Haswell overclocked: the Core i7-4770K at 4.7GHz

We’ve already seen what the Core i7-4770K can do at its default frequency. Now, it’s time to go beyond the chip’s stock speed.

K-series Haswell processors provide overclockers with three ways to crank the CPU frequency. Raising the Turbo multiplier is the easiest path to higher clock speeds. This variable targets the CPU cores exclusively, so changes won’t affect other system components. Multiplier tweaking has been the preferred overclocking method for the last few generations of Intel CPUs, and it remains so for Haswell.

The CPU frequency is the product of the multiplier and the base clock speed. Increasing the latter can also produce higher core speeds. However, the processor’s DMI and PCI Express interfaces also derive their frequencies from the base clock. Fiddling with the base frequency causes those interfaces to run out of spec, potentially compromising stability.

Haswell supports base clock control, but we’re told by multiple folks in the motherboard industry that the range of useful frequencies is similar to Ivy Bridge: an additional 5-10% at best. That said, Haswell improves upon Ivy by adding a base clock strap inherited from Intel’s ultra-high-end Sandy Bridge-E processor. This strap acts as a reduction gear for the DMI and PCI interfaces. It’s capable of dividing the base frequency by 1.25, 1.66, or 2.5, allowing that clock to be raised to 125, 166, and 250MHz without messing with the chipset and peripheral links.

Well, that’s the theory, anyway. Asus tells us it hasn’t found a single Haswell CPU capable of running a 250MHz base clock. The majority of chips will do 166MHz, it says, and 125MHz should be a lock for all of them. You should see a similar +/- 5-10% adjustment range at each strap setting.

With Z87 boards supporting CPU speeds up to 8GHz in effective 100MHz increments via multiplier boosting, there’s little need to touch the base clock or its associated strap. Only extreme overclockers looking to set benchmark records should worry about those settings.

Asus has tested hundreds of Haswell CPUs as part of its effort to profile the chip for auto-tuning algorithms. According to the motherboard maker, Intel’s new hotness has a little less overclocking headroom than Ivy Bridge does. Perhaps more importantly, Haswell apparently has more variance from chip to chip, especially in the voltages necessary to hit specific speeds.

Of the processors Asus has tested, 70% hit 4.5GHz, 30% reached 4.6GHz, and 20% made 4.7GHz. Only 10% were stable at 4.8GHz. Heat is reportedly the limiting factor, and Asus recommends using a dual-fan water cooler to prevent thermal throttling past about 4.5GHz or 1.25V. Going beyond 1.35V is apparently problematic even for high-end water coolers.

Since we have a high-end water cooler in-house, we decided to see how far it could take our Core i7-4770K. This is a different chip than the one Scott used for his CPU benchmarks, and the test configuration differed slightly from his, as well. The system was based on Asus’ Z87-PRO motherboard and a GeForce 680 GTX DirectCU II graphics card. Corsair provided the Force GT 120GB SSD, the AX850 power supply, and 16GB of Vengeance Pro memory.

Although the RAM is rated for operation at speeds up to 2400MHz, we confined ourselves to testing the limits of the CPU. Both Asus and Gigabyte tell us that higher memory speeds can limit CPU overclocking, so it may be worth exploring that dynamic in a separate article.

To keep the CPU cool, we used Corsair’s H80 water cooler. The radiator isn’t a double-wide affair, but it is sandwiched between dual fans. We also swapped the stock spinners for Corsair’s Air Series SP120 units.

We kept things simple for this round of tests and limited ourselves to manually tweaking settings via the motherboard firmware. A combination of AIDA64’s CPU stress test and the Unigine Nature benchmark was used to test stability.

To start, we let the motherboard select CPU voltages automatically as we raised the multiplier. We made it up to 4.2GHz without issue, and CPU-Z reported a CPU voltage of 1.2V at that speed. The system blue-screened at as soon as we started our stress test at 4.3GHz, though. Setting the CPU voltage to 1.25V kept the loaded system stable for a couple of minutes before the next BSOD, so we added more. In the end, 4.3GHz required 1.275V.

That voltage bump was enough to sustain our CPU up to 4.5GHz. However, blue screens at 4.6GHz forced us to nudge the CPU up to 1.3V. At 4.7GHz, the chip needed 1.35V, and its core temperatures spiked up to 84°C with regularity—even with the water cooler’s fans and pump going at full blast. Thermal throttling didn’t rear its head until we tried for 4.8GHz, though. Keeping BSODs at bay at that speed required 1.375V, and the additional voltage sent temperatures into the 90s. No amount of further tweaking produced a stable, throttle-free config at 4.8GHz.

At 4.7GHz, the system was stable enough to run a handful of benchmarks. The x264 test crashed on the first run but completed two subsequent three-loop sessions without issue.

(Although the screenshot above shows a CPU voltage of 1.376V, the CPU was set to 1.35V in the firmware. Asus’ AI Suite software agreed with CPU-Z’s reading.)

