Kaby Lake Benchmarks Might Have Been Leaked

Subject: Processors | May 9, 2016 - 12:51 PM |
Tagged: kaby lake, Intel

Fudzilla claims that they have a screenshot of SiSoft benchmarks belonging to the Intel Core i7-7700k. I should note that image only mentions “Kabylake,” not any specific model number. It's possible that the branding will change this generation, and there's an infinitesimal chance that this is not highest level SKU of that specific chip, but it should be safe to assume that this is the 7700k, and that it will be branded as such. I'm just being over-cautious.

View Full Size

Image Credit: Fudzilla

In terms of specifications, Kaby Lake will be a quad-core processor that runs at 3.6 GHz, 4.2 GHz turbo, backed with 8MB of L3 cache. The graphics processor has 24 CUs that can reach a clock of 1.15 GHz. If Intel hasn't changed the GPU architecture since Skylake, this equates to 192 FP32 processors and 442 GFLOPs. Apart from a lower CPU base clock, 3.6 GHz versus Skylake's 4.0 GHz, Kaby Lake seems to be identical to Skylake.

I was hoping to compare the benchmark results with Core i7-6700k, but I'm not sure which version of SiSoft they're using. The numbers don't seem to line up with our results (SiSoft 2013 SP3a) or the SiSoft 2015 benchmarks that I've found around the net (and even those 2015 benchmarks varied greatly). It might just be my lack of experience with CPU benchmarks, but I'd rather just present the data.

Source: Fudzilla

May 9, 2016 | 01:05 PM - Posted by johnc (not verified)

Intel has become probably the most boring technology company out there.

May 9, 2016 | 01:36 PM - Posted by Anonymous (not verified)

Well hell, the only reason why AMD isn't is because they had to go out and acquire ATI and take on their product line. If they hadn't done that, I'm sure AMD wouldn't even have anywhere near the performance in their APUs or if they would even HAVE APUs to spice up their old lineup. Aside from APUs, the most recent CPU they've launched was like 2011.

May 9, 2016 | 02:30 PM - Posted by Anonymous (not verified)

That ATI product line kept AMD in business in spite of AMD's crappy management, and really crappy management decisions! Going with CMT, especially with the pre DX12/Vulkan graphics APIs unable to take advantage of multi-core CPUs. With DX12/Vulkan here even some of the older FX CPUs/APU are able to better perform on games.

So now AMD is going with SMT on its CPU cores and the Zen cores will have better IPC metrics, they still may not beat Intel's latest, but single core performance is not that important now that DX12 and Vulkan are able to do more on more than one CPU core. And new to the DX12/Vulkan APIs is those graphics API's ability to take advantage of any asynchronous compute ability on the GPU to move more of the non-graphics gaming compute off of the CPU and on to the GPU, in addition to the graphics compute, and reduce the VR gaming latency issues associated with CPU to GPU over PCIe communication latency issues for VR gaming/games! Even Nvidia is improving its hardware asynchronous compute ability so the CPU core/cores will not be so taxed anymore.

May 9, 2016 | 01:47 PM - Posted by Barack_McBush (not verified)

Yep, and it will stay that way until AMD beats them for performance.

May 9, 2016 | 02:08 PM - Posted by Anonymous (not verified)

AMD's integrated APU graphics beats the hell out of Intel's for more than just gaming uses, and Intel is quick to brag about its best graphics, but never uses it in its lower cost SOC SKUs. Even Nvidia's Tegra K1/M1 graphics gives Intel a hard time, for the price, Ditto for IT's integrated graphics on Phone/Tablet SKUs!

May 9, 2016 | 01:59 PM - Posted by Anonymous (not verified)

Intel management's shirts are so full of starch that if you where to dip them in an Olympic swimming pool the water would become so congealed that you could easily drive an M1A2 Abrams main battle tank over it and not bog it down a millimeter on its way across!

Intel graphics/integrated GPUs so devoid of execution resources that it can only run at much higher clocks than normal AMD/Nvidia/PowerVR GPUs and with only the low polygon meshes used for gaming! Don't try any high polygon mesh scenes or high polygon mesh editing for non gaming graphics uses on Intel's SOCs, so gimped of integrated GPU execution resources!

Intel the company that has forever use questionable market tactics to keep its SOCs with their crappy graphics on the majority of laptop SKUs, with users unable to get the Intel CPU without the crappy Intel graphics and get the much better graphics solutions from Nvidia and AMD. Even high end gamers are only Interested in Intel's CPU cores but not Intel's graphics!

May 9, 2016 | 06:55 PM - Posted by Anonymous (not verified)

Indeed, it is boring to count on your foundries to shrink dies instead of making an efficient new design but AMD which split its foundries to Global Foundries didn't really cut the link with its outdated foundries owned by arabs from Qatar too.

