The Intel 6th Gen Skylake Review: Core i7-6700K and i5-6600K Tested
by Ian Cutress on August 5, 2015 8:00 AM ESTWhat You Can Buy: Linux Performance
Built around several freely available benchmarks for Linux, Linux-Bench is a project spearheaded by Patrick at ServeTheHome to streamline about a dozen of these tests in a single neat package run via a set of three commands using an Ubuntu 11.04 LiveCD. These tests include fluid dynamics used by NASA, ray-tracing, OpenSSL, molecular modeling, and a scalable data structure server for web deployments. We run Linux-Bench and have chosen to report a select few of the tests that rely on CPU and DRAM speed.
C-Ray: link
C-Ray is a simple ray-tracing program that focuses almost exclusively on processor performance rather than DRAM access. The test in Linux-Bench renders a heavy complex scene offering a large scalable scenario.
NAMD, Scalable Molecular Dynamics: link
Developed by the Theoretical and Computational Biophysics Group at the University of Illinois at Urbana-Champaign, NAMD is a set of parallel molecular dynamics codes for extreme parallelization up to and beyond 200,000 cores. The reference paper detailing NAMD has over 4000 citations, and our testing runs a small simulation where the calculation steps per unit time is the output vector.
NPB, Fluid Dynamics: link
Aside from LINPACK, there are many other ways to benchmark supercomputers in terms of how effective they are for various types of mathematical processes. The NAS Parallel Benchmarks (NPB) are a set of small programs originally designed for NASA to test their supercomputers in terms of fluid dynamics simulations, useful for airflow reactions and design.
Redis: link
Many of the online applications rely on key-value caches and data structure servers to operate. Redis is an open-source, scalable web technology with a b developer base, but also relies heavily on memory bandwidth as well as CPU performance.
477 Comments
View All Comments
boeush - Friday, August 7, 2015 - link
A good point, but I think you missed this page in the review:http://www.anandtech.com/show/9483/intel-skylake-r...
The other pages where all CPUs are normalized to 3 Ghz are for generational IPC comparison, not memory scaling. The later "what you can buy" pages repeat all the same tests but with all CPUs at full default clocks, as well - to gauge the combined effect of IPC and frequency scaling across generations.
Still missing and hopefully to be addressed in a future follow-up, is a study of generational overclocked performance, and performance under DDR4 frequency scaling with and without CPU (other than memory) overclocking.
MrSpadge - Friday, August 7, 2015 - link
Well, on the page before the one you linked Ian says:"For these tests, both sets of numbers were run at 3.0 GHz with hyperthreading disabled. Memory speeds were DDR4-2133 C15 and DDR3-1866 C9 respectively."
I think this applies to both memory scaling pages.
You've got a good point, though, that the "what you can buy" section compares DDR4-2133 and DDR3-1600 (latency unspecified) at default CPU clocks. And from a quick glance the differences there are not that different from the ones obtained in the dedicated memory scaling section.
Nutti - Friday, August 7, 2015 - link
Left out all the AMD FX processors? Looks pretty bad for AMD this way. FX is still much better than 7870K. Zen will nicely catch up with Intel. AMD needs 40% improvement over FX8350 and they will sure get that through better IPC and multithreading.Bambooz - Friday, August 7, 2015 - link
Wishful thinking from a fanboiOxford Guy - Friday, August 7, 2015 - link
Ad hominem isn't a rebuttal, bud.Oxford Guy - Friday, August 7, 2015 - link
The FX does nicely in a modern game like Witcher 3 that uses all of its threads as can be seen here: http://www.techspot.com/review/1006-the-witcher-3-...Anandtech has been doing the "let's throw in a dozen APUs and completely ignore FX" for some time now. The only thing it accomplishes is obscuring the fact that the FX can be a better value for a workstation (rendering and such) that also has some gaming requirements.
nils_ - Friday, August 7, 2015 - link
You probably should have run the Linux Tests through Phoronix Test Suite, the Linux Bench seems rather outdated with Ubuntu 11.04 (we are on 15.04 now).eeessttaa - Friday, August 7, 2015 - link
Great article as always. I wish intel would leave the fivr in it. i know how hot it got but instead of removing it they should've improved on its design.Nelviego - Friday, August 7, 2015 - link
Seems it might finally be time to OC my i7 2600k and give it another 4 1/2 years. ;)Oxford Guy - Friday, August 7, 2015 - link
Intel made everyone think Skylake was going to be a massive improvement on all fronts. Massive IPC increase. Massive technological advance. People shilled for Intel by claiming it was highly likely that Skylake wouldn't need a new socket and would just use LGA 2011-3.Instead, we get ... what? Chips that aren't significantly better than Haswells, let alone Broadwell?
I guess Intel is sandbagging even more than usual since AMD isn't doing anything new on the CPU front. So much for all of the intense Skylake hype. It amazes me, too, how people are blithely now saying "I guess I'll wait for Kaby Lake" -- the same people, often enough, who said Skylake would revolutionize computing.
It looks like this is what happens when Intel has minimal competition. The FX chips are still clinging to relevance now that consoles have 8 threads and weak individual cores (not that you'd know it based on the way this site never puts even one of them into its reviews in favor of weaker APUs) -- and because rendering programs like Blender can use their threads which can make them a decent value for budget workstation use, but their design is from 2012 or so. Overclocking is also keeping old chips like the 2500K viable for gaming.
I admit I feel for the hype a bit. I was expecting at least some sort of paradigm-shifting new tech. Instead... I don't see anything impressive at all. A new socket... a small gain in efficiency... rinse repeat.
An article I read recently said that overclocking will become increasingly non-viable as process nodes shrink. It seems we're seeing that already. The article says an Intel executive said Intel is taking overclocking seriously but the company may not have much choice.
Intel should have included hardware devoted to h.265 encoding for Skylake at least. Maybe it did, but it's not like I can tell by the charts provided. What is the point of putting in that h.265 encoding chart and not including the fastest non-E Haswell (4790K) and a Haswell-E (5820K)? It makes it look like your site is trying to hype Skylake. Don't you think people who are doing a lot of tasks like that which require serious performance (like the "slowest" setting in Handbrake) are going to need to see a comparison with the best available options?