The numbers are pretty close to what one might expect. Pushing the Core i7-4770K to 4.7GHz increases performance by 19-26%, which is in-line with the rise in clock speeds. Our 4.7GHz overclock works out to increases of 21% and 27%, respectively, over the chip’s maximum Turbo frequencies for single- and all-core loads.

As is always the case with overclocking, your mileage may vary. That said, it’s worth noting that we hit 4.9GHz with a similarly early Ivy Bridge sample a year ago. That CPU also required 1.35V, but it got by with a dual-fan air cooler.

All indications point to overclocked Haswell processors requiring more aggressive cooling than their Ivy predecessors. The Core i7-4770K does have a higher TDP than the 3770K, but the associated heat is also spread over a larger die area. The 4770K’s TDP per area works out to 0.47W/mm², while the 3770K’s is 0.48W/mm². Haswell and Ivy seem to be on even footing in that regard. The die layouts follow the same basic blueprint, as well.

Haswell and Ivy Bridge also use a similar interface material between their dies and external heat spreaders. Intel used to employ a fluxless solder between those two pieces, but it switched to thermal paste with Ivy.

We don’t have a definitive explanation for Haswell’s apparent need for most robust cooling, but the chip’s integrated VRM may play a role. Voltage regulation was handled off-chip in Ivy Bridge, but Haswell brings it—and the associated heat—onboard the die. Integrated voltage regulation is a big part of Haswell’s appeal for mobile platforms. Unfortunately, it may also limit the processor’s overclocking potential on the desktop.

Comments closed
    • indeego
    • 6 years ago

    Kinda whack that TR posted a review of a CPU without power draw measurements or heat measurements. That seems…wrong?

    • RachelGat7
    • 6 years ago
    • bcr5784
    • 6 years ago

    Having just purchased an overclocked 3770k I was at first a bit miffed that Haswell was out so soon at a similar price, but having seen the tests I was relaxed about it.

    Clearly few people (not including me) have run enough tests on enough samples to give a definite view. However I find it revealing that the supplier I bought from – who doesn’t actually charge for overclocking, but does require you to buy a water cooler and appropriate case, only offers a 4.4
    Ghz overclock for a 4770k whereas he offered a 4.6 one on the 3770K AND requires a better water cooler as well. Since they don’t charge for overclocking I must assume they believe that the level of overclocking they provide can be achieved by pretty much any sample – and therefore represents a worst case. If that is typical then it looks like Ivy Bridge offers slightly more bang per buck than Haswell.

      • Airmantharp
      • 6 years ago

      ‘Out so soon’

      We knew Haswell was ‘coming out’ before we could buy Ivy Bridge.

    • kileysmith31
    • 7 years ago
    • tootercomputer
    • 7 years ago

    What’s up with newer intel chips and heat? For a long time I followed an ASRock forum for overclocking z68 and z77 mobos and was stunned at how hot the IBs got compared to SBs including my SB 2500K that was OCed on air stable and cool at 4.5GHZ. Now it sounds like Haswell is following suite with high heat. I know some folks took off the IB heat spreader and used better thermal paste and attained heat performance like SBs, obtaining really significant drops in temparatures.

    • LauraLasher43
    • 7 years ago
    • jamsbong
    • 7 years ago

    Interesting to see that the Haswell and Ivy bridge do not do well above 4.4 to 4.5GHz. Per CPU cycle, the Intel CPUs have become more and more efficient; so Haswell > Ivy > Sandy per clock. That is expected.

    However, despite the die shrinking from 32 to 22nm and 3D die tech, it is still not a high-ghz design.

    I think Haswell is worth upgrading for notebooks, tablets and other portable computer forms. Desktop users can skip this CPU. I’m still a proud owner of 2500K, it has been o/ced since day one. Has never skipped a beat.

    Speaking of notebooks, Haswell has shown to last 9.5 hours on battery! plus, with the decent on-die GPU, it will make all the difference in portable computing. I’m sure the next version of macbook will either have tons more battery life or 30 to 50% lighter…

    • jokinin
    • 7 years ago

    I wonder when there will be a CPU worth upgrading from my i5 3550. I feel like CPU developement is getting slower all the time.

    • DeadOfKnight
    • 7 years ago

    I wonder if turning off hyperthreading would allow you to push it a little more.

    • DavidC1
    • 7 years ago

    The possible reason for Ivy Bridge and Haswell using TIM rather than solder was discussed at Anandtech.

    There’s a Intel presentation where it shows there may be a long-term reliability issue with thermal stress on soldered processors. If I remember it correctly, they were talking 3-4 years or more.

    I’m pretty sure folks at Intel are well aware of what we are seeing right now, but doesn’t know a solution yet. There’s no artificial reason to cripple enthusiast-class processors like the K chips, especially when they acknowledged multiple times that enthusiast-class desktops like the K are the only other growing segment for the Desktop market(other being AIOs). It would be in their best interest to keep the K chip a good overclocker.

    That’s of course no problem for BGA or Notebook chips that doesn’t have a heatspreader in the first place.

      • Krogoth
      • 7 years ago

      I still think it simply comes down to cost. TIM is cheaper and does the job. The only people make a fuss over are extreme overclocker types.

