I want to belive, but we have to face facts. Yields are still too low, the pricing was too agressive and only having the top 1% cpus be good enough to have the 3dcache implant means this will never improve.Im calling it, its a paper launch. it was too good to be true.Prove me wrong.
>>103240379It's actually what the 7800x3d should've been, but they throttled and gimped it because they wanted to sell the 7900x3d.
>>103240379Don't forget the few that are actually out there have been burning up
>>103240441>why yes i love spreading misinformation for no reason
>8% faster 18 months later is too good to be trueok retard
>>103240379You must not be America.
>Prove me wronghttps://www.amazon.com/Best-Sellers-Computer-CPU-Processors/zgbs/pc/229189#2 after the affordable 5700XIt's selling
>>103240448>it's okay when AMD does it
I am challenging the "it" part of this comment.
>>103240441Proof?
>>103240441From what we have seen so far that has been user error.
>>103240379Yep, it looks like a near-paper launch. The 5800X3D and 7800X3D never had such issues at their debuts despite having nearly peerless gayming performance at the time of launch. >>103240477It is the 8700K and 9900K all over again. There weren't nearly enough units in channel to supply demand.
>>103240379>Yields are still too low> the top 1% cpus be goodsource?>>103240441fake news>>103240784I think it took a while for word to spread about how legendary the 7800x3d is in performance and efficiency - it beat everything that came after, 13900, 14900, 9xxx, 285x - so now the wider masses joined in and want the 9800x3d. There's no proof that it's a paper launch, it sold a lot but amd didn't expect this demand. The 7800x3d didn't sell a lot initially and the price fell far below msrp.
>>103240379>just built a 9800x3D system for a gamer>Most of the games I tested run well>thinking i might buy a chip for myself>boot up Rust and test it>performance is absolute dogshit>wtf is going on?>reload to new servers over and over>same behavior>1%'s crash for ~35seconds before they recover>move a significant distance across the map>crash for another 35 secs>test my tuned 14900k in same servers>crash for 7-8seconds before recoverySo basically when new texture data needs to be loaded in the cache it takes 35 seconds for the AMD cpu and only about 7-8 seconds for the intel cpu necause its a monolithic die with way lower latency in the memory subsystem.if you play normie faggot games like Forspoken and concord the 9800x3d is the best chip ever but if you play any AA game, indie game, Early access game etc. Your performance is an illusion and when the game goes outside the cache for data fetching your performance will shit the bet and you will get horrible stutters.
>>103241665enjoy your instability
>>103241677>admitting that you dont know how to tune your coresIntel instability is only from core boosting which does not improve your performance at all especially in games. Its a scam designed to make the artificial benchmark numbers higher to sell newer chips to retards like you. Disable boost and lock your cores and instability disappears and your performance in games improves. Overclock a 14900K to about 5.7-6.0ghz all core, Lock cores, Tune your memory at 7600mhz-8200mhz with tight timings. Run Karhu memtest for stability check and your doneHow are you on /g/ and you do not know this you dumbfuck?
>>103241722>>103241665>Exact same talking points as Frame Chasers with the same gameJufes pls, nobody wants to buy your $500 tuning course to pay off your mortgage, you bald faggot.
>>103241665both systems seem fucked. ask other rust players with 9800x3d about their experience.>>103241722all that work and it's still slower than amd
>>103240784>never had such issues at their debuts despite having nearly peerless gayming performance at the time of launchintelfags werent switching over to the new x3d en masse because of intel's own fuckups back then
>>103240379lol stocks are replenished daily, my buddy literally just bought one off neweggintel plebs just have to cope AMD has a hot product meanwhile core ultra is snot
>>103240379>paper launch Wait for the 5090 maybe the whole 5000 serie, the stocks will be abysmal. No way they will let a consumer product compete with pro cards and take precious production slots.
>>103241665Retard-take, it is not cache that is problem. It is when the program needs to feed into main memory. The infinity fabric is the bottleneck here. Intel chips always struggle if you give it fast enough memory or endure the same issues with their first go at chiplets ie Arrow Lake.
>>103246067There will be plenty of sub-optimal Blackwell yields, the big versions of silicon with no flaws cannot be easy to produce in vast quantities. You know that 4090 were sub-optimal yields of AD102?
