Back to Article

  • vastac13 - Friday, July 11, 2014 - link

    Just when I'm about to go to bed... Good thing I checked 1 more time :D Happy reading folks! Reply
  • Iketh - Friday, July 11, 2014 - link

    I just hope Intel has another massive breakthrough in performance soon because rendering a 4K mp4 is gonna SUCK

    Don't need AMD for this case, the industry will drive it this time...
  • ddriver - Friday, July 11, 2014 - link

    Looks like desktop CPU performance is hitting a brick wall, the last 4 generations are barely incremental in performance. Could very well be the product of AMD ultimately failing to compete in the high end. I don't complain, this way I don't feel the urge to upgrade my 3770k. Reply
  • Antronman - Thursday, July 17, 2014 - link

    IBM says silicon will end up limiting CPU capability, and are investigating alternatives. Reply
  • jhawkinstx - Tuesday, March 03, 2015 - link

    Yes, when I heard that in the news I went out and invested in companies with alternative tech (Gallium-Arsinide based processors) which did not have the limitations of silicon.......that was in the late 1990s....seems the end of silicon, like peak oil, has been preached for quite some time. For now, I will just wait patiently with my 4 year old Intel boxes. When something more than incremental improvement arrives I will buy. Reply
  • Chaser - Monday, December 29, 2014 - link

    Nor do I with my 2600K. It would be throwing money away. I am a single monitor 2440P higher end gamer that does some office work as well. I have no CPU shortfalls. Reply
  • jhawkinstx - Tuesday, March 03, 2015 - link

    I agree, I have 4-intel 2600K boxes and which run my business jobs at roughly the same rate as the Xeon based box I bought a year earlier, but with the 2600K costing 1/6 the price of the xeon. Current benchmarks show me nothing worth spending the money on. If anything I would add another 2600K box for roughly $200/processor. Reply
  • JlHADJOE - Saturday, March 21, 2015 - link

    On the contrary, CPU performance has been steadily increasing, but Intel has been choosing to spend those IPC gains on lowering the TDP instead of increasing outright performance. Reply
  • wrkingclass_hero - Sunday, July 13, 2014 - link

    It's funny, I'm actually rendering a 4K mp4 right now! My 4.6 GHz 3930k looks like it's going to take a grand total of 21 hours to render it... and it's a 3 minute clip...
    But there are a lot of effects on it (stabilization, layers of videos, etc.) It also doesn't help that it is a two pass.
  • ddriver - Monday, July 14, 2014 - link

    What do you expect, throwing such a workload on a measly single cpu - such tasks are performed by rendering farms with thousands of quad socket machines. Reply
  • Braincruser - Sunday, July 13, 2014 - link

    Rendering should be left to GPU cores in shaders. They scale much much better than cpu. Reply
  • Mark-Benney - Thursday, August 14, 2014 - link

    I wrote out a the Manderoot program on a Acorn Electron 32k, Back in the very early days. Took 48hrs to complete up on books with fan placed under it. Lol bet you were not even born when i first wrote program in Dos/Machine code/Pascal/BBC Basic. And still in nappys when I was overclocking a Intel Celron from 233mhz to a stable 24/7 367mhz Reply
  • CrystalBay - Monday, July 14, 2014 - link

    Thanks Dr. Ian I love my Intel 4790K @ 4.8 Ghz , I also love Asus Z97 Deluxe . This isa the simplest way to OC in my 25 years of building PC's ... Screw it being AVX stable capitol BS never will be used instruction . Any modern chip fails at it anyway ... Go Devils ,go AMD Reply
  • pcjoeyd - Wednesday, February 18, 2015 - link

    crystalBay,,, not ANY modern chip fails...just your unstable cpu setup. why run at 4.8ghz and run your cpu super hot? bump it down to 4.6-4.7ghz (you won't notice any difference), run your cpu much cooler and be 100% stable. Reply
  • superjim - Tuesday, July 15, 2014 - link

    Intel hasn't made a good OCing chip since Sandy Bridge. Devil's Canyon just reinforces how good SB was. Nearly every i5 and i7 chip could hit 4.4 without issue with most at 4.6+ on a good air cooler. Raise your hand if you're still on SB only because there is nothing better... Reply
  • AlucardX - Tuesday, July 15, 2014 - link

    same here. I'm on 2500k at 4.5 GHz. I kept reading this article to try to determine if it was time for a rebuild, and i guess the answer is no. this would be more appealing if the i7 had true 8 core CPU's in it Reply
  • jloutz - Thursday, July 24, 2014 - link

    Completely agree with above posts. I'm at 4.5 GHz on closed loop water on a 2500k and this thing has been rock solid for four years (by far the longest upgrade cycle since I started DIY builds in 1998. I believe I got a Microcenter CPU/ASUS MB bundle for ~ $300 (wow!) I've added SSD's and a new graphic card, but still itching for a CPU/MB chipset upgrade. Let's see what Skylake brings to the party?? Reply
  • pcjoeyd - Thursday, February 19, 2015 - link

    yep, no need to get anything greater than sandybridge as it stands now. I have my 2600k i7 clocked at 4.4ghz super stable and cool. My daughter's pc I have her 2500k running at 4.7ghz. I'm using Noctua NH-U14S coolers in both systems. Rock solid, runs cool and super QUIET!! Reply
  • MrSpadge - Saturday, July 26, 2014 - link