      • mesyn191
      • 7 years ago

      Sounds like BS to me. I remember back in the K7 days Intel was saying exposed die CPU’s might be subject to early failure due to fan vibration being directly transmitted to the CPU die through the heat sink.

      Now exposed die CPU’s have issues with mounting a HSF on them but once they were on they ran just fine for far longer than their useful life.

      • chuckula
      • 7 years ago

      I’m not buying the whole thermal stress instability thing. Intel has been soldering on IHS caps since the Core 2 days at a minimum (proof link showing a blow-torch (!) being used to remove an IHS from a core 2: [url<]http://www.legitreviews.com/article/402/2/[/url<] ) Have there been any real reports of Core 2 failing due to thermal stress? As just one datapoint, my E8400 has been OC'd to 3.6GHz for over 5 years now for regular operation with daily suspend-resume cycles (hot-cold-hot cycles) without any instability. If they *really* *really* are too cheap to solder on the IHS, then I propose another solution: Sell the IHS as a separate component in K-series overclocking parts. It's still there, you say that it has to be placed on the CPU to maintain warranty, but there's no need to delid, and no globs of glue and low-rent thermal goop screwing up the thermal transfer from the CPU die.

    • End User
    • 7 years ago

    My 3770K @ 4.7 scores 1.93/9.10 in Cinebench R11.5 using a H100.

      • chuckula
      • 7 years ago

      Thanks for the datapoint!

    • Bensam123
    • 7 years ago

    Interesting… So it’s about the same performance increase you’d get out of IB taking into account the clock improvements with Haswell.

    I’m guessing having VRMs on board like Geoff said is responsible for the heat increase. I know VRMs dump a lot of heat, especially when OCing, and that’s when they’re sitting next to the processor… with their own heat spreaders… spread out over the motherboard.

    Also sorta interesting why Intel decided to continue on with thermal paste between the heat spreader and the processor instead of the fluxless solder. That was like the single biggest complaint from the OCing community and if I remember right Intel was just like ’cause’ when they were asked about it. Is there any reason to use paste when almost no one is going to delid their processor anyway (risking limb and wallet)?

    I like the nice little concise article snippet as OCing really deserves it’s own attention and you can really write a article about it. I can only imagine you guys also have FCAT and streaming articles in the lineup too. Definitely a good idea since articles have gotten so large.

    Undervolting? I’m not sure if you guys have another article planned for that, but do you plan on looking at it? I think the results would be just as interesting as OCing.

      • HisDivineOrder
      • 7 years ago

      Intel decided to stop using the fluxless solder because it’s cheaper to use the alternative. I think they probably saved a buck.

        • Bensam123
        • 7 years ago

        Sounds like a pretty expensive dollar to me. :l

          • deathBOB
          • 7 years ago

          What are you going to buy instead?

            • derFunkenstein
            • 7 years ago

            This is Bensam123 we’re talking about. he’s going to buy AMD without reading any reviews.

            • chuckula
            • 7 years ago

            Don’t accuse Bensam123 of not reading the reviews!

            He reads them, and then ignores them unless they say AMD wins! That’s what I call an open mind!

            • Bensam123
            • 7 years ago

            I bought a 3570k after reading the review and a 8350 quite awhile after reading the review on it too. I’ve owned a Intel system since I first built a computer and this is the first time I’ve touched AMD besides video cards.

            Not sure where you’re getting such a implication from besides people that seem to keep trying to dump it on me.

            • chuckula
            • 7 years ago

            Speaking of overclocking, since you ran out and bought the 3570k, what did you do in the way of overclocking to that chip before you decided it was a failure?

            I’m curious because with your latest “price-performance is the only metric that matters” kick, I want to know if you intentionally bought a more expensive part than was necessary from Intel just to “prove” that AMD is superior.

            • Bensam123
            • 7 years ago

            I’m not sure… Perhaps there are other alternatives out there… like ARM.

            • chuckula
            • 7 years ago

            Yeah Bensam123… the ARM side is calling to you… just imagine the price/performance* graph we could draw for a Cortex-A15 compared to the 3960X..

            * Where “performance” is: Can it run a web browser? Sorta? OK! It has enough performance!

    • ULYXX
    • 7 years ago

    Why do you guys even feed ABW?

    Thank you creating a separate and straight to the point mini article about the Overclocking of the 4770k. Although the performance increased about 15% from Sandy to Ivy to Haswell, the overclocking space seems to be decreasing each update. Unless people care about idling use, I wouldnt even feel bad Considering 3770k. Thats the opinion I have anyway.

      • derFunkenstein
      • 7 years ago

      Re: looking at Ivy Bridge, it also depends on what you want out of a motherboard. The Z87 also has 2 extra USB 3.0 ports and all of the SATA ports are 6gb/s ports. If you can give that up, and if you want to save a few bucks, then sure. Or if you have an existing 1155 setup and want a CPU upgrade, then sure – going from an i3 2100 on a Z68 board to a 3570K, for example, might make a ton of sense.

