Over the holiday weekend a leaked Geekbench benchmark result on an engineering sample AMD Zen processor got tech nerds talking. Other than the showcase that AMD presented a couple weeks back using the Blender render engine, the only information we have on performance claims come from AMD touting a "40% IPC increase" over the latest Bulldozer derivative.
The results from Geekbench show performance from a two physical processor system and a total of 64 cores running at 1.44 GHz. Obviously that clock speed is exceptionally low; AMD demoed Summit Ridge running at 3.0 GHz in the showcase mentioned above. But this does give us an interesting data point with which to do some performance extrapolation. If we assume perfect clock speed scaling, we can guess at performance levels that AMD Zen might see at various clocks.
I needed a quick comparison point and found this Geekbench result from a Xeon E7-8857 v2 running at 3.6 GHz. That is an Ivy Bridge based architecture and though the system has 48 cores, we are only going to a look at single threaded results to focus on the IPC story.
Obviously there are a ton of caveats with looking at data like this. It's possible that AMD Zen platform was running in a very sub-optimal condition. It's possible that the BIOS and motherboard weren't fully cache aware (though I would hope that wouldn't be the case this late in the game). It's possible that the Linux OS was somehow holding back performance of the Zen architecture and needs update. There are many reasons why you shouldn't consider this data a final decision yet; but that doesn't make it any less interesting to see.
In the two graphs below I divide the collection of single threaded results from Geekbench into two halves and there are three data points for each benchmark. The blue line represents the Xeon Ivy Bridge processor running at 3.6 GHz. The light green line shows the results from the AMD Zen processor running at 1.44 GHz as reported by Geekbench. The dark green line shows an extrapolated AMD Zen performance result with perfect scaling by frequency.
Continue reading our preview of AMD Zen single threaded performance!!
There are three tests in this benchmark, the AES encryption, Canny and Gaussian Blur, where the AMD Zen extrapolated results have an edge over Intel's Ivy Bridge. The AES results are probably the most compelling though we have definitely seen AES performance increases in Haswell, Broadwell and Skylake since this Xeon processor was released in early 2014.
Basically all of the other compute task based tests show the advantage going to Intel's Xeon processor running at 3.6 GHz. On average, if these results are at all indicative of what AMD Zen will be in the next 4 months, then it will be ~70-80% of the IPC of Ivy Bridge.
This relative performance graph removes the last three memory tests as it is very unlikely that memory bandwidth or latency will shift by simple linear scaling with clock speed.
What does this mean for AMD and the Zen architecture? It could mean absolutely nothing if the benchmark results we have here from a Zen CPU running at 1.44 GHz are simply not accurate, or not optimized. It might tell us that AMD and its partners have some work to do to optimize the platform in general for better overall performance, but with only months until stated product availability and partners already receiving samples this month, it looks less likely that we'll see dramatic changes from the numbers above.
This is Zen going up against Ivy Bridge and a Xeon processor that was released in Q1 of 2014. AMD needs to make its next CPU competitive not just against this but against Skylake-K and Broadwell-E systems if they want to impress the consumer and the press like they did back at IDF in August. Is it possible that AMD will be able to get Zen to 4.0 GHz and beyond in 8-core systems and thus will utilize clock speed to make up some of this IPC performance gap we show here? It's possible – though AMD needs to do it while maintaining processor thermals within a reasonable level. Doing some quick napkin math, in order for the AMD Zen results to swap wins and loses with the 3.6 GHz Xeon running on Ivy Bridge, this pre-production setup would have to run at around 4.8 GHz.
Again, I must reiterate, the results and graphs we have here are nowhere near finalized and could be an AMD red herring or the result of very early prototype hardware.
If the summer of 2016 was the season of GPUs, it looks like winter 2016/2017 is about do the same for CPUs.
Ryan…
Can we go back to
Ryan…
Can we go back to AMDMB yet? Been waiting a decade.
Extrapolating 1.44ghz results
Extrapolating 1.44ghz results to 3.6ghz? Wtf is this garbage? I expect better from PCPer. Lay off the bourbon while at work.
It’s a simple
It’s a simple multiplier:
Performance = IPC * Clock * Number of Cores
Easy to extrapolate when three of the four variables are known; just solve for IPC, then redo the calculation using a different Clock and resolve for Performance.
For many scientific
For many scientific applications memory bandwidth is the single most important determinant of performance. AMD Zen seems to do very well there. I’m not saying the other benchmarks aren’t important because they really are. But for certain workloads bandwidth to memory is really the main determinant of speed.
Is this the magical desktop
Is this the magical desktop 4C8T Zen CPU that we’ve been waiting for?
http://browser.primatelabs.com/v4/cpu/199148
That doesn’t look Right? The
That doesn’t look Right? The cache is missing in L3?
No.
See:AuthenticAMD Family
No.