    Switched from Sandy to Ivy for the better power consumption under moderate OC (4.1 GHz @ 1.03 V). Running 24/7 load on it, so even stock voltage (1.18 V) is far too expensive. Reply
  • bhima - Sunday, August 03, 2014 - link

    Still rocking my i5-2500K at 4.2 on air. I could go higher, but I'm pretty conservative when it comes to OC. I have no need to upgrade as games aren't bottlenecked by my processor, and using the Adobe suite is still a trivial task for Sandy Bridge. Reply
  • FlushedBubblyJock - Thursday, November 20, 2014 - link

    Raising 2500K hand acquired near release.
    4.8 oc from the motherboards auto setting no sweat.
    Saw that and never tried anything higher as I figured the chip is golden to me.
    It's been an extreme pleasure to never worry about the system being bogged down to a page turning crawl, which prior to SB was always a possibility, for years, on anything and everything.
  • Chaser - Monday, December 29, 2014 - link

    I'm at 4.4 with my 2600K just with simple MB automated O/Cing. /raise Reply
  • DanNeely - Friday, July 11, 2014 - link

    What did you use as cooling for the tests? A stock cooler or something aftermarket? Reply
  • Stuka87 - Friday, July 11, 2014 - link

    I was wondering the same thing. I did not see it mentioned, unless I just missed it. Reply
  • MooseMuffin - Friday, July 11, 2014 - link

    He mentions a Corsair H80i in the overclocking section. Reply
  • DanNeely - Friday, July 11, 2014 - link

    Thanks, I was expecting to see it on the overclocking page itself instead of in the first page's intro/summary information. Reply
  • Ian Cutress - Friday, July 11, 2014 - link

    Corsair H80i in a push-pull configuration. Reply
  • bill5 - Friday, July 11, 2014 - link

    kind of surprising how well the 8350 holds up in some of these tests, given it's age and price. granted, in some tests it's nearly doubled, but in others it's surprisingly competitive.

    if amd ever gets anything going at all they'll give intel a run. intel performance has been stagnant for ages.
  • DrMrLordX - Friday, July 11, 2014 - link

    Forget the 8350. Go A10-7850K! Anywhere OpenCL shows up, it kicks butt at its price point. You can get one off the Egg for around $180 now. And that's just OpenCL. HSA where are you?

    To me, the results where the 7850K wins are clear indications of what could come to pass once OpenCL and other GPU compute options become more mainstream. Anyone writing software would have to be insane not to look at the advantages offered by Kaveri (and Carrizo, and Skylake).
  • sr1030nx - Saturday, July 12, 2014 - link

    I'm glad to see opencl and hsa coming along, there's a good chance they'll make AMD relevant in higher end computing. Reply
  • GeorgeDBartlett - Saturday, July 12, 2014 - link

    hi Reply
  • Spirall - Friday, July 11, 2014 - link

    Thanks for the article! I undestand Intel has no chance to go back with the improvements (including base frequencies) with Broadwell Ks. Would like to see memory speeds (including BLCK change impact on them). These are what Haswell Ks should have been from beggining. Reply
  • CrystalBay - Friday, July 11, 2014 - link

    Thanks Dr. Ian it was worth the wait ! I had already bought a 4790k ,but your breakdown hypothesis of the TIM has put this article above others... Reply
  • Kevin G - Friday, July 11, 2014 - link

    This is the Haswell Intel should have introduced a year ago for the enthusiast. There is a bit more tangible benefit for owners of Sandy Bridge and Ivy Bridge chips to jump to Haswell now. Still the main reason for the upgrades in my view is the improved chipsets with Z97 being rather nice over the older Z68 and Z77.

    As for coming Haswell-E chips, it'll remain rather niche like the current socket 2011 chips. You either have an explicit need for more cores, more PCI-e lanes or more than 32 GB of memory.

    The delays of Broadwell make me wonder just how long it'll be on the market before SkyLake arrives. At first Intel was targeting Broadwell as a pure mobile part but eventually recanted. Then the delays hit with desktop parts looking seemingly appearing in 2015. If SkyLake is a late 2015 part as envisioned, then Broadwell will have a short life span and ultimately worth passing to jump directly to SkyLake.
  • FlushedBubblyJock - Thursday, November 20, 2014 - link

    I can see the nm performance wall hitting... for general use gamers... some fancy software specialty is going to be the near future selling points... looks like it to me anyway. Reply
  • Muyoso - Friday, July 11, 2014 - link

    I wish there was an overclocked 3770k @ 4.5ghz in the charts so that I could compare my current setup to what is new. I have a feeling that it would be fairly competitive still. Reply
  • beginner99 - Friday, July 11, 2014 - link

    Why don't us test multiplayer games? Seriously single-player games on 1080p will be GPU limited basically always with the current top end intel chips. and these 5 fps differences are pretty much irrelevant. What would be interesting is BF4 64 player maps for example and with mantle or dx. These benchmarks here are useless for choosing a gaming cpu or even worse they can be misleading in some cases. Reply
  • ZeDestructor - Friday, July 11, 2014 - link

    It's harder to make multiplayer completely reproducible.

