TimerBench: Ein Benchmark für Windows Timer

Thanks for taking the time to post here! It depends a lot on the platform. Intel platforms up to Kaby Lake including Broadwell-E are not affected. They have a very fast HPET timer that can be called up to 1.4 million times a second. Performance goes to hell with Skylake X where the number of possible HPET calls goes back to 200.000 per second.

With AMD it is difficult to say currently. I don't think the APUs are affected. Ryzen shows performance differences though, but only if a lot of timer calls are made like during a game and the GPU is strong enough to shift the bottleneck to the CPU. It would cap an already high framerate, that's why this has gone under the radar for so long. The PassMark tests could show a difference in 3D with the right setup, but not with an IGP in a higher resolution. That's always GPU bound.

Btw a normal CPU benchmark won't be held back, because these benchmarks use timing very little. Sometimes it's just one call at the beginning and one in the end.

You should read my new article about this, it's finally in English: https://www.overclockers.at/article...nd-what-it-isnt

I am curious what your problem with TimerBench was. A screenshot or an error message would be appreciated!

https://www.anandtech.com/show/1267...ryzen-results/5

Zitat

Matthias from Overclockers.at reached out to me and linked me to his article on how they have previously encountered the issue. The article is a nice read, and well worth clicking through.

Matthais explains how during their X299 testing, they were experiencing slowdown in their game benchmarks, and pin-pointing the problem with HPET. (We also had similar issues, and didn’t post results, but never got to the bottom of the issue.) As a result, the team over at Overclockers.at developed a tool called TimerBench in order to determine the effect of HPET. As noted, HPET has a much longer latency, but is more accurate.

In the results from overclockers.at one metric stood out: moving from Broadwell-E to Skylake-X meant that the number of theoretical peak HPET calls per second reduced from 1.4 million to 0.2 million – the latency to make a HPET call suddenly became 7x longer with Skylake-X. TimerBench, the tool developed, provides an Unreal 4.7.2 scene and measures timer calls between a system running a game, and one without.

Wir haben geplaudert, aber den Link hat er von meinem Facebook-Account.

nice

Not sure how to post a picture on here so I shared it from my google drive.

https://drive.google.com/open?id=1O...hV-xPHCXCncCTZf

Falls es jemanden interessiert: Wir hatten auf unseren englischen HPET-Artikel in den letzten 24h 6188 Zugriffe. Er wird also fleißig gelesen. Ich bin mittlerweile auch in Kontakt mit jemandem vom Benchmark-Team bei Intel. Mal schauen, was dabei herauskommt.

Edit: Und fast 500x wurde TimerBench heruntergeladen.

Zitat aus einem Post von mat

Thank you! It looks like there is a problem with Unreal Engine 4 on your system. Is this a DirectX 11 compatible graphics unit?

Funny thing is, my other machine ran it fine twice then on the third time it failed in the same manner. As for DX11 compatible, I believe so but thats a good question.

Bin auch endlich dazu gekommen