1 Like
Side channel attacks on x86 is a train wreck in general. I noticed a i/o speed bump when going from a 9900k to a 10700 where the silicon had the various fixes for side channel attacks.
Until we get stacked memory CPUs none of the CPUs are safe. There will be a way to attack these desktop oriented cpus that just happen to also serve servers.
2 Likes
Anyone seen any updated benchmarks with the fix enabled?
https://openbenchmarking.org/embed.php?i=2104035-PTS-RYZEN75825&sha=2ac737a8f2da&p=2
0.003% slower with the patch on Zen3. By comparison Intel’s fix has a 40% reduced i/o performance.
3 Likes
Oh I doubt he’d take your word for it tbh
Griefs refuses to believe the Linux community, Oracle, Microsoft, AMD, Nvidia or Intel.
He admits he doesn’t have any of the hardware he talks about but claims he’s knowledgeable and everyone else is lying 
The benchmarks are from the Linux community who have the patch in their hands. If he wants to believe his fantasy world instead of the people who provide technology that’s on him.
I don’t know what his experiences are so I can’t speak for that. All I can do is speak for myself. My rule of thumb is it’s best to not state things about Hardware results that you have not personally tested. Which is why everything I mention on AMD vs intel are things I have personally tested and discovered on a test bench or in my own personal rig through real world use.
Unless it’s pure opinion in which case I will state that. The only thing with AMD from my end that is pure opinion is the extreme dislike for the current iteration of AGESA. That is pure opinion and I dislike dealing with it and working with it when it comes to ram OCing and CPU OCing
1 Like