Leaked slides tease Ivy Bridge performance improvements

According to the latest round of leaked Intel roadmaps, Ivy Bridge is due in the second quarter of next year. We’ve already discussed the selection of models being prepared to replace the existing Sandy Bridge lineup of Core processors. Now, Xbit Labs has published purportedly leaked slides detailing how the new silicon stacks up against its predecessor.

A couple of official-looking Intel slides compare the performance of the Core i7-2600 to that of the upcoming i7-3770, which has the same core count, base clock speed, and cache size. The i7-3770 is purportedly 7-25% faster across a range of tests that stress the CPU, although it’s fair to note that the chip’s 3.9GHz Turbo peak is 100MHz faster than the i7-2600’s maximum speed.

Perhaps more interesting are the performance results from tests that tap the integrated graphics. "ArcSoft Media Expresso," which is perhaps what happens when the folks making the presentation slides confuse ArcSoft’s MediaConverter software with CyberLink’s MediaEspresso product, is said to be 56% faster on Ivy Bridge. Both of those applications support Intel’s QuickSync transcoding tech. Ivy’s integrated graphics component scores three times higher than Sandy’s in 3DMark Vantage, as well. Keep in mind that the Core i7-2600 has the slower HD 2000 version of Sandy Bridge’s integrated graphics, while the i7-3770 is expected to have the fastest version of the new IGP, dubbed the HD 4000.

What’s especially remarkable about these results is that Ivy Bridge’s CPU cores should be able to deliver performance improvements while at the same time drawing less power than their predecessors. The i7-2600 has a 95W TDP, while the i7-3770 is supposed to consume less than 77W. That lower TDP suggests Ivy Bridge will have more relaxed cooling requirements than its predecessor, and the chip could prove to be a more potent overclocker as a result.

Comments closed
    • Abdulahad
    • 8 years ago

    SPONSORS OF TOMORROW….
    THIEVES OF TODAY……

    • ronch
    • 8 years ago

    How about a leaked slide detailing Piledriver’s performance improvements? Hello, DonanimHaber?

      • xcore
      • 8 years ago

      huh, the amd gossip generator :)))
      they will manage to get something, but I wonder, Have AMD one?! :)))

    • Chrispy_
    • 8 years ago

    ROFL, the “2.99x” HD Graphics 2000 is in 3DMark Vantage, where you get bonus points just for bringing a DX11 part. Intel is only claiming “up to 30%” and 30% more of a lousy IGP doensn’t bring anything new into playable territory.

    • Abdulahad
    • 8 years ago

    The problem is that every year Intel is making huge leaps in graphics (their weakest link) while AMD is stagnating in CPU performance.
    As it is Intel is already over charging for their products and I cannot understand how we could get hyped up talking about them in a such a Godly manner….

    If AMD is out of the picture, fanatics (whom I respect a lot) would be the first to bite their fingers( while hiding) for ridiculously high prices and we can already see this.

    Whenever huge conglomerates (like Intel) make big advances, the first thing they think of is how to nail the wealth of consumers by any means and crush the competition dead…..

    Of course better products have its place, but if Intel is giving us better CPU tomorrow with the intention of over-pricing today, we have a big issue…..all of us, Fanatics,Tech savvy, Tech-layman, Tech-fearing person ect..

    Intel Inside…. Think Twice of the Price!!!

      • NeelyCam
      • 8 years ago

      That’s the funny thing about the attitude towards competition…

      “Competition is good!..But don’t compete so hard that you’ll win. Win is bad. If it looks like you’re winning, we gotta do something to stop that. We gotta make your competitor win!”

      Most of you agree, I’m sure, that these monopolies need to be stopped to protect the consumers. But that’s essentially socialism, is it not? A pure free market would let all companies go crazy and do anything they can to beat the competition.

      So, Abdulahad: do you think there should be regulation in place to set the prices of CPUs? Or regulation forcing Intel to let other companies use their manufacturing facilities?

        • khands
        • 8 years ago

        We haven’t lived in a purely capitalistic market during almost the entirety of the US’s history. Monopolies get broken up only when considered “necessary” for the betterment of the consumer, and it’s still in Intel’s best interest to keep AMD afloat, at least till ARM gets competitive on the desktop (which won’t be for at least 10 years). Unless they want monopoly regulators down their throats, which we all most assuredly don’t want because the splitting process would hurt everyone at least short term.

      • indeego
      • 8 years ago

      My CPU expenses are less than .5% of my total yearly budget, and even less for work. I work in a field that depends on them, and a CPU gives me thousands of ways to get other things done. I donate to people that can’t afford the latest processors after I’m done with them. How is this taking advantage of me or society?

    • Arclight
    • 8 years ago

    I’ll be shocked, shocked i tell you if IVy bridge IGP is equal or better than Llano, let alone Trinity. You guys give Intel slides wayyyy too much credit.