    Then again, now that I think of it.. if you have a LAN server, it should be possible t do that...

  • Ian Cutress - Friday, July 11, 2014 - link

    Benchmarks have to be consistent. Just applying FRAPs to a few online games of BF4 for each CPU has no guarantee of consistency and it is an apples to oranges comparison. One match could be heavy in explosions for example, or if I decide to camp out as a sniper. There is also local variability based on the server you connect to and the number of individuals logged in.

    Until an intensive online multiplayer game has the option to record a time demo and emulates network delay accurately, the only way we can test gaming is via single player. BF2 came pretty close back in the day, but no modern titles that I know of have this feature and remain consistent.
  • et20 - Friday, July 11, 2014 - link

    Starcraft 2 games can be quite intense and the replay functionality is solid.
    I doubt it provides network delay emulation. Why do you want that anyway?
  • ZeDestructor - Friday, July 11, 2014 - link

    Couldn't you use something with a dedicated server combined with kb/mouse emulation on a bunch of extra PCs running at 640x480 nonsense kind of thing?

    It's potentially more work, especially the potential synchronisation and timing issues, but it should be doable to within 10ms of latency (on my LAN I see ping roundtrips in the sub-ms range), should it not?
  • Ryan Smith - Friday, July 11, 2014 - link

    In theory yes. But in practice most first-person games spawn you at a random point, which makes any kind of input track playback ineffective.

    The games where such a thing would work would tend to be games that already have better benchmark capabilities anyhow, such as racers and RTSes.
  • ZeDestructor - Friday, July 11, 2014 - link

    Couldn't you get a modified DS from a dev? Reply
  • FlushedBubblyJock - Thursday, November 20, 2014 - link

    They just don't want to do it.
    They live for their claimed accurate scientific method.

    It will take a genius with guts or a brute with some money, then we can see the results we never get that we all want somewhere else.
    Let's face it, it could be done and there would be some +/- low end percentage variability, so frames could be rounded to whole numbers removing the tenths ( which are outside the current bench errors and variability ).

    It just isn't going to happen here, but someone should definitely do it, and we'd all love going there.

    Real life scenarios are just too scary for the cold and removed and protected elite. Politics.

    We all play our online games and know what frames we can count on given our vid cards and systems and current clocks and we all have our favorite maps and servers... etc.

    Remember all these websites run the highest end overlcocked cpu and boards they possibly can, and that also is deceptive for most readers.
    They run SSD's now with clean installs and used to go with defragged and rebooted but now methods are equivalent so nothing extra is running best case scenario stuff...

    Yes another type of review site is needed, but then again the general idea is given for what info is available here, so downsize accordingly is the answer.
  • wallysb01 - Sunday, July 13, 2014 - link

    If this is something that’s really needed for a real test, why not just do it 50 times per processor and do some stats. Reply
  • doggghouse - Friday, July 11, 2014 - link

    Out of curiosity, what do people consider to be safe voltages for Haswell? I recently had to replace my 4770K with a 4790K because the chip started to BSOD even when not overclocked... I don't know if I helped speed its demise after having tested it at 1.4V several times, and I think I settled on 1.3V 4.4GHz daily use (it was a mediocre chip apparently).

    I apparently lucked out on my 4790K because it is running stable at 4.6GHz with only 1.25V, and 4.7GHz at 1.3125V. I was thinking about testing its upper limits for fun--try for the mythical 5GHz--but I don't want to accidentally burn out an otherwise great chip. I very briefly ran it at 5GHz with 1.424V and HT disabled, it was stable enough to run a few benchmarks. But if I play around in the 1.4V range, am I potentially going to wreck it?
  • TheinsanegamerN - Saturday, July 12, 2014 - link

    Typically, for 22nm intel (sandy bridge and haswell) the typically regarded "safe 24/7 voltage" is only 1.3 volt on air, and 1.4 volt on water. for a very short period, higher than 1.4 can be used if you have VERY good cooling, but you may damage the vhip even with ln2 cooling at anything above 1.4v.
    Also, it sucks you had to do 1.3v to get to 4.4Ghz....I hit 4.2Ghz at only 1.075v. apparently got a REALLY good chip somehow, although i heard that the costa rica chips, of which mine is one, always did overclock better.
  • doggghouse - Monday, July 14, 2014 - link

    Interesting... so is it the temps, and not the voltage directly, that eventually kills the chip? If so, would running lots of synthetic benchmarks that brings temps into the 90-100C range shorten its lifespan? I have an AIO water cooler, the Kraken X60, which can keep normal temps cool, but anything above 1.25V will still hit 100C on the latest Prime95 Small FFTs test. Reply
  • FlushedBubblyJock - Thursday, November 20, 2014 - link

    Voltage can kill the cpu. You won't necessarily get the temp reading it can happen so fast.
    Secondarily, a higher voltage is more volatile with higher temps, so the combination can also cause electromigration blowouts.
  • Dustin Sklavos - Friday, July 11, 2014 - link

    "A contact at Corsair."


