cross-posted from: https://lemmy.ml/post/19143537
Last Wednesday was the review embargo for the Ryzen 5 9600X and Ryzen 7 9700X Zen 5 desktop processors that proved to be very exciting for Linux workloads from developers to creators to AVX-512 embracing AI and HPC workloads. Today the review embargo lifts on the Ryzen 9 9900X and Ryzen 9 9950X and as expected given the prior 6-core/8-core tests: these new chips are wild! The Ryzen 9 9900X and Ryzen 9 9950X are fabulous processors for those engaging in heavy real-world Linux workloads with excellent performance uplift and stunning power efficiency.
I have been very much enjoying my time testing out AMD’s Zen 5 wares from the Ryzen AI 300 series to the Ryzen 9000 series. The Ryzen 5 9600X / Ryzen 7 9700X were great for whetting my appetite while awaiting the Ryzen 9 9900 series. I had been very much enjoying them to the extent I was rather surprised myself last week when hearing of some reviewers not finding much excitement out of these new Zen 5 processors but typically those just looking at Windows gaming performance or running only a few canned/synthetic benchmarks. Following the Ryzen 5 9600X and Ryzen 7 9700X Linux testing when the Ryzen 9 9900X/9950X arrived, they were put immediately to my gauntlet of hundreds of Linux benchmarks and indeed living up to expectations.
So … all reports of the CPUs performance being shit or way below expectations are just due to windows. As always.
I hate reviewers doing that. It’s like buying a sports car and rating it on what you are legally allowed to drive. It doesn’t make any sense. Also, windows performance will vary extremely depending on privacy settings, edition and country (eg. due to the EUs privacy regulations). A minimal Arch setup will ALWAYS be the same in every aspect. No new shit no one needs. No performance-wasting malware by default. Just an OS.
Reviewers like Gamers Nexus and Hardware Unboxed serve the Windows gamer market first. If you game on Windows you want to know the best price/performance for your purposes. Benchmarking kernel compiles and database transactions on Linux has zero relevance to a Windows gamer, particularly if Microsoft bugs cause the performance not to translate.
If we only looked at raw hardware performance and ignored platform support we might evaluate Nvidia only on Windows and determine they are the best graphics cards for Linux users which would be insane. Platform support matters to an audience.
The argume t would be that their job is to advise buyers on the performance they would see. If most buyers are using windows and gaming then they should report windows gaming numbers.
However, there’s a lot of people who aren’t doing that and I’m not sure where we are meant to get our numbers from. Phronix I guess.
Yeah really should be testing on freebsd and minimal linux distros. Could much easier automate the entire test process then too
I would hold on the conclusion for now. Steve from HW Unboxed tested both Zen 4 and Zen 5 with the “supposed” fix and both had improved performance so the rough difference between Zen 4 and Zen 5 remained almost the same as the issue was affecting both. We will need to see more tests though to draw a reasonable conclusion. We don’t yet know if this also affects older Zen 3 at all or not.
Something is wrong with the Windows scheduler and these new chips. The Linux results aren’t revolutionary, but they’re about what you’d expect from what AMD marketed in terms of IPC uplift.
More reviewers should benchmark hardware on multiple operating systems.
The new gamer’s nexus review outlines some pretty specific prerequisites that AMD sent to fix performance on Windows, and AMD didn’t communicate those until they’d had the review units for days.
On a slight tangent, Phoronix is effin awesome
Michael has written more than 20,000 articles since 2004, plus he is the lead developer of the Phoronix Test Suite, Phoromatic, and OpenBenchmarking.org. This guy is a one man army!
This is… Interesting. I would love for gamers nexus to investigate this tbh. Means something is horribly wrong in windows ( shown by the wtf steps reviewers had to go through ).
Im also curious at the performance uplift of zen5 in linux in regard of handbrake. Amd claimed a 40% uplift there which i guess might have been in linux and with a very specific clip?
Handbrake uses avx512 and zen5 significantly improved on avx512
That’s encouraging, 3ven though these models are out of my price range.
I’m planning to build a new system pretty soon. With Intel 13 an 14th Gen woes, AMD CPU releases and upcoming (Septmber) AM5 Motherboards, my planning is in constant flux.
