AMD sleep bug boosts Ryzen benchmarks?

S

Sith'ari

Gawd
Joined
Oct 13, 2013
Messages
573
I'm not sure if someone has posted this already, but @Linus Tech Tips , they are talking about a very interesting "bug" during AMD's RYZEN benchmarks. Check at 33:15 at Linus video :

 
Last edited:
Pieter3dnow said:
Hpet enabled in the bios would counter that.
Click to expand...

yeap, but if you see the video you will notice that the problem lies with the reviews that were made so far. How can i be sure that the data presented at these reviews aren't affected by this "bug" (*and thus, giving me a deceiving conclusion between AMD's and Intel's CPU real differences ? )
As Linus stated, with Hpet off, if the pc enters sleep mode, the benchmark deviation at scores can be huge!! (for example, 2000 instead of 1750 @cinebench benchmark as was mentioned by Linus)
 
Sith'ari said:
I'm not sure if someone has posted this already, but @Linus Tech Tips , they are talking about a very interesting "bug" during AMD's RYZEN benchmarks. Check at 33:15 at Linus video :

Click to expand...


Yes, this was posted in another thread before. There are two issues that affect Ryzen in a way that the benchmark scores reported do not reflect real performance. HWBOT has officially banned certain submissions and investigating both the RTC bias and the sleep bug

http://hwbot.org/newsflash/4335_ryz...bias_w88.110_not_allowed_on_select_benchmarks
 
And here is another interesting video describing large communication latencies among Ryzen's certain cores , compared to Intel's. (*not exactly related with thread's topic, but i feel it's related from the perspective that we are analyzing the way that Ryzen CPUs work )

 
Yeah I can't say that there saying anything useful. The whole debate about NUMA is rather silly. Windows scheduler is not broken according to them, which is weird because they are themselves suggesting that it should be improved for Ryzen.

That is a contradiction if I ever saw one.
 
You must log in or register to reply here.
Back
Top