overclockers.at Timer bench results

[William Gaatjes]

I thought it would be fun to see the effects in a large variety of computers.
Thus i downloaded the timer bench program and see what the results would be just for fun.
https://www.overclockers.at/articles/the-hpet-bug-what-it-is-and-what-it-isnt

No need to point fingers. Just results.

Here is my old piledriver system :
WIndows 10 home , 1709. build : 16299.371
Latest updates applied:

And HPET on :

TSC vs HPET on a piledriver system with rx480 card.
I lost 2 % when looking at the framerates.
476.26 fps vs 466.13 fps.
50788.74 calls/s vs 39942.93 calls/s
is 27% less timer calls.

Anyone else want to post their results ?

 

[.vodka]

No performance loss from using HPET here.

 

[stAbb]

System:
CPU stock R5 1600x
GPU Nvidia 1060 6gb
RAM Hynix AFR at 3200mhz and custom timings.
Windows 10, Asus x370 pro, latest updates applied, specter and meltdown protected.

Synthetic test:
Timer calls around 27^10x6 with HTEP disabled, running Invariant TSC (3.51 Mhz).

Timer calls around 0,4^10x6 with HTEP enabled.

Game test done at 640x480:
Timer calls around 0,8^10x6 with HTEP disabled, running Invariant TSC (3.51 Mhz).
Average frame rate around 890.
Max frame time of 8ms

Timer calls around 0,2^10x6 with HTEP enabled.
Average frame rate around 840.
Max frame time of 11ms.

With kind regards,
stAbb

Edit: Maybe these tests just really likes ram tuned for low latency? Or it doesn't scale well with more then 6 cores / 12 threads?
Perhaps someone with an 8 core Ryzen and some free time on his hands feels like looking into this.