    Here are the complete results for my Devil's Canyon and Pentium chips:
    4790K #1: 4.7GHz @ 1.275V
    4790K #2: 4.7GHz @ 1.31V
    4690K #1: 4.7GHz @ 1.375V
    4690K #2: 4.8GHz @ 1.375V
    G3258 #1: 4.9GHz @ 1.4V
    G3258 #2: 4.7GHz @ 1.375V
  • maximumGPU - Friday, July 11, 2014 - link

    Thanks for the info! Reply
  • xeizo - Friday, July 11, 2014 - link

    I quit overclocking this generation, I have a SB 2600K that does 5GHz+ stable for 24/7 operation if needed and it sits fine in my gaming rig. But, Haswell is interesting exactly for that 24/7 operation thing, I like to have my home server always on and it doubles as a Linux workstation. The lowest end i5 aka 4570S is perfect for that kind of use, very conservative power draw means great opportunities for a silent rig and at the same time snappy enough performance for almost all tasks. I did some bclk on mine so that it turbos to 3.83GHz and it has fast 2133MHz memory. And it is virtually silent and very cool(max 50C at 4x100%). It replaced a Q9450 and was quite a big upgrade in all aspects, so Haswell has it's points other than maximum performance at any cost.
    My gaming rig waits sceptically for Haswell-E, looks more to be Maxwell frankly speaking ...
  • jihe - Saturday, July 12, 2014 - link

    I quit overclocking when one has to pay a premium to do it. Long live FSB and BCLK overclocks. Reply
  • ZeDestructor - Monday, July 14, 2014 - link

    Everything LGA2011 i-series is unlocked ;)

    Ofc, if you're going there, you have other reasons for doing so, not just overclocking.
  • PICman - Friday, July 11, 2014 - link

    Would you please elaborate on the relationship between outer shell electrons and thermal conductivity? Is more better?

    Incidentally, polymers have poor thermal conductivity. Nearly all of the thermal flux is due to the filler (composition, concentration, aspect ratio).
  • ZeDestructor - Friday, July 11, 2014 - link

    From what (admittedly little) I know about chemistry, outer-shell electrons don't matter much in thermal conductivity. Not here at least, since they are not free (and thus electron motion heat transfer (can't remember the proper name)isn't possible as in pure metals and alloys). All they would do is change the colour of the TIM as well as other properties of the associated salt (oxides are a specific type of salt).

    If you check back around Sunday evening/Monday morning, I may update this thread: I have a DnD game with some chemist friends who may be able to answer that question, and well, is a mildly interesting question.
  • PICman - Friday, July 11, 2014 - link

    ZeDestructor, thanks, but I was being sarcastic. Thermal conductivity is much more complex than simply counting valence electrons. Note that a diamond (carbon has 4 valence electrons) is the best solid non-conductive material. Liquids and conductive solids (like silver or copper) are a different story. Reply
  • Ian Cutress - Friday, July 11, 2014 - link

    The outer shell d-electrons was more a comment towards color. A number of d10 oxides are used as pigments for paint due to their strong white color. If the ceramic matrix suspension had been something more vibrant, we might be able to pin it down easier. But you don't see any Vanadium based TIMs anywhere :D

    Thermal conductivity usually benefits from rigidity of the material. Hence why diamond is good and a sponge is bad. If you want the best thermal conductivity between the CPU and heatsink, then something like Coollaboratory Liquid Ultra, which uses a binding metal (Gallium, which eats Aluminium - check Youtube for videos) to reform between the two metal surfaces and bind them. This binds the heatspreader to the heatsink (or the die to the heatspreader) without any chance of removal without severe force.
  • DiDaDaDi - Friday, July 11, 2014 - link

    Does anyone knows why 4670k/4770k had TSX/VT-d removed? Of course apart from the obvious reason Reply
  • DanNeely - Friday, July 11, 2014 - link

    They consider it an Enterprise feature and don't want enterprise customers buying cheap consumer CPUs and Overclocking them for more performance when they could force them to buy Xeons for ten times as much. Reply
  • ZeDestructor - Friday, July 11, 2014 - link

    Non-overclockable LGA-115x CPUs have VT-d (assuming your chipset + mobo allows it) as does the LGA-2011 platform as a whole nowadays, when intel added it to the platform on unlocked CPUs in a later revision (i7-3930K C2 and newer, available on all i7-3820) while keeping overclocking.

    That's reason number 3 for me getting Haswell-E later this year, since I do a fair bit of VM-related work on my desktop.
  • DiDaDaDi - Saturday, July 12, 2014 - link

    We don't care what you want. I was just wondering why Intel hasn't pushed an update to the MB BIOS to "unlock" VT-d/TSX Reply
  • ZeDestructor - Saturday, July 12, 2014 - link

    No idea... Most likely it's written in the CPU itself, but then again, we can get microcode updates to the CPU as well... so really, who knows?