        • ULYXX
        • 7 years ago

        Ah, that’s very true. I have totally forgotten about the motherboard enthusiast crowd. On that note, there will be some very nice and plentiful options for every PC user this holiday season. 🙂

      • abw
      • 7 years ago

      Lolz……

      • smilingcrow
      • 7 years ago

      “Why do you guys even feed ABW?”

      It’s a bit like as a child seeing an insect with a wing or leg missing bumping around in circles. There can be a morbid fascination to poke it for some strange reason to watch its reaction. So in other words it’s a childish fascination with the mildly bizarre.

        • abw
        • 7 years ago

        I loled because he did exactly that by naming me in his post….

        That said i m open to any valuable tech debate but once
        what you say doesnt suit the fanboism the crowd will
        systematicaly displace the debate on a purely emotionnal
        round….

          • brute
          • 7 years ago

          my god

            • Chrispy_
            • 7 years ago

            “iz it can be troll tiem now?”

            • chuckula
            • 7 years ago

            Don’t worry brute, your trolls are about a million times more clever.

            • abw
            • 7 years ago

            All that is exagereted is insignificant…..

      • Chrispy_
      • 7 years ago

      Exactly the point I was trying to make, you just worded it better than me:

      [b<][i<]Per clock, Haswell is 7% faster than Ivy, which is 7% faster than Sandy. However, Sandy clocks 7% higher than Ivy, which clocks 7% higher than Haswell.[/i<][/b<] (7% being situationally very approximate) Original buyers of the K-series Sandy Bridge are probably feeling quite smug.

      • jihadjoe
      • 7 years ago

      I upmod trolls to keep them at zero, but only if there are no replies to the troll post.
      =)

    • Chrispy_
    • 7 years ago

    I never owned a 2700K myself to OC but from reports of colleagues and friends, as well as my own 2500K experience, most Sandy-bridge K-series chips would run at over 5GHz if pushed up to 1.4V (still just about possible to cool with air) and forums suggest that many risk-takers were able to hit 5.5GHz with water loops and more voltage.

    Power-efficiency be damned; Sandy Bridge is still the best overclocker IMO, and I’m wondering what speed an ‘old’ 2700K has to be pushed to to beat a 4.7Ghz 4770k 😉

      • Waco
      • 7 years ago

      As a 2600K and 2700K owner…BS. 5 GHz on air isn’t sustainable by any means unless by “air cooling” you mean a massive cooler with obnoxious fans and even then running in the 90s under load.

      Any most won’t run at over 5 GHz. Even with a good board many chips won’t get 4.8 GHz without some serious voltage (well over 1.4V).

      2500Ks clock much higher with less voltage though.

        • brute
        • 7 years ago

        depends on binning, too.

        i had a xeon e3110 (core 2 duo e8400) C0 stepping that did 4ghz at 1.18V, iirc.

        some chips are simply cherries!

        • Chrispy_
        • 7 years ago

        Just FYI, I’m not talking about “sustainable” overclocks, and neither is this article about sustainable overclocks – you’re crying ‘BS’ over stuff that I neither wrote nor implied:

        TR’s article clearly states that 4.7 was at the temperature limit of their H80 (running unbearably full-blast), and even then it crashed running the x264 test and was only stable enough for a handful of tests.

          • Waco
          • 7 years ago

          If it’s not a sustainable clock, what’s the point?

          A bit of tweaking would probably net TR a stable 4.7. Benching on the bleeding edge of stability is pointless and I thought it was pretty obvious by my post that those weren’t the kind of clocks I was talking about.

          No crying was done or implied. Posts like this are what give people unrealistic expectations and comparisons.

            • Chrispy_
            • 7 years ago

            I agree that there isn’t much point to an unsustainable clock;

            I just object to you calling my factually correct statement BS based on some nonexistant context that you’ve incorrectly implied yourself. I have seen many overclocked 2500K and a small number of 2600K/2700K overclocks at various LANs in the UK and abroad. Of those willing to push voltage up to 1.4V and above, all ran at 5GHz and certainly a lot of the i5’s were on air cooling. Whether you disagree with it or not does not change the fact that it has happened and I have seen it, nor does it change the relevance of this article which quite clearly states that 4.8Ghz was unstable and 4.7GHz was borderline unusable with their current cooling and settings.

            No misleading or malicious intent to deceive was ever on the cards, you just decided call BS based on your own expectations and slate the article in the process – that, to me, is a biased opinion, incorrectly based on a limited sample of data.

            • Waco
            • 7 years ago

            I guarantee I’ve seen more 2600/2700Ks overclocked than you have. I never said that some couldn’t do 5 GHz but the vast majority can’t even with good cooling and a lot of voltage. It’s certainly not “most” and while there are a few gems that will boot at 5.5 GHz almost none of the chips can actually complete any real work (read: complete benchmarks) at that frequency without some form of sub-ambient cooling…and even then they aren’t totally stable.

            So be offended all you like but what you posted was not fact.

