holy shithttps://www.phoronix.com/review/amd-9950x3d2-windows-linux
>>108887183>9950X3D2what's going to be their next naming scheme?
are any of these benchmark simple basically one click done deals? ill run some in about an hour if yes and if anybody cares.literally never trust reviewers btw.
>>108887183why does he use uubuntu it's dogshit
>>108887253>literally never trust reviewers btw.it's literally just pts dudehttps://www.phoronix-test-suite.com/https://openbenchmarking.org/
>>108887422cool. imma do et.
>>108887183assuming you obsessively encode av1 video files all day and night, sure.why do you do that?
>>108887681You don't have to encode av1 video files all day and night, ffmpeg is just one of the tests that shows the biggest difference. If you actually read the article, you'd notice that the geometric mean of all test results still shows a 15% performance difference, even on something as shitty as ubuntu.
>>108887238AMD Ryzen™ 9 9950X3D2ULTRAPROMAXAI
>>108887183that's such a huge difference, what's even going on there? something must be going very wrong with windows 11 on this ryzen chip
>>108888168Remember the huge Drama, when Windows nerfed AMD in purpose? Windows and Intel were always close + gov Deals.
>>108887183>>108888168ryzen needs correct scheduler setup or you lose a ton of performance for no reason. its a mess on both linux & windows. in this bios theres three options for cppc on am5. frequency,cache,driver. driver is technically the best but windows requires a working vcache service driver and most setups have it broken because amd cant make fucking chipset drivers. cache is for dual ccd cpus and just throws it on the cache ccd. frequency and soon highest frequency are supposed to replace them all.https://www.tweaktown.com/news/111469/amds-new-cppc-highest-frequency-feature-could-improve-cpu-scheduling-and-boost-behavior-in-windows-11/index.html
>>108887238Super Ryzen 11950X3D II Turbo
>>108887183Everything provides better performance than windows 11. Even if you don't look at the numbers the thing itself runs like absolute shit.
>>108887183Reminder that pixDAIZ was a Winfag.But it's never too late to change. You may in fact be a little penguin fetus in an egg that needs to be cracked open.
>>1088872389950 is the modelX3D means the L3 cache is bigger and physically located in a slightly different place (hence 3D)2 presumably means it's a newer version of X3D
>>108887681Yeah for that use case, *buntu would certainly be a hueg upgrade over *doze.>24 * 115.10 / 294.83 ~= 9.3695So with 9950x3d2, Noobuntu only needs 9h 22m 10s to encode what would take 24h on Windoze.
>>108887183fake and gayleenox is not >2.5 times faster than win11, under no circumstances, no way.
>>108888579have you tried right clicking on your desktop
>>108888904Holy fucking shit..
>convert video to av1 on cachy>it’s slow as shit>convert video to av1 on winblows>it’s also slow as shitit’s just av1 that sucks
>>1088872389950X3D2 is the last desktop AMD CPU
>>108887183This has been true for the last 2-3 years.
>>108887238they're going to AI 9 590X3Dsomething like that
>>108888168>what's even going on there?Windows CPU scheduler moving threads between CCDs?
>>108888168AMD is at home on Linux since they're more pro open-source.
>>1088883892 means literally two as in there are two X3D dies instead of one
>>108887183It's probably something really stupid like different compiler flags ruining the comparison
>>108891556oh, the L3 cache is a twice as big?
>>108887183But I use Win 10 LTSC
>>108887183That's cool, but has the X3D2 provided substantive general performance improvements over X3D? It is a halo chip, but I don't need to calculate fluid dynamics or optimize Borderlands 4.>>108888579yes and no, the X3D2 needs software and firmware that utilize it, this finding is only stating that a bleeding edge Linux kernel can run it well in a specific circumstance, eventually Windows should do so as well.>>108887280One of the most popular distros and the perk of having Canonical behind it, is that it often has the most bleeding edge hardware enabled for it, if Canonical is interested in it; see their experimental ARM image of Steam for bleeding edge benchmarks.
>>108887183so does my toaster.
>>108887183microshit can't possibly compete on desktop with its current workforce
>>108888267Interesting, thanks