    Also, iirc the 4670K/4770K never had VT-d to begin with, like all previous LGA115x K-series cores...
  • wintermute000 - Saturday, July 12, 2014 - link

    I thought 4790k has vt-d? (unlike 4770k). Stupid intel segmentation Reply
  • Roland00Address - Friday, July 11, 2014 - link

    I would buy the 4790k and just keep it at stock. Why get the K model? It is clocked 400 mhz higher and some people say that is worth $40 more dollars which is a 13% increase of cpu price, and even less when you figure in the cost of all the other computer components.

    I understand these chips are not marketed to me, but why would you deal with all the hassle of overclocking for less than a 10% difference in most tests? We are talking like 28 to 40 degrees celsius hotter cpu even though you are using an $100 cpu cooler, 40 watts more power draw and thus a hotter room, a louder computer, and worse a chance of data corruption and all the potential hassles overclocking brings in some instances.

    I understand if we were getting 30 to 50% performance increases like we used too, and still do with the new unlocked pentium g3258, but for less than 10% in most tests...why bother?

    (If you answer why bother? With its fun! That is okay and I understand. But to me overclocking is just work. I rather spend the 2 hours overclocking playing a game or watching a movie, and you will spend at least 2 hours tweaking and testing, etc.)
  • ZeDestructor - Saturday, July 12, 2014 - link

    Overclocking is easy if all you want id a 10-15% boost: took me all of 15minutes worth of reboots on my 3570K to settle on 4.4GHz capped and +0.03V. Sub-80°C with my cheap-ass $30 cooler, until such time I build an awesome Novec-cooled (Novec is a line of non-conductive cooling fluids by 3M) liquid-cooling loop. Reply
  • Communism - Friday, July 11, 2014 - link

    Intel Math Library Optimized Linpack or no balls :D

    But seriously. My Ivy Bridge can do Intel Match Library Optimized Linpack @ 4.6ghz without throttling.

    Haswell and/or Haswell refresh are not upgrades for me if they cannot do the same.

    Stability must mean no errors of any sort ever. Period. This includes any and all throttling.

    Hopefully Haswell-E delivers.
  • willis936 - Friday, July 11, 2014 - link

    I don't think it's possible to have no errors ever. Reply
  • JimmiG - Saturday, July 12, 2014 - link

    The Ivy Bridge only supports AVX. It's AVX2 that absolutely kills Haswell overclocks. We're talking a 10C increase over AVX (v1) easily. Reply
  • wurizen - Friday, July 11, 2014 - link

    not tempted by the 4790k at all. i have a i7-3770k oc'd to 4.1. i thought i was safe at 4.2 until sony movie studio crashed and i realized it was the oc so i downclocked it to 4.1. i have it offset too to -.065. not really an OC maestro but i am happy with it and i think i have one of those underperforming cpu batches. i also have a push-pull tpc 812 cpu cooler on there which is a high-end air cooler from cooler master. whatev.

    any news from AMD upgrading the FX series? and a new chipset? why is amd just focusing on the F2 series and A10 APU's?
  • prime2515103 - Friday, July 11, 2014 - link

    Dolphin Benchmark: "...and general reports tended to suggest that Haswell provided a significant boost to emulator performance."

    Boy, I'd say... My i5 2500k at 4.5Ghz took 9.87 minutes. Beat out by an i3, sheesh!
  • wurizen - Friday, July 11, 2014 - link

    oh, also want to add that haswell-e seems to be what people are waiting for since it will offer 6-core as base and a new 8-core cpu as well as supporting ddr4. but, i'm also waiting for AMD's answer to desktop cpu. they already have an 8-core processor. i feel like if they update the chipsets, add pci 3.0, maybe even ddr4 support as an option, improve the FX series 8-core processor to be equivalent or near haswell-e or even ivy-e 6-core,--it would really give intel a run for the innovation/awesome cpu maker spot--but i know i'm dreaming. Reply
  • Gigaplex - Saturday, July 12, 2014 - link

    AMDs chip lies between a true 8 core, and a 4 core with hyperthreading. It has enough hardware for 8 full integer cores but behaves more like hyperthreading with the 4 shared floating point modules. Reply
  • wurizen - Friday, July 11, 2014 - link

    oh, does anyone know if AMD has better TIM on their FX series cpu's?and/or have the IHS soldered? Reply
  • hahmed330 - Friday, July 11, 2014 - link

    What I want in a discrete graphics or a APU is High quality HEVC 1080p encoding with the speed of quick sync as I have to rely on X264 encoding. Otherwise buying a new CPU has no point for me. Sandy Bridge still provides great performance at 5.5Ghz good enough for gaming and hopefully at the emergence of DX12 CPU limited scenarios will be in very few circumstances. Reply
  • dj_aris - Friday, July 11, 2014 - link

    Somehow it feels like Intel are really bored and struggle to present something interesting. I mean, One would expect Broadwell to be around the corner or at least some Iris Pro love in a socketed chip just to compete with AMD in the only area they are not dominating right now. At least in Atom chips things are more interesting lately. Reply
  • TrevorH - Friday, July 11, 2014 - link