            • Chrispy_
            • 7 years ago

            Since you’re happy to troll me I’m happy to feed you, but you seem to have basic reading comprehension or interpretation fail:

            [quote<]what you posted was not fact[/quote<] How is "from reports of colleagues and friends, as well as my own 2500K experience" not fact? I stated that (my generalisation) that "most Sandy-bridge K-series chips would run at over 5GHz if pushed up to 1.4V," and disclaimed it in the same sentence with the preceding phrase "from reports of colleagues and friends, as well as my own 2500K experience". At no point whatsoever, did I ever claim outright that "most Sandy-bridge K-series chips would run at over 5GHz if pushed up to 1.4V" as you seem to be incorrectly implying. Neither I nor you (unless you work at Intel in the chip-binning department) has seen enough chips to accurately back that up. It is therefore totally irrelevant whether you can or can't guarantee that you've seen more 2600/2700Ks overclocked than I have (and in a willy-waving kind of way, just a little bit childish).

            • Waco
            • 7 years ago

            I was simply pointing out that your generalization is not true. Take that how you will.

            Most SB K-series chips will [i<]not[/i<] do 5 GHz no matter the voltage. Most you've seen? No doubt. Not most though. Hence my original post...

            • Chrispy_
            • 7 years ago

            Heh, we’ll agree to differ I guess 😉

            You say my generalization is not true.
            I am saying that [b<]I did not make that generalization[/b<], only a much more specific statement (qualified with limits). Ambiguity can be the source for the mother of all off-topic arguments though and there is some scope for ambiguity in my original post so let's leave it at that.

        • travbrad
        • 7 years ago

        [quote<]2500Ks clock much higher with less voltage though.[/quote<] Even with my 2500K when I started approaching 5ghz it required a lot of extra voltage and therefore generated a lot more heat. I can run 4.5ghz (1.31vcore) at 60-65C (prime95) but to run @ 4.9ghz temps were 80C+. "Settling" for 4.5ghz and 20C lower temperatures seemed more reasonable to me. Mind you this was with a $20 HSF (CM Hyper 212+). Perhaps with water cooling temps would be a bit more manageable.

          • tootercomputer
          • 7 years ago

          Hmm, my 2500k experience almost exactly, even including the cooler. 4.5 was painless and on air max temps to mid 60s (whereas IB users were in the upper 80s and even the 90s) but to go beyond that it got ugly, i.e., high vcore and high temps. 4.5 was fine with me.

      • Stickmansam
      • 7 years ago

      Depends on the binning a lot. Friend has a 2500k that hits 4.8ghz on the stock cooler and stock volts but it hits about 90c+ on full load prime95. He never needed to try for anything higher

      • ColeLT1
      • 7 years ago

      Each chip is different.
      I the only sandybridge (2600k) I ever got to overclock, got to 4.5 (4.6 was prime stable but hot) (corsair A70)
      My ivy I got to 4.7 (4.8 was stable, but was hot) (dtek fusion 3×120 loop)
      Another Ivy I also got to 4.6 (4.7 was stable, but was hot) (CM 212 evo)
      I have a 4670k on order, will also be on a CM 212 evo.

      Note: when I say hot, I mean hitting 80c on a single core on coretemp is my limit with sandy and ivy under prime. The speeds I show above, after 24hrs on prime never go above 79c. Because these are computers for friends, since they don’t seem to be as keen to keep their computers dust free I dial it back a notch.

      • gmskking
      • 7 years ago

      5GHz on my 2600K just by upping the multiplier. Yea, Sandy can overclock.

    • slaimus
    • 7 years ago

    Will increasing the base clock to 125mhz work on non-k processors? If so, that is a good 25% performance increase that anyone can do.

      • wizpig64
      • 7 years ago

      I’m pretty sure you can fiddle with the multiplier a little bit too, just not as much as with a K part.

        • derFunkenstein
        • 7 years ago

        With SB and IB you could get +4 on all 4 Turbo settings. So for example with an i5 3470 you should be able to get 3.2 with 3.4/3.5/3.6 turbo (4, 3, and 2/1 cores) up to 3.8/3.9/4.0. Or something along those lines. So even if you’re not buying a K part, it makes sense to buy a Z-series chipset.

      • Pholostan
      • 7 years ago

      No.

      You need the clock strap, as several other important clocks are derived from the bclk. The PCIe clock for example, probably would be very unstable at 125 MHz. The clock straps though only works on K series processors. And on them you don’t really need it.

    • brucethemoose
    • 7 years ago

    [url<]http://www.chinadiy.com.cn/html/24/n-9024-18.html[/url<] As old/shady/badly translated as that article may be, I remember getting excited about the return of NB overclocking (pushing the memory controller on my phenom ii x4 actually boosted my FPS in some games). Since it wasn't mentioned, I don't suppose NB overclocking/tweaking has much of an effect on these processors, does it?

    • chuckula
    • 7 years ago

    [quote<]The system was based on Asus' Z87-PRO motherboard . . .[/quote<] I see you have my motherboard there! Hey Geoff, will it be part of any upcoming Z87 motherboard roundup reviews?