See:AuthenticAMD Family 21 Model 16 Stepping 1
Zen is Family 23(or Family 17h in hex).
The processor ID
The processor ID (AuthenticAMD Family 21 Model 16 Stepping 1) points to that processor being an A10-4600M.
The Samsung model number given (SAMSUNG ELECTRONICS CO., LTD. 355V4C/356V4C/3445VC/3545VC) and the Motherboard model number given (SAMSUNG ELECTRONICS CO., LTD. NP365E5C-S03US) points to it being a Samsung-branded Windows 8 laptop. That particular model appears on Samsung’s website as an A6-equipped model, so likely whoever bought it had the APU upgraded.
So were these test’s for the
So were these test’s for the Server CPU’s coming from the Zen line? My understanding is there’s different CPUs for this whole lineup of AM4 socket.
Like Bristol Ridge. Are these test’s for their server cpu and not Desktop?
I just want to go team red on my new build at the end of the year, and if they can give me i7 4770k performance or close for 200$ I’m in.
If they are more in line with lower i5, then I guess I go intel again and pay their ridiculous prices.
you should post the scores of
you should post the scores of a FX-8370 for reference.
Here’s
Here’s one.
http://browser.primatelabs.com/v4/cpu/216470
To Ryan and all of you:
Take
To Ryan and all of you:
Take a nice look at memory bandwidth…ST will be higher than that.
well, for kicks i lookt
well, for kicks i lookt up
AMD Athlon X4 845
result on geekbench. actually ifound it only on geekbench 3.
single core wise it scores on 1.9 ghz something roughly 2400 points.
the leaked info shows 1140 points at 1.44 ghz.
clearly a non-winner, even if i would scale it up by the freqvency.
what i suspect is, this is a sample made to see how well performance scales with core count.
probably a test chip to see what is going on if you squash 64 cores into 1 machine.
what i allso did, is compare it with the oldtimer fx series.
the fx 6300 can be found in geekbench4, so it was a not so hard thing. surely its verry old, but maybe it will give something for us who use that old am3+ platform.
single core the fx 6300 avarages arond 2500 points at 3,52 ghz, ad i do have to add that this is totally made up batshit crazy bullshit.
in real life things don’t scale as well, its not 1:1 clock wise.
but it does sortha gives a bit of room for speculating.
surely the final product will give the final word, but if i was to upgrade from my curent fx 6300 that i bough utterly cheap, then i guess if the performance would scale cock to clock like this, i would hazard i would choose it.
will see how high will it closck, and how well will a desktop with 8 cores work.
as long as it can do
as long as it can do rendering as fast as the same thread count of an intel cpu i will consider it.
I think your article was
I think your article was great, very informative explaining these supposed leaked benchmarks. People have to understand that engineering samples even if they are a few months away from launch, can and usually do have significant performance increases but this is the first benchmark I’ve seen where the sample is supposedly clocked @ 3.6 Ghz because the engineering samples are clocked @ 2.8 with a 3.0 turbo clock. But the consumer can make his or her own tweaks to increase IPC, I had a Phenom Iix4 980 with an overclocked north bridge @ 2.8 and CPU frequency @ 4.0 and it out performed my Intel 6 core Gulftown clocked @ 3.2 HT enabled in Cinebench, Passmark and AIDA 64.
If you look at the actual
If you look at the actual leaked page : http://browser.primatelabs.com/v4/cpu/105227
You will see that the results are actually FAKE. They mixed up the following:
L1 Instruction Cache 32 KB x 32
L1 Data Cache 64 KB x 32
If you look at the HotChip slides and PCPers article you will notice that the cache sizes are:
L1 Instruction Cache 64 KB x 32
L1 Data Cache 32 KB x 32
Geekbench doesn’t scale
Geekbench doesn’t scale linearly wityh frequency, thus the Zen results extrapolated to 3.6GHz are invalid. This can be observed comparing single thread score for Haswell at different clocks
Intel Haswell @1.6GHz: 1804 pts / 1.6 GHz -> 1127 pts / GHz
Intel Haswell @3.6GHz: 3075 pts / 3.6 GHz -> 854 pts / GHz
Linear scaling [(1804/1.6GHz)*3.6GHz] would give a wrong value 4059 for the Haswell chip at 3.6GHz. Linear scaling is giving fictitious IPC advantage to the low-clocked chips. You can do two things:
(A)
Apply an approx correction factor to linear scaling: e.g. 3075/4059 ~0.76 to all Zen scores at 3.6GHz.
(B)
Repeat the comparison using instead an Intel Xeon chip clocked similarly to Zen
https://browser.primatelabs.com/v4/cpu/117877
http://ark.intel.com/products/83349/Intel-Xeon-Processor-E5-2603-v3-15M-Cache-1_60-GHz
This is a 6-core Haswell Xeon clocked so low as 1.6GHz without Turbo and without Hypertreading.