    You know, they make marker pens that have nibs that are less than 0.5cm across ;-) How do you ever read what you scribble on cpus? Reply
  • wurizen - Friday, July 11, 2014 - link

    yeah. it's called a sharpie. lol! Reply
  • AnnonymousCoward - Friday, July 11, 2014 - link

    He only had a Sir Marks A Lot Reply
  • ncrubyguy - Wednesday, July 16, 2014 - link

    someone needs to send him and Ultra-Fine Point Sharpie
  • leoblaze9 - Friday, July 11, 2014 - link

    WOW! what happened in the particle simulation test? that difference is STAGGERING!!
    just one of the prime examples of the benefits of using both cpu and gpu together as a cohesive whole rather than two discrete entities. I am genuinely amazed at AMD's engineering! kudos deserved that one!
  • MamiyaOtaru - Saturday, July 12, 2014 - link

    possible reason they don't use solder anymore:

    "At higher thermal cycles the cracks have propagated across the entire center region of the die and significantly degrade thermal performance.This failure mechanism is predominant on small die form factors. The center degradation mode is not observed, as the die size is increased to the large and medium sizes"

    Smaller die sizes are more likely to see cracks in the solder spread to the middle, significantly harming thermal performance. Ivy Bridge / Haswell with their process shrink have smaller die sizes, perhaps making solder unworkable as a TIM
  • ZeDestructor - Monday, July 14, 2014 - link

    Interesting.... Reply
  • Galatian - Saturday, July 12, 2014 - link

    I agree: 4,6 GHz seems to be the sweet spot. Mine uses 1,25V core and a 0,175 V offset uncore to reach Ringbus parity. I needed 1,294V for 4,7 and 1,345 for 4,8. Temperatures were off the hook though and this is with a custom-liquid loop with 2 480mm radiators (one monsta and one 25mm) as well as Liquid Ultra. Stability tested with Prime95 24h and IBT 20 runs on maximum (kudos to IDK in the forum for is stability guide).

    It is an improvement as it seems to be easier to get a rock stable CPU, but temperatures still are a problem. The thing is: I don't believe a better solution would ultimately change anything. I mean when you clearly see that the sweet spot is 4,6 GHz with 1,25V one would assume that 5 GHz probably needs something like 1,5V and that's just dangerously high IMHO.

    Also what I realized is that DC needs a lot less Input Voltage. Stock was at 1,8V, but even with my overclock I could dial it all the way down to 1,5V, which is amazing. Ian have you checked how low you can go? (I can't believe I just said that...). Also how exactly can Prime95 ruin a CPU? I still consider it gold standard, as it would show problems within minutes when other stress tests would happily run for hours. I know it is a hassle to let it run for 24h but nothing gets close to rock stable setting then Prime95.
  • pierrot - Saturday, July 12, 2014 - link

    Now we just need the OC'd Anniversary Pentium benchmarks added in! Reply
  • andrew-1983 - Saturday, July 12, 2014 - link

    The image is blurry, but the batch number of the 4790k looks
    like L336... which if I'm decoding it right means 2013 36th week,
    around 2013 September. Does that matter that you are testing
    a 10 month old CPU for overclocking?
  • Galatian - Saturday, July 12, 2014 - link

    That number stand for the time the silicon was started to be manufactures or something alike...(read the has been mentioned several times now)...there are a lot more step until a final chip is put together, so it really is not a 10 month old processor. Reply
  • andrew-1983 - Sunday, July 13, 2014 - link

    Ok I get that, but when the retail cpu we buy has an L418C164 that's still
    an 8-month difference, whatever the numbers actually mean, and there is
    a new microarchitecture or die shrink in every year..
    What I'm trying to say is overclockability should be tested with
    silicon similar to what people will buy after reading the review.
  • bludragon - Saturday, July 12, 2014 - link

    My chip need 1.237v to be prime95 stable at 4x4.4GHz. This is the stock vid for aturbo on 1 or 2 cores. Using an NH-D14 it would throttle on cores 2 and 3 (so hitting 105+deg?) withing a few minutes. Delidded it maxes at 91 degrees, so at least 15 degrees better and I used the Noctua TIM. I tried 4.6 but the voltage nneds to be 1.27 or higher as it throttled again there. Reply
  • ratbert1 - Saturday, July 12, 2014 - link

    With an i5-2500k I still do not see the need to upgrade. I have run at 4Ghz with speedstep and voltage on auto for 3 years. My synthetics will be slower, but using an SSD my everyday usage is fine. Just ran Bioshock Infinite benchmark and got 115 fps average and 41 fps min. @1080p on GTX770 running stock speed(card is OC'd version though). Hopefully Broadwell early next year. Reply
  • Timur Born - Saturday, July 12, 2014 - link

    My 4770K needs 1.25 volts just to hit stable 4.2 GHz on all cores. So a 4790K looks tempting. On the other hand I can get 4.4 GHz Turbo out of my 4770K at single/dual core operation, so in essence it is the same thing, just using higher voltages than a stock 4790K. Worth paying lots of money for that? Not so sure. Reply
  • doggghouse - Monday, July 14, 2014 - link

