>hire semi-competent engineers>release good enough CPUs>wait for Intel to ruin itself>win
>>101552354>do your job instead of chasing DEI metrics
>>101552354>purposefully gimp GPUs in order to maintain a monopoly with your cousin
>>101552409Nobody complains when the chinks do it, but if they were JEWS, oh boy...
>>101552409 AMD drivers are open source can you show the exact line of code that does that? or are you just going to run away after being asked to elaborate on your brand tribalism faggotry
>>101552354but why are they going the cuck core route
>>101552471>Nobody complains when the chinks do it, but if they were JEWS, oh boy...naw, /g/ hates them equally.
>>101553201The "C" cores AMD is using are quite a bit different than the "E" core approach Intel has. Intel's E cores are a different design than their P cores, and do not support all the same instructions. That of course can cause all sorts of scheduling headaches and inconsistent performance. AMD's C cores on the other hand are the same basic design as the performance cores, and support all the same instructions - AMD basically just squished all the transistors and such closer together, letting them cram more cores on the same die space, while using less power. The only downside is that these C cores can't clock as high, but for mobile and many server uses that trade off is worth it.
>>101552354>hire semi-competent engineers>release good enough CPUs>wait for Intel to ruin itself>still manages to lose somehowKinda impressive, really.
>>101552477i have no fucking clue what point you think youre making
>>101553608AMD already screwed themselves with their cuck cores. Only one gen to go from 35% size reduction to 25%. Mobile Zen5 also uses cucked dual pumped 256b FPU instead of the 512b FPU like desktop.
>>101553727He is claiming that because AMD has pretty good open source drivers, AMD could not be "gimping" their cards as part of a collusion scheme with Nvidia. While that does of course rule out software gimping, it would not cover claims of AMD gimping cards by simply designing worse hardware on purpose. That said, I personally think there really isn't collusion here; Nvidia has simply been spending a a lot more on GPU R&D and related software than AMD for years now, and Nvidia's management is competent enough to make sure that money is not being wasted. So its no wonder that Nvidia has better graphics cards, no collusion needed.
>>101554036>Nvidia has simply been spending a a lot more on GPU R&D and related software than AMD for years nowYeah and thats not a coincidence.AMD is allowed to have the CPU market and the low end gayming GPU market.Nvidia keeps the high end GPU (and importantly) the entire datacenter/AI marketThere is no rational explanation for why AMD would abandon the largest and fastest growing market for computer hardware ever other than some sort of arrangement like this
>>101554036>>101554146and secondly, the driver being open source does absolutely nothing to rule out software gimping. Let me know when you finish parsing hundreds of millions of lines of low level driver code and identify all the optimization deficiencies FOR FREE
>>101552354she learned the ways of gabe. sit by do nothing and watch the opposition shoot themself in the foot
>>101553976I wouldn't say they "screwed themselves", it more like they did a really good job of cramming more transistors into the same space on the regular Zen 5 cores (the 8 core desktop chiplet has 28% more transistors than Zen 4 chiplet, while having the same sized die on only a moderately enhanced process, and maintaining clockspeeds), so there was simply less low hanging fruit to be found when they squishing stuff down for the C version. As for the mobile chips and AVX512, that is a separate issue from whether or not they used C cores; the 4 performance cores on the upcoming Strix Point are also cut down in the same manner. I am guessing AMD decided that extra performance for AVX512 applications was worth the die space and power draw on laptops. That said, we can also throw some blame at Microsoft and their whole Copilot+, with MS pushing CPU vendors hard to include larger NPU units, which of course is eating into die space. Of course, AMD (and every other chip maker right now) should get some blame too for following through on all this AI marketing stuff.
>>101554269They absolutely screwed themselves, because datacenters don't want a core that is 25% smaller with mostly the same instructions. They want a core that is 70%+ smaller with massive increases in PPA and efficiency. AMD's failure with A1100 was what directly led to every hyperscaler making their own ARM server CPUs. AMD might have taken a bit of Intel's server marketshare for the short term, but the whole market will get destroyed by ARM within a decade. By the time AMD realizes their cuck cores won't cut it and try to develop a smaller core, it'll be way too late.
>>101554511>the whole market will get destroyed by ARM within a decade"People" have been saying this for more than 15 years now.Better luck next year. And on the next. And the next...
>>101555504ARM won years ago in mobile due to Intel's failures.ARM will win again in servers due to AMD's failures.