    • dragosmp
    • 7 years ago

    Thanks for relaying the info from ASUS about BCLK OCing. I don’t agree with:

    “With Z87 boards supporting CPU speeds up to 8GHz in effective 100MHz increments via multiplier boosting, there’s little need to touch the base clock or its associated strap. Only extreme overclockers looking to set benchmark records should worry about those settings.”

    If you don’t have a K you have to do all these things, without being extreme at all. A 2.5GHz CPU does 4.2GHz @166BCLK if stable.

    That said we may have been spoiled since the C2D days. Before the fastest CPU in a generation was close to what the architecture could do, OCing was limited. Only low end chips could OC 80% over nominal. With Haswell it seems those times are back and it’s good there is the BCLK option to push locked CPUs

      • Pholostan
      • 7 years ago

      Only the K parts support the clock straps AFAIK.

        • dragosmp
        • 7 years ago

        Here’s where I found it, hope they’re right:

        “It also means that even LGA1150 CPUs with locked multiplier can be overclocked, but only by 25% or 66% above the default frequency”
        [url<]http://www.xbitlabs.com/articles/cpu/display/core-i7-4770k_12.html#sect0[/url<]

          • chuckula
          • 7 years ago

          I’m hoping against hope that xbitlabs is right, but I’m not optimistic. Several reviews (TR included) have expressly stated that non-K parts only work with a BCLK of 100. If there is a way around that limitation, it could be very interesting though.

      • abw
      • 7 years ago

      Check Hardware.fr review for accurate info….

    • anotherengineer
    • 7 years ago

    Nice chart with power consumption??

    • Yeats
    • 7 years ago

    …”have a high-end water cooler in-house”…

    So why didn’t you use it?

      • axeman
      • 7 years ago

      Maybe it was the kind for drinking water…

      • DrCR
      • 7 years ago

      Same thought by me I’m afraid. I would love to know how Geoff could go if he had an actually high-end water cooler setup.*

      *This actually started me on a duckduckgo session to see what’s out there these days since I have not been into watercooling since the Cathar G5, Iwaki MD-20Z era. I found DangerDen apparently closed up shop a few months back.

      Watercooling was truly a different era for me, but I’d be interested to see if Haswell increases the demand for genuinely high-end watercooling setups.

      Geoff, maybe you should get yourself a couple of Thermochill 3x120mm, Iwaki MD-30Z, an a D-TEK Fuzion v2 or the like. That would be a good starter setup at the least. I recall 7/16″ tubing being a good compromise size, but I don’t know who may be using that nowadays.

    • Star Brood
    • 7 years ago

    I wonder how well the cooling would work on this guy:

    [url<]http://www.newegg.com/Product/Product.aspx?Item=N82E16835709011[/url<] According to reviews I've read, it's better than almost every water cooler.

      • derFunkenstein
      • 7 years ago

      LOOK AT THE SIZE OF THAT…

      [spoiler<]heatsink[/spoiler<].

        • Ditiris
        • 7 years ago

        Oblig: That’s what she said.

      • Dashak
      • 7 years ago

      This looks nearly identical to Noctua’s NH-D14, which also has one more heat pipe.

      • slowriot
      • 7 years ago

      I rather use one of the closed water cooling kits. Chances are the 120mm kits have greater compatibility than that monster. Not to mention a great deal less stress placed on the motherboard.

        • Airmantharp
        • 7 years ago

        And a great deal easier to work around, too. Not that you should be working around the CPU socket often.

        • Star Brood
        • 7 years ago

        Hmm you have a point. Reading off the reviews these big monsters have clearance issues left and right, forward and back.

      • Klimax
      • 7 years ago

      How about this one:
      [url<]http://www.zalman.com/eng/product/Product_Read.php?Idx=449[/url<] (Currently cooling my 3930k on 4.4GHz...)

    • ca_steve
    • 7 years ago

    Would love to see what overclock you can get with stock voltage as well as the associated stock/overclock temps. Heck, add undervolting to the list 🙂

    • derFunkenstein
    • 7 years ago

    So with your fancy enclosed-water setup, how bad were the CPU temps?

      • Arclight
      • 7 years ago

      [quote<]That voltage bump was enough to sustain our CPU up to 4.5GHz. However, blue screens at 4.6GHz forced us to nudge the CPU up to 1.3V. At 4.7GHz, the chip needed 1.35V, and its core temperatures spiked up to 84°C with regularity—even with the water cooler's fans and pump going at full blast. Thermal throttling didn't rear its head until we tried for 4.8GHz, though. Keeping BSODs at bay at that speed required 1.375V, and the additional voltage sent temperatures into the 90s. No amount of further tweaking produced a stable, throttle-free config at 4.8GHz.[/quote<]

        • derFunkenstein
        • 7 years ago

        I can’t read. My mistake.

          • heinsj24
          • 7 years ago

          I couldn’t read it either, I was looking for a nice little graphic with temps and power consumption from the wall.

            • Arclight
            • 7 years ago

            You guys have been spoiled. Back in my days we had to mine the data and make our own pie charts.