    If you already have a 4770K, then buying a 4790K is basically just buying another ticket in the silicon lottery, unless you have a need for the extra enabled features (ex. VT-d). One good thing about 4790K is it's a higher binned chip, guaranteed to run at least 44-44-43-42 with default voltage (1.1v?). However, beyond that it may or may not have any overclocking headroom. Reply
  • Timur Born - Tuesday, July 15, 2014 - link

    I agree. My 4770K needs 1.1v stock, hardly reaches those 44-44-43-42 at 1.3v (jury is still out on stability) and doesn't go to 43-43-43-43 at *any* voltage. So any lottery can only be better than what I already got. ;)

    I'm haggling with a friend who might want my CPU. Else I might just wait for Broadwell.
  • wrayj - Saturday, July 12, 2014 - link

    I work in media production, and when dealing with 6k raw footage that must be decoded and debayered, our editing performance scales tremendously with more cores. The idea of an upcoming overclockable 8-core CPU with the fresh x99 platform is much more exciting to me than Devil's Canyon. Even at $1k, we're talking about a huge amount of value compared to xeon platforms. Reply
  • JanieEMontgomery - Saturday, July 12, 2014 - link

    hyyy Reply
  • mfenn - Saturday, July 12, 2014 - link

    Ian, you can't have your cake and eat it to when it comes to proselytizing your overclocking viewpoints. For example, in the very same page of this review, you make the following statements:

    " A fair number of regular end-users ... can be concerned about overclocking..."

    "To put the concept of 'overclocking death' into perspective: ... Out of the 250+ CPUs I own, I have only ever had one CPU fail."

    "Personally, I shudder when a user suggests a system is not stable unless it passes ‘72hr Large FFT Prime95’, because I have seen users irreparably damage their CPUs with it. "

    On the one hand you say that you've only broken one CPU by inputting wildly out of range values, the purpose of that statement is to tell people that they're probably not going to break their machine. On the other hand you're telling users that running a piece of application software could damage their (overclocked) CPUs.

    So is overclocking a safe an fun hobby or isn't it? You need to make up your mind before trying to convince people of your position.
  • xeizo - Sunday, July 13, 2014 - link

    Well he didn't, or did he? He just shared his reality which isn't black or white. It's grey, like the reality is. Of course there is danger for damaging the hardware when overclocking, what Ian means is that when sensibly doing so there is very little risk but he also warns about ridiculous stability testing(not "application software") which makes the chance of damaging the hardware much more likely. When in doubt, don't OC, and when doing it for fun the risk is part of the fun. Really. Reply
  • DeathMetroll - Sunday, July 13, 2014 - link

    I'm currently using a I7 920@3.7Ghz, is it worth to go for a I5 4690K ? I'm just a gamer (BF4 and other 3D games) Reply
  • ZeDestructor - Monday, July 14, 2014 - link

    Wait till Skylake would be my opinion. You'll get an even better boost, and move up to DDR4 and other good stuff. Reply
  • DeathMetroll - Monday, July 14, 2014 - link

    Thanks for the answer, that's what I was thinkng about. But I don't know if I'm able to wait till then !!! ;) Reply
  • FlushedBubblyJock - Thursday, November 20, 2014 - link

    What vid card are you using is my question.
    If you've got a 970 or a 290x already then it might be worth it to jump early.
  • HollyDOL - Monday, July 14, 2014 - link

    Umm, please fix the pseudo-code on page 2:

    try {
    } catch {
    CPU.CoreVoltage += 0.025;
    continue; // exit do while loop

    Increasing multiplier only makes sense when you successfully passed the exception block ;-)
  • Djibrille - Friday, July 25, 2014 - link

    I was thinking the same thing.
    That continue statement does not exit the do loop it just returns to the beginning, so the comment is not accurate either. ;)
  • HollyDOL - Friday, July 25, 2014 - link

    ye true, actually incrementing that voltage float (in case it is a floating point variable) is wrong too since it leads to FP errors, so it should be like:

    CPU.CoreVoltage = baseVoltage + (++voltageMultiplicator * 0.025);
    continue; // skip to next loop iteration

  • Djibrille - Saturday, July 26, 2014 - link

    Maybe that's taking it too far, it's just pseudo-code albeit pretty close to real c/c++/java code.
    Logical errors I think he needs to fix though, otherwise including the code is pointless.
  • rms141 - Tuesday, July 15, 2014 - link

    "It would seem to offer higher limits for overclocking, although your mileage may vary. For Ivy Bridge users, the rewards are even greater. However, gamers at 1920x1080 resolutions gain very little."

