The [url=https://en.wikipedia.org/wiki/Running_of_the_Bulls<]Running of the Bulls[/url<] has started.

        • prb123
        • 3 years ago

        [url<]http://www.cnn.com/videos/us/2017/02/22/bull-escapes-new-york-slaughterhouse.wcbs[/url<]

        • Neutronbeam
        • 3 years ago

        You know, I’ve ALWAYS thought of you as the chief running bull.

        EDIT: Took out , between You and Know.

          • morphine
          • 3 years ago

          MOO MOO MO********!

Pin It on Pinterest

Share This