            • derFunkenstein
            • 7 years ago

            And that begs the question: just how old ARE you? I mean, I remember the Celery 300A days so I should remember this, too. :p

            • Arclight
            • 7 years ago

            Have you heard of the Enigma?

            • derFunkenstein
            • 7 years ago

            Is this a repeat of my parents telling me to never tell anyone my age over the phone? 😆

            • heinsj24
            • 7 years ago

            I blame tv. And public schools… but mostly tv.

            • derFunkenstein
            • 7 years ago

            So was I, but now that I think about it the graphics would have had one bar each. :p

        • heinsj24
        • 7 years ago

        84C @1.375V?

        Pffft, my Bulldozer can take 1.6V and doesn’t come anywhere close to that temperature – um, thanks to throttling.

    • Arclight
    • 7 years ago

    4,7 GHz at 1.35V is a far cry from the reports we got prior to launch. I must say that it is rather dissapointing.

      • Airmantharp
      • 7 years ago

      I am too- and that’s more than what other sites seemed to be getting.

      Maybe this CPU will need it’s lid popped and TIM upgraded even more than Ivy?

        • HisDivineOrder
        • 7 years ago

        Or you could just stick with/buy SB/IVB and get more with less hassle. Honestly, nothing about Haswell is advantageous to those unconcerned with lower power and/or improved IGP, so overclockers really don’t care about Haswell that much anyway.

        I suspect Intel might have been better off releasing this one as a mobile chip first and sitting on it six months before putting it on desktop. I can see why they’re thinking about not putting Broadwell out on LGA systems. It’ll likely be even worse about this.

          • Firestarter
          • 7 years ago

          haswell still has higher IPC, so it might be worth it even at lower clocks

          • chuckula
          • 7 years ago

          [quote<]Honestly, nothing about Haswell is advantageous[/quote<] Assume that Haswell has zero performance benefit over Ivy for the sake of argument. If you are in the market for a new system and you aren't just upgrading from a 3770K or something similar, then go with Haswell every time since the Z87 platform is a very nice upgrade over the Z77 platform including more USB 3 ports and more 6 GBit SATA ports.

      • Pholostan
      • 7 years ago

      Not really surprising. The truckload of engineering samples Intel gives away for free usually are hand picked chips that overclock very well. It is marketing pure and simple. Retail chips that you actually can buy in the stores always are much worse when it comes to overclocking.

      The old engineering samples of Sandy Bridge Xeons have been on ebay for a while, there must be many, many thousands around.

        • eloj
        • 7 years ago

        Yes, Sweclockers reported that their retail CPUs overclocked to around 4.4 while their engineering samples hit 4.6-4.7

    • chuckula
    • 7 years ago

    [quote<] Intel used to employ a fluxless solder between those two pieces, but it switched to thermal paste with Ivy.[/quote<] Le sigh.... That's your main reason for why Haswell can't hit 5GHz right there. Thanks for the good OC report anyway though. It looks like the 5GHz OC rumors were way too optimistic, although 4.5 - 4.7 GHz shouldn't be too crazy with a good closed-loop cooler. I'm not insane enough to try a delid, but I'd be very curious to see the results from somebody who is....

      • Firestarter
      • 7 years ago

      Over at Anandtech, someone figured out that the poor temperatures of Ivy Bridge weren’t so much due to poor TIM or thermal paste, but due to excessive clearance between the CPU die and the heatspreader. By removing the heatspreader and clearing out the glue, the whole CPU with heatspreader got about 0.06mm thinner in one test, which was the gap between the die and the heatspreader. After reintroducing that gap with a shim, they found that the temperatures were very comparable with the stock, unmodified CPU.

      Source: [url<]http://forums.anandtech.com/showpost.php?p=34053183&postcount=570[/url<]

        • chuckula
        • 7 years ago

        Another major advantage of solder: The gap between the IHS and the die is effectively zero since the solder bridges the gap.

          • Firestarter
          • 7 years ago

          *assuming that solder has the same thermal resistance as the IHS

            • chuckula
            • 7 years ago

            Ah… thermal impedance matching 🙂
            Even if they aren’t perfectly matched, I’m willing to bet that a decent solder bridge has a much closer thermal resistance to the metal in the IHS than cheap TIM goop or… much much worse… a thin air gap.

            • abw
            • 7 years ago

            Thermal resistance can be lower or higher , saying that it s close is valid
            when compairing two matters thermal conductance , you can then say
            that a given matter has a thermal resistance that is close of another
            matter s one , otherwise the thermal resistance provided by the solder
            must be as close to zero as possible for the die-IHS temperature difference
            being as close as zero as well….

            Theses details tells you if one know what he s talking about…

    • abw
    • 7 years ago

    Why not overclock the other CPUs on the charts , it s not
    like they dont allow this possibility but then the thing
    would look less good in respect of previous gen and
    competing products….

    Edit : when overclocked in its review the FX8350
    was compared to an even more overclocked 3770….

      • derFunkenstein
      • 7 years ago

      Why are you still here?

        • abw
        • 7 years ago

        There s other bars if the air in this one doesnt suit you….