    If this is true, then why are there no Ivy Bridge CPUs included in the benchmarks? Not even the venerable i5 3570k? It's very strange to include a single Sandy Bridge CPU, but no Ivy Bridge CPUs.
  • FlushedBubblyJock - Thursday, November 20, 2014 - link

    The 3570K is faster on the 1155 SB socket, so like it's a given it's covered and in better shape, but yeah, it might be too close to the new breed... lol... leave it out so everything looks like real progress. Reply
  • Hrel - Tuesday, July 15, 2014 - link

    Seems like Intel is slowing down their upgrade cycle. These are basically stepping revisions (A1, A2, A3 etc...) being marketed as new CPU's. I guess their tick/tock is going bi-annual. Cause Haswell came out last year, unless you wanna be technical, then it came out in 2012. Reply
  • johnpombrio - Sunday, July 27, 2014 - link

    I might just remove my Haswell that is o/c to 4.2GHz and stick in a Devil's Canyon. Why? To get rid of the noisy water cooler fans. The higher I o/c, the noisier the fans and 4.4 is max anyways without occasionally bumping into the 90 C headbump. I also would not bother to o/c the 4790K chip and put the CPU back to an air cooler. This would be my FIRST computer build that I have NOT o/c in the past 25 years! Reply
  • namec - Monday, July 28, 2014 - link

    You hit the nail squarely on the head, I think, JP. Devil's Canyon is the first stock-clocked 4.0 ghz CPU, and it has a 400 mhz on-demand overclock (Turbo). This all comes with a factory warranty!

    I was waiting for Broadwell, but aside from 30% better efficiency (which I covet), it may have little to offer. At least until the 5790K comes out.

    The days of the 40% overclock are past. Sure the G3258 (factory at 3.2ghz) can oc to 4.7 ghz (which btw is about the same stable top clock for 4790K), but Intel could have released the G3258 at 4.2 ghz stock. Then some would complain that it does not oc well.
  • SeanJ76 - Tuesday, August 05, 2014 - link

    Hitting 4.7ghz@1.290Vcore with my 4690k and Msi Z97 Gaming 5 deck. I'm satisfied with that. My 2500k@5.0ghz is easily beaten by a mild overclocked 4690k(4.6ghz is when I started passing my 2500k scores in Firestrike and Valley Bench). Reply
  • SeanJ76 - Wednesday, August 06, 2014 - link

    ^^That's 4.7ghz on air, and I didn't see temps above 52C!! I went with the Hyper Evo/Msi Z97 Gaming 5 deck. Reply
  • Mark-Benney - Thursday, August 14, 2014 - link

    I have intel i5 4670k very nice chip good upgrade from i7 920, But in defence of the i7 920. It was well overclockable as long as you keep temp down came close to running it stable at double its base speed. The new i5 I have is not really stable much past 3.9mhz with overclock ram 16gb 2400mhz. But to this day no game including B4, Crisis 3 in ultra 1080 has slowed it , GPU GTX770 4gb Reply
  • Mark-Benney - Thursday, August 14, 2014 - link

    I am thinking about getting the i7 4790k. But really do not think it would be worth wild. I think better wait 12month with system am running and hope for new chip from Intel 6-8core on 1150 would be very nice and worth wild upgrade Reply
  • Damage125 - Thursday, August 28, 2014 - link

    This is my system build : MB : Msi Gaming 5 Z97 / Corsair 8 gb dual channel @2400 mHz / Downgrade to 1.600 Mhz (i want lower latency on memory ) , Cpu Cooler Scythe Mugen Max 100 Air cooling / 550 PSU Cooltek Bronze 85% + / Cooltek Case .
    Ambience Temp around 25-26 ( Living in romania and its kinda hot here ) This is the frequency that my i7 works now '''' , look at the voltage :> 1.290 @4.8 Ghz ..the record i could push its 5.3 Ghz (But i`am limited cuz of my PSU i think i can get 5.5 ) On all Cores ..on the winter i plan to keep it at 5 Ghz Constantly ..but now on the summer 4.8 Its enough . And this is Screen is while i keep refresh F5 active . 2 cores Active @ Look this SHIT .. Thats 5.6 GHZ on 2 Cores . Overwall Devil Canyon despite what haters say of SB...SB cant comper with Devil Canyon its much more better , In every category , its faster , cooler , more overclockable , Cheers
  • ZenithSal - Sunday, September 14, 2014 - link

    There is a reason why Larabee was a good concept, too bad intel killed it. Desktop processors based on the current architecture are getting stale and they have hit a dead end a while ago. Competition isn't helping either since AMD is trailing way back and their only advantages right now are lower prices and a unified socket.
    On the other hand GPUs are breaking performance grounds constantly driven by fierce competition yet very little of that power is tapped into.
    Software needs to adapt, the CPU must change its role to become merely a task scheduler and do simpler tasks while any expensive processing should run on the GPU.
    AMD realized that and developed HSA for their Kaveri architecture, too bad the raw processor performance is not up to par yet compared to intel and software isn't taking advantage of its heterogeneous features.
  • Dimetime35c - Tuesday, October 13, 2015 - link

    I have a question. Okay my i5 4690k has been able to OC to 4.8Ghz and passed 12+ hour stress testing. I am wondering if I've been pushing it further then I should be seeing as in your tests you BSOD at 4.8? I'm newish to overclocking and I'm worried that sense I'm newish I am not seeing the warning signs that my oc isn't stable and even though I'm running. In fact I'm slowly killing my cpu as I'm either under volting it or doing something else that someone that knows oc would see as not right. Reply

Log in

Don't have an account? Sign up now