>>101555597>Google server CPUThey can't even get phone CPUs right and have to recycle the Exynos 2100.
>>101555658A pitiful final cope from the AMJeet.
>>101552354lol
>>101552409Who removed them?
You don't even need competent engineers at this stage.A psychotic monkey on hallucinogenic drugs could defeat intel at this stage.
>>101555859>delayed for 2 weeksoh noooooAMD is collecting more good samples to give intel refugees the best reception
>>101555859>s-s-surely t-there's hope kike-tell bros...
>>101552354what actually happened is that intel collapsed from wokeshit and amd just had to do nothing but keep course to succeed.
>>101554511Zen4c has 100% instruction parity with Zen4.Zen5c has 100% instruction parity with Zen5, the only caveat here is that mobile APUs are distinct for the sake of maintaining low power consumption, which is also why RDNA3 discrete cards had doubled ALUs but the RDNA3 IGP bearing APUs did not.Your posts are worthless because you're ignorant of basic information. You typed all that for nothing.
>>101556060>muh instruction paritythe customers that might have cared about your cuck cores before don't care anymore
>>101556276>okay I was literally 100% wrong>but look at this irrelevant imagelollmaoyou stupid autistic retard loserThe EPYC SKU in these tests isn't showing Zen4c. Genoa is not Bergamo. Literally nothing about this bench is relevant.
>>101556322SAR, please calm down. It's okay if Zen(c)uck cores are a flop. Hyperscalers will still redeem your non-cucked cores.
>>101556376Poop skin, lower your tone when you're talking to a white man.
>>101556421Sar, please do not redeem the cuck cores
>>101556453>amazon that has to recoup the investment on their own internally designed arch>isn't impressed in the chip that you are too stupid and cowardly to show benchmarks of>everyone is a pajeet but me!!!!!LOLAmazon good employer SAAAAAAAR, they paying me 3 rupee to post on four channels!!!! INDIA SUPAR POWAR SAAAR
>>101556478SAR, please redeem AMD
>>101553976>new SoChaha, they threw the towel and copied Apple M series.
>>101552409So they are only cousins after they've been removed? Who removed them?
>>101552354>lisa su>semicompetent engineersyou guys have no idea who she is, do you?
>101558159getting paid per (You), huh?
>>101552354
>>101559220
>>101559232
>>101559240
>>101559220>>101559232>>101559240>>101559247why do intcels cope by showing mistakes other companies make?they fucked up, they know and they are not fixing anything
>>101555859what’s the point of rushing it? They’ve been in the lead for four years and intel chips are literally rusting and burning to death
>>101552354Isn’t the cpu architecture merc they hired to design Zen now working for Intel?
>>101562790He left years ago and is working on his own startup tenstorrentHis contribution to intel is the royal core that wont be released for a couple more years (~2 gens away)
>>101562819huh, it looks like his 2nd chip is already shipping color me surprised
>>101552354>implying Israel Inside was ever in the business of computingLook, Intel is audibly more about geopolitics and spyware than anything their front door implies.
>>101555863>>101558261bot or ESL.
>>101563994Or a joke, dumbass
https://www.anandtech.com/show/21489/amd-delays-ryzen-9000-launch-1-to-2-weeks-due-to-chip-quality-issuesDuring final checks, we found the initial production units that were shipped to our channel partners did not meet our full quality expectations.AYYMDEAD IS FINISHED & BANKRUPT
>>101554198>driver being open source does absolutely nothing to rule out software gimpingso your argument is that they would do gpu gimping without a single line of code just on silicon? so they increase failure rate just to build instructions on silicon just to scratch gpu performance a little? LMAO LOOK AT THIS NIGGER TOURIST PLAYING TRIBALISM HAHAHAHAHAHAHAHA
>>101552354>Be intel>Demolish AMD in the late 2000s and early 2010s>Continue to not increase core counts for almost 9 years>AMD finally releases a cpu that isn't nearly as fast in single core but destroyes intel when it comes to core counts>Conveniently start increasing core counts right after that>Fail to catch up in multi core and lose the server and workstration market>AMD eventually overtakes you in single core as wellAll because of greed
>>101555859Intel shat the bed so bad Amd can throttle the release schedule and I am not happy that this means also postponing 9800x3d since the only competition for it are own older x3d models.
>>101562819Keller had nothing to do with CPU arch at intel. He was working on SoC integration and fabrics. He left the company before the project he was working on was even finished.