          • chuckula
          • 7 years ago

          OK TR! Abw wants us to take a vote: Who should stay, derFunkenstein or abw?

            • derFunkenstein
            • 7 years ago

            Can I vote for none of the above? 😉

            • abw
            • 7 years ago

            I sense cowardice through your strawman……

            • SomeOtherGeek
            • 7 years ago

            I vote they spit in hands, shake and be TR blood brothers!

            • derFunkenstein
            • 7 years ago

            +1 but I’m a germophobe.

            • abw
            • 7 years ago

            Sieg heil…..

            • derFunkenstein
            • 7 years ago

            That’s germs, not Germans.

            [url=http://politicalhumor.about.com/library/images/blpic-moran.htm<]GET A BRAIN! MORANS[/url<]

            • abw
            • 7 years ago

            Why no Chucktrolla in the choice list.??…Indeed Coward…

            • chuckula
            • 7 years ago

            Pshaww… if you hadn’t been living in your AMDzone echo-chamber* for the last 8 years, you’d know that they voted me out YEARS ago. I’m kind of like the Agent Smith of the TR forums.

            * Speaking of which, I hear they miss you there. Please go back to AMDzone where you can be with your own kind.

            • abw
            • 7 years ago

            In short a miserable destiny , as for me i have the same name
            in all tech forums where i hang , find me if you wish at Hardware.fr
            and Anandtech……

            • chuckula
            • 7 years ago

            Oh so your French! That explains why you are an AMD fanboy!
            You see, AMD makes its chips at GloFo fabs in Germany, so the instant you saw one, your innate surrender instinct took over!

            Vichy Abw! Vive le surrender!

            • abw
            • 7 years ago

            I also talk french , wich is not the same thing as being french ,
            as well as a couple other languages…

            On the same note i generaly find the intel club fan quite
            illiterate and far less open than AMD dudes , that s a given
            that surface in any linguistic area….

            • maxxcool
            • 7 years ago

            hahaha, now it makes sense, nonsense posts, all wind, hides on the web to strike terror autonomously, then hides when wrong then cashes the amd checks 🙂

            • abw
            • 7 years ago

            You are living on childish myths , 9 years old psyché at first glance….

            • maxxcool
            • 7 years ago

            ohhh now i get it, your snakeoil!!!!

            • abw
            • 7 years ago

            Edit : 6 year old intellect at most , i gave you 50% more IPC
            than you actualy deserved…

            • maxxcool
            • 7 years ago

            Where’u been Snake? it has been boring around here without your poor posts and terribly easy to goad amd paid love ?

            • abw
            • 7 years ago

            I guess you are TRs village most idiotic troll..?….

      • puppetworx
      • 7 years ago

      The 8350 overclocked managed roughly a 10% performance boost when it was reviewed.

      [url<]https://techreport.com/review/23750/amd-fx-8350-processor-reviewed/13[/url<] Edit: I just noticed the 8350 was tested on air, not water, and the heatsink fans were pointing at the back of the graphics card, WTF? The FX-8350 would probably do a little better on a level playing field then it did in this test, but it's a rough guide.

      • nanoflower
      • 7 years ago

      For a simple reason, time. There’s a lot going on (as they’ve said) and only so much time in a day so they can’t do everything that they would want due to those time constraints. In this case you can find all the overclocking results you want if you go back and look at past reviews. They just aren’t included here for the likely reason of not enough time.

        • abw
        • 7 years ago

        Time ?..When TR review was released one day after other sites.?.

        It doesnt take much time to make a few copy/past
        with a few previous ocking results.

          • Star Brood
          • 7 years ago

          Wow you actually want a review site to copy paste from other review sites? I don’t even know why you visit TR. Might as well go back and crawl up into the cavities of the likes of Tom’s Hardware.

            • abw
            • 7 years ago

            From their own database , like they did for the Cinebench preview…

        • Bensam123
        • 7 years ago

        Not sure time fits this. With the original article they had to get it out the door right away, but with these sub articles they have quite a bit more time to polish them as there isn’t something directly comparable being released on the same day.

      • albundy
      • 7 years ago

      have you ever seen a review with that happening? i havent. secondly, haswell isnt just about the cpu. the new z87 boards have started including much nicer features than even the top of the line 990fx chipsets.

        • abw
        • 7 years ago

        I guess that you did read one line before jumping to answer….

        • heinsj24
        • 7 years ago

        It’s not about the [s<]Doctor[/s<] cpu anymore; it's all about the [s<]companions[/s<] motherboards.

      • travbrad
      • 7 years ago

      I agree some overclocked IB/SB test would be nice but I suspect the results would line up very similarly to the stock clocked results, since they all have fairly similar overclocking capabilities. An overclocked Haswell is going to be about as much faster than an overclocked Ivy Bridge as a stock Haswell is faster than a stock Ivy Bridge.

      • NeelyCam
      • 7 years ago

      The first quasi-reasonable argument I’ve heard from you, and delivered without the usual poison..

      But the thumbs have ruled that you’ve overstayed your welcome

Pin It on Pinterest

Share This