      • khands
      • 8 years ago

      It’s 3DMark too, which means nothing. That being said, they’ve tripled the IGP cores and made a bunch of other tweaks, I’m guessing Intel’s “round” won’t be as round as AMD’s, but performance wise we could very well see them catching up. Close to Llano (given best case scenario’s) isn’t an impossibility. But part of me feels that it’s because the Llano IGP isn’t being fully utilized (hamstrung by the CPU part), the exact opposite of Intel’s IGP problem.

      • xcore
      • 8 years ago

      the slides expert at the all times is AMD! their green ppt presentations are unbeatable. see the “magnificent” “dust-pan” (some call it bulldozer) pictures!
      but where is the GREAT cpu, huh? 🙂

    • FuturePastNow
    • 8 years ago

    I’m glad the focus is on decreasing power consumption rather than increasing clock speeds. Of course, Intel can afford to slack a bit on the performance front given their lack of competition in that regard.

    • kamikaziechameleon
    • 8 years ago

    Heres to hoping Ivy bridge E will have all 8 cores unlocked.

      • OneArmedScissor
      • 8 years ago

      Don’t be too disappointed when they announce it’s a 10 core chip with 1/5 the cores and cache disabled. :p

        • DeadOfKnight
        • 8 years ago

        That actually sounds rather likely.

    • UberGerbil
    • 8 years ago

    Did I miss the thread where people went off on the inevitable rant about Intel’s numbering system for Ivy Bridge? Or have we as a people finally moved past that?

      • OneArmedScissor
      • 8 years ago

      I think people have just accepted at this point that you need some sort of decoder ring. AMD’s scheme is almost as convoluted. I just look at the lists on Wikipedia, as I always had to do for graphics cards, anyways.

      • Thesprunk
      • 8 years ago

      Actually I think people realized its not as convoluted as it sounds.Really, it’s pretty straight forward, it just doesn’t follow the way that we think about processors (in terms of cores and ghz, but rather market segment and speed intervals). It loosely follows the bigger is better/pricier/faster philosophy.

      In case you’re still confused
      Core ix = Market Segment. 3 is entry, 5 is mainstream 7 is high end/enthusiast
      Thousands digit = Generation
      Hundreds digit = Entry/Enthusiast end variant for market segment, but essentially, higher the number, the higher “class” and faster the chip.
      Tens digit = Multiplier speed increment (that’s right, GHz no longer are representative of speed these days)
      Ones digit (if not 0) and any Letters that follow designate special versions.

      I’ll admit that something like

      xyz GHz
      x=Gen
      y=Threads
      z=Core
      Giving you 142 2.6GH

      or something would be a bit more “transparent” for those that care about specs, but those numbers aren’t as neat and marketable to the masses who don’t know enough about specs to know whats better. In fact that would quickly get convoluted as well if you started to factor in what graphics it has or if its low power, unlocked etc etc. It’s a give and take, but complaints about “It makes no sense” are clearly because you didn’t think it through too hard.

        • NeelyCam
        • 8 years ago

        Yeah, except that SB-E and IB are now in the same “generation” according to your decoding scheme.

        • ronch
        • 8 years ago

        I think It’s actually very similar to how AMD gives model numbers to its Radeons.

    • omegajvn1
    • 8 years ago

    Hey Geoff, the i7-2600 actually has the 3000 integrated graphics vs the 2000 of which u probably typo-ed lol

      • chuckula
      • 8 years ago

      No, Geoff is right. The 2600 has 2000 series graphics. The 2600[i<][b<]K[/b<][/i<] has 3000 series graphics... the two chips are not the same.

    • Jigar
    • 8 years ago

    This news is 6 days late guys…

      • Peldor
      • 8 years ago

      It was in the shortbread a few days ago too.

      • Rakhmaninov3
      • 8 years ago

      *shuffles through the stuff on his desk, trying to find Jigar something better to do with his time*

    • Duck
    • 8 years ago

    Who cares about quicksync when you cant even use it? It’s a broken feature unless you are not even runny any kind of discrete graphics card.

      • cmircea
      • 8 years ago

      Lucid Virtu, NVIDIA Optimus, AMD (Dynamic) Switchable Graphics.

        • Duck
        • 8 years ago

        …all require a new motherboard?

          • NeelyCam
          • 8 years ago

          Not necessarily a new one – just the right one. AFAIK, H67 and Z68 support Lucid Virtu.

            • MadManOriginal
            • 8 years ago

            I think a certain aquatic avian is pissed he did a crazy-early adoption and got a P-series board for Sandy Bridge, even before reviews were out making it clear that QuickSync won’t work with those boards.

          • jensend
          • 8 years ago

          It’s not as though Sandy Bridge and Ivy Bridge fit into some older socket. So when you buy a processor with QuickSync for the first time you’re already buying a new motherboard anyway. If you really wanted both QuickSync and a discrete graphics card but bought a mobo that didn’t support that, you have only yourself to blame.

          People and their first-world problems. Sheesh.

      • bcronce
      • 8 years ago

      IB’s IGP is DX11, which means it can be used for any direct compute work loads.

      • jabro
      • 8 years ago

      Can’t you use QuickSync with integrated graphics using the Z68?

