Request: Prime95 bench on Kaby Lake

Can I get some volunteers to run a custom Prime95 benchmark on Kaby Lake? I just want to check that KL does have same IPC as Skylake in this use, as there is still one claim otherwise. They have no credibility or evidence so I don't believe it, but I just want to have hard data to know for sure.

Use version 28.10 http://www.mersenne.org/download/

Add the following lines to the top of prime.txt
MinBenchFFT=128
MaxBenchFFT=128
If the file doesn't exist, it will get created after you run it once, but do make sure to exit Prime95 completely before editing otherwise it will overwrite it on close.

The lines fix the benchmark to one FFT size which is small enough to run out of CPU cache, so ram speed doesn't matter. Results should scale ideally with CPU clock. I'm just looking to get some sample points on peak IPC here. One run should be ok, but I would recommend running it a few times and pick the best. Higher throughput is better. It is pretty quick.

Results are saved in results.txt. I just need the CPU model, the clock it is running at during the benchmark, and the "throughput" results line e.g.
Timings for 128K FFT length (4 cpus, 4 workers): 0.57, 0.52, 0.47, 0.45 ms. Throughput: 8017.20 iter/sec.

There may be two lines if you have HT on, if so please post both.

Predicted results based on my Skylake system, assuming KL has same IPC:
GHz iter/sec
4000 8893
4200 9338
4400 9782
4600 10227
4800 10672
5000 11116

Above assumes a quad core. If dual core, halve the expected values. If we get repeatable and different results from these (more than a % or so), that would be interesting and worthy of further investigation.

I only did one run on my i7 7700k @ 4.7ghz:

Timings for 128K FFT length (4 cpus, 4 workers): 0.35, 0.32, 0.39, 0.39 ms. Throughput: 11128.26 iter/sec.
Timings for 128K FFT length (4 cpus hyperthreaded, 4 workers): 0.39, 0.34, 0.34, 0.33 ms. Throughput: 11522.21 iter/sec.

*on edit*
Seems to be fairly variable, I tried it a couple of more times and got this:
Timings for 128K FFT length (4 cpus, 4 workers): 0.36, 0.32, 0.32, 0.32 ms. Throughput: 12106.94 iter/sec.
Timings for 128K FFT length (4 cpus hyperthreaded, 4 workers): 0.38, 0.33, 0.33, 0.32 ms. Throughput: 11767.90 iter/sec.

Thanks for the results. There will be some variation depending on what else is running at the time. Still, that's above expectations, so either my 6700k was running below par, your 7700k is above par, or there is some other factor. Now I'm wondering if L3 cache speed makes much difference. I'll re-run on my 6700k first to make sure it wasn't having a bad day at the time.

I ran it on my 2nd 6700k (4.2 GHz) and got a best of 11719 in 3 tries. So... the 7700k isn't looking so fast after all, and my main even slower. Maybe this is more sensitive to background CPU than I thought...

Yeah I think background processes mess with it. I closed Outlook and One Note and this is what I got:
Timings for 128K FFT length (4 cpus, 4 workers): 0.30, 0.30, 0.30, 0.30 ms. Throughput: 13230.78 iter/sec.
Timings for 128K FFT length (4 cpus hyperthreaded, 4 workers): 0.31, 0.31, 0.31, 0.31 ms. Throughput: 12993.96 iter/sec.

*on edit*
For fun I ran it at 4.2ghz:
Timings for 128K FFT length (4 cpus, 4 workers): 0.34, 0.34, 0.34, 0.34 ms. Throughput: 11896.42 iter/sec.
Timings for 128K FFT length (4 cpus hyperthreaded, 4 workers): 0.35, 0.35, 0.35, 0.35 ms. Throughput: 11570.17 iter/sec.

Thanks again. Did you pick the best of multiple runs or was it just one? Your 4.2 result was 1.5% faster, and the 4.7 result was 0.9% faster after clock adjustment. Now, I might still not have got a great run, or KL might really have a fractional % improvement over Skylake, but either way the difference isn't significant and I think I'll leave it here.

mackerel said:

Thanks again. Did you pick the best of multiple runs or was it just one? Your 4.2 result was 1.5% faster, and the 4.7 result was 0.9% faster after clock adjustment. Now, I might still not have got a great run, or KL might really have a fractional % improvement over Skylake, but either way the difference isn't significant and I think I'll leave it here.

Click to expand...

I did the best of 3 at 4.2ghz (I also closed some background programs like steam which I didn't do before). I don't think there is any doubt that kaby lake isn't a huge improvement over skylake. For me the upgrade was good going from an i5 to an i7, especially with the overclock potential of kaby lake.