Now foundries evolution is rapidly slowing even for Intel which uses an expensive multigate design for its new processors, AMD have an opportunity to compete with Intel.

I'm afraid but despite the market have a lot of expectancies on the future Zen architecture, AMD will disappoint a lot for its overhyped products.

I think if Intel or AMD can't provide an efficient new design, it is an opportunity for new competitors to enter into the market (nVidia? Apple? Google? Amazon? Tesla?).

May 9, 2016 | 07:17 PM - Posted by Anonymous (not verified)

GlobalFoundries is using a licensed from Samsung 14nm process for Zen/Polaris so its coming from Samsung via some help from IBM. And IBM, Samsung, and GlobalFoundries have been a chip fab IP/technology sharing consortium/foundation for a more than a few years now! IBM's out of the chip fab business for its power8, power9s and GlobalFoundries got that 14nm process licnese from Samsung because they are all in with IBM(GlobalFoundries is fabbing for IBM)to get more 14nm chip lines up and running for IBM/other third party openpower licensees benefit!

It's about IBM covering its power8+ and power9 fab capcity needs and for the openpower licensees' 14nm power9 fab capcity needs! And the entire market for 14nm processors is benefiting.

May 10, 2016 | 03:15 PM - Posted by BlackDove (not verified)

Knights Landing, Knights Hill, integrated FPGAs, XPoint, NVMe, Omnipath, Skylake Purley, buying Altera, silicon photonics is BORING?

May 10, 2016 | 03:39 PM - Posted by Anonymous (not verified)

All of the Knights of Larrabee, are higher clocked power hungry x86(ATOM) cores that utilize way too much energy for the FP performance that they provide, and Nvidia's and AMD newer more power efficient and relatively lower clocked GPU accelerators provide better DP-FP performance/Watt than any one of those Knights who say: (Ni)-Need more parallel execution resources in your gimped down GPUs Intel! More parallel GPU execution resources at lower clocks for the Exascale computing market, and the consumer non gaming/gaming graphics markets!

Stop beating that dead x86 ISA only, one ISA trick pony, business model, Intel! And get your GPU's lack of enough parallel execution resources at lower power saving clock rates issues fixed!

All that money wasted on Contra Revenue, Intel, what a waste!

May 10, 2016 | 08:51 PM - Posted by BlackDove (not verified)

I am pretty sure a SPARC machine will be the first real exascale supercomputer, but ok.

PEAK theoretical FLOPS and Linpack are pretty irrelevant these days. Let's wait for the HPCG results of a GP100 or GV100 system, compared to a Knight's Landing or Knights Hill system, or some Rmax vs Rpeak ratios before you say that either chip has won the race.

May 11, 2016 | 04:49 PM - Posted by Anonymous (not verified)

There will be no exascale supercomputers without GPU accelerators from Nvidia/AMD, OR APUs on an interposer From AMD.

Those AMD exascale APUs on an interposer with not only 4096+ traces to the HBM stacks, but also Way MORE traces from Zen CPU cores DIE complex to GPU DIE complex for thousands of traces wide direct CPU to GPU on interposer TBs of lower clocked power saving High effective bandwidth connection fabrics! No NVLink or other connect technology will be able to best the APU on an interposer's CPU to GPU effective bandwidth at many TBs, or the power savings. No only CPU(CPUs are Mooks at massively parallel computations) designs for exascale computing!

Those AMD exascale APUs on an interposer will have plenty of wide parallel traces etched into the silicon interposer's substrate, and the interposer will go from the being a passive design, and just hosting traces, to a more active design with whole memory buffers/stores etched into the interposer, and/or intelligent fully coherent logic circuits to go along with the 10 of thousands of traces.

Look for Navi based exascale systems where the GPU part will be made up of smaller modular GPU units all wired up like they where a single GPU on a single monolithic Die! So AMD with Navi will have a modular GPU system in which to build APUs on an interposer module for all markets just by scaling both GPU modular die units, and CPU modular die clusters!

And do not forget the extra in the HBM memory stacks FPGA compute that AMD will be adding to the HBM stack sandwiched between the HBM's bottom controller logic die and the HBM's memory die stacks above, AMD has a patent pending for that FPGA added to the HBM stack/s for in memory compute!

U-Sam is awarding billions in Exascale grants to get those Exascale systems built and not needing their own Nuke plants for a power source!

CPUs are Mooks! CPUs are Mooks!

May 9, 2016 | 03:21 PM - Posted by Anonymouse (not verified)

>Intel the company that has forever use questionable market tactics to keep its SOCs with their crappy graphics on the majority of laptop SKUs...<

And to think they were fined only $1.25 billion (IIRC) for their collusion and dirty practices over the years. The loss to AMD was surely greater.