      I also recall that QuickSync and discrete graphics can be used together with the next P-67 series equivalent chipset when Ivy Bridge comes out.

    • chuckula
    • 8 years ago

    Good news for Intel: It looks like IB’s graphics will be competitive with mobile versions of Llano while having good power usage.

    Good news for AMD: It looks like Trinity’s graphics will still be comfortably ahead of IB, especially on the desktop.

    Unknown: Will AMD get a version of Trinity into a < 20 watt power envelope and have graphics that still beat IB by a good margin…. ???

      • NeelyCam
      • 8 years ago

      Yep – that pretty much sums it up.

      • sschaem
      • 8 years ago

      I dont share that view on the GPU.
      AMD claimed Trinity to be 33% faster then llano.
      And Intel benchmark show IB to be 26% faster then llano.

      The difference is minimal. ~5% is not going to give AMD the Edge.
      I would buy llano over SB in a laptop, but noway would I consider Trinity over IB.

      IB is going to be >30% more power efficient then Trinity and be able to offer much high peak performance.
      Trinity is going to be a valid choice for those cheap OEM building those 7 pounds big plastic laptop with shitty screens,
      but otherwise Trinity is DOA.
      BTW, If the PC grows stall, its possible that Intel production capacity will be able to supply everyone and AMD is out.

      AMD in 2011 took a HUGE step backward while Intel R&D soared… This will become even more evident in 2013.

      As AMD stated they are out of the race and will focus on other markets.

        • NeelyCam
        • 8 years ago

        [quote<]IB is going to be >30% more power efficient then Trinity and be able to offer much high peak performance. Trinity is going to be a valid choice for those cheap OEM building those 7 pounds big plastic laptop with **** screens, but otherwise Trinity is DOA.[/quote<] I think Trinity will be a fine chip if AMD gets BD bugs and issues figured out, and GloFo gets the process stable and fast. If the price is right, Trinity will sell well for a certain segment of the market. I don't see Trinity being competitive in the Ultrabook market segment (it cannot possibly compete with IB in CPU performance at <20W TDP), but beyond that it should be OK.

          • tay
          • 8 years ago

          I like your posts in general, but those seem like two big if’s especially the latter with Global Foundries failing at everything they do.

            • NeelyCam
            • 8 years ago

            They are ifs, and I’ve made my opinion about gate-first pretty clear, but these people (AMD/GloFo engineers) are smart and can figure stuff out when there is enough of a panic. Don’t forget that AMD/GloFo 45nm was pretty good.. they’ve had successes. With 32nm they probably put their gate-eggs in the wrong basket, but they’ll figure out how to fix the situation to some degree.

            So, I won’t rule out Trinity just yet.. AMD has managed to survive while pricing high-cost items cheap, and they’ll price Trinity just were it needs to be.

        • khands
        • 8 years ago

        Llano was pretty much 50% faster than what Intel had going IGP wise overall with SB, so this catches up for the most part (and their image quality is likely still shitty, but I’ll reserve judgement till proper reviews come out). That would make Trinity about 30% faster still than Ivy.

          • maroon1
          • 8 years ago

          According to anandtech, ivy bridge GPU is going to have improved anisotropic filtering quality
          [url<]http://www.anandtech.com/show/4830/intels-ivy-bridge-architecture-exposed/5[/url<]

          • sschaem
          • 8 years ago

          Intel show a 300% improvement in graphic performance.
          So if IB is 3x faster then SB and SB was half the speed of llano… you do the math.

          From the leaked Intel benchmark and AMD PR, both should deliver about equal graphic performance.
          Its even possible that IB comes out on top in term of graphics performance.

          Unless AMD intentionally sandbagged its 33% faster GPU statement about trinity over llano, or Intel leaked fake slides,
          you can bank on Trinity not having a faster GPU then IB.

          Tricky: I just noticed Intel slide show the HD2000, not HD3000…
          I dont have those numbers… what happen when you scale HD2000 score by 3x?
          A8-3850 class ?

            • chuckula
            • 8 years ago

            It’s 300% faster than a Sandy Bridge *2000* series core according to Intel’s slides. That means it should do pretty well against Llano, maybe winning some benchmarks and losing a few others within a few % here and there. Trinity, however, is going to be ahead of IB in graphics, especially for the 125 watt chips that AMD is going to be pushing out. Whether a *low power* version of Trinity for ultrabooks can be competitive with IB is still an open question.

            • khands
            • 8 years ago

            Actually, looking back at techreport’s reviews Llano was more like 2x the 3000 (which was at best around 50% better than the 2000), on both desktop and mobile.

            [url<]https://techreport.com/articles.x/21099/14[/url<] [url<]https://techreport.com/articles.x/21208/15[/url<] Regardless, I think we'll just have to wait till we get real reviews to see one way or the other.

    • Atradeimos
    • 8 years ago

    Meh. Krogoth not impressed.

      • bittermann
      • 8 years ago

      He should be…

Pin It on Pinterest

Share This