May 9, 2016 | 07:15 PM - Posted by Anonymous (not verified)

AMD accepted $1.25 billion plus a 5-year patent cross-licensing agreement to continue the production of x86 compatible processors.
Without the latter agreement AMD would be out of business even with a big check from Intel.

May 9, 2016 | 07:38 PM - Posted by Anonymous (not verified)

Not really with AMD being the creator of the x86 64bit ISA that even Intel is using! Itanium with its 64 bit ISA was a big failure for Intel, and Intel was forced to license AMD's x86 64 bit ISA!

Those Chip Pimps at Intel, with their terrible graphics, and terribley high priced at that, for what you get from Intel. Nvidia and AMD have nuch better graphics at a much more affordable prices!

What an Itanic Flop from Intel.

May 9, 2016 | 09:36 PM - Posted by Anonymous (not verified)

"Chipzilla veteran joins IBM's OpenPOWER

Rani Borkar borged by Big Blue"

There you go Intel "Redundent" employees! Get some "Chipzilla veteran joins IBM's OpenPOWER

Rani Borkar borged by Big Blue"

There you go Intel "Redundent" employees! Get some OpenPOWER employment going, that market is just starting up with the ARM holdings style Licensing of Power8s/power9s and some Green Team GPU accelerator business, better get into some of that OpenPOWER market youself Red Team GPU accelerator maker!

http://www.theregister.co.uk/2016/05/10/chipzilla_veteran_joins_ibms_ope...

May 9, 2016 | 10:36 PM - Posted by Anonymous (not verified)

Great! IBM will at least die recruiting Intel incompetent employees...

May 9, 2016 | 11:53 PM - Posted by Anonymous (not verified)

Who the F_ck cares about IBM, it's the third party licensees that will be licensing the OpenPower power8, and power9, designs that will matter! IBM will never go completely out of business their R&D and patent licensing part will continue on no matter what! For all I give a F_ck about it IBM could become just an IP licensor like ARM Holdings! It's the the Power8/Power9 designs in the hands of the third party licensees that will give those Chip Pimps at Intel some well deserved headaches, with some Nvidia GPU accelerators to top that pain off!

May 10, 2016 | 11:56 AM - Posted by Anonymous (not verified)

"AMD's chips aren't good enough! Fucking AMD!"

"Intel's chips aren't improving enough! Fucking AMD!"

"My dog was 17 years old and died last night! Fucking AMD!"

"Oh no, I overslept and I'm late for work! Fucking AMD!"

May 10, 2016 | 04:00 PM - Posted by Anonymous (not verified)

"AMD's chips aren't good enough! Fucking AMD"

That's right, AMD and all of YOU filthy suppliers of parts to the OEMs! Including M$'s OS part to the third party PC/Laptop OEMs! The third party laptop OEMs need only the windows 7 OS with the under the hood improvements from you, M$, you are a lowly supplier of the OS part, ditto for Intel, ditto for Nvidia, and any other parts supplier to any OEM PC/laptop/tablet/phone market!

Good work on that Zen by the way, AMD! You APU/GPU parts supplier scum to the various OEMs, AMD scum! Now get it even better at a lower cost! And AMD is right where any lowly supplier of parts should be, as any healthy and non-sycophant parts supplier market should ever let its part suppliers ever become! Down and Out an engineering Just to stay in business, hear that Intel, Nvidia, and other lowly parts suppliers to the OEM market-place of part supplies by parts supplier scum(M$)!!!

May 10, 2016 | 04:56 PM - Posted by Anonymous (not verified)

Psst... hey... that thing that flew past your head and you completely missed it?

It was the point.

May 10, 2016 | 06:07 PM - Posted by Anonymous (not verified)

No I saw your point, and I hijacked it to post my diatribe on just how much the market FOR supplies of parts to the third party OEM PC/Laptop Makers is sick and has been in a sycophant state for around the last 35 years! Those SOC/CPU/OS supply markets have been so sycophant for the last 35 years with M$'s/Intel's and others' sycophant domination of the OEM's supply chain of those very necessary parts!

For Normal markets AMD/Intel/Nvidia/others must be kept poor(relatively) and struggling for the good of competition and the consumers of the third party OEM produced end products! The OEM markets have been so disrupted(in a very bad way) that some very serious action needs to take place to enforce fair markets!

Suppliers of parts do not need to be allowed to acquire such a large market share of the essential parts supply chain that they hurt the overall market, Both Intel and M$/others need to be put back into their proper supplier of parts roles!

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Lines and paragraphs break automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd> <blockquote><p><br>
  • Web page addresses and e-mail addresses turn into links automatically.

More information about formatting options

CAPTCHA
This question is for testing whether you are a human visitor and to prevent automated spam submissions.