It's up.https://youtu.be/RDpBuiu0ql8
>>108667281>slower than 9800X3D in many vidyawtf were they thinking? With the twin CCDs it should at least have been to emulate an unbothered 9800X3D by reserving an entire CCD to the vidya game and putting the shitty bloatware and gaming-unrelated interrupts on the other
>>108667359I also wonder if Threat Interactive will still keep believing Steve is a marketing slave after this
>>108667281Between 1.6 TO 5.1% GAINS WOW amaze
>>108667281its a direct upgrade from the 9950x3d for those with money to burnseems fine to me but i'll stick to 1-ccx cpus for the foreseeable future
>>108667359That's literally what the 9950X3D did, you retarded faggot. It was the entire point of the CPU. AMD repeatedly told people that adding v-cache to the second CCD would not improve gayming performance due to cross-CCD latency penalties, yet people kept begging for it anyway. Now they've released such a CPU and everybody's claiming they never wanted it and it's pointless. Gaymer scum must be rounded up, held in camps and taken in groups to be gassed until none remain.
>>108667359>0.01% wants dual X3D>Give 0.01% dual X3D at 0.01% pricesIs it that difficult to understand?
>>108667271They also neutered ECC memory speeds to try and upsell the X3D lol
>>108667364>Threat InteractiveWhy do you care what that obnoxious retarded faggot believes in lmao
>>108667383It only just removes the hassle of fiddling around with core affinities on OS that can't get into thread scheduling. $200 extra premum for that? No thanks. This SKU is really just a niche production unit for selected use case that benefit from the cache and threads but cannot justify or have budget for fully blown Threadripper/Epyc platform.
>>108667482WTF are you talking about? ECC memory adheres strictly to JEDEC-spec (No XMP/EXPO). Any overclocking has to be done maunally by the end user if their platform supports it. Overclocking defeats the entire reason for ECC memory.
>>108667531That hassle never existed in the first place, you can configure the preferred CCD in the BIOS lmao
>>108667271
>>108667510Because threat interactive and gamers nexus both are pro consumers. One focuses on software/dev incompetence the other on hardware/corporate corruption. They are both useful
>>108667601Poorly coded OS overrides it anyway. This is why 9950X3D does better in *nix.
>>108667271says he got blacklisted by amd. what a liar.
>>108667646This is not true, I have a 7950X3D and when I bought it I tested all possible configs for game (Factorio specifically) performance: AMD's Windows-side automatic driver thing, the CPPC preferred cores option in the BIOS and also manual affinity setting on the game process. All of them basically performed the same, within regular testing variance of each other, so the BIOS option works fine. I also tried setting it to frequency priority and it totally nuked performance, so clearly the OS obeys the setting.
https://videocardz.com/pixel/amd-deploys-war-room-to-monitor-our-website-for-leaksImagine actually buying from this garbage company
Did Intel continue to ignore Bar Chart Guy over the 13th/14th gen overheat issues? I would've considering what an autistic idiot he was about it. My replacement 14900K has been fine.
>>108667359Any variation if "what were they thinking" just makes me think of pic rel
>>108667682he got his review chippy from another YouTube slopper
>>108667359just don't buy it thenyou people have been told time and time again that even more cache means nothing in latency bound scenarios when it has to jump across to the other ccd
>>108667954>1 metric out of 40 is betterwaste of silicon
>>108667682Lienus, despite being the biggest advertisement billboard for hire on YT, didn't get one either. Except Steve went insane over his obsession with how Peter Thiel fucked kids with Epstein and he thinks Lisa Su has personally crossed him off the list for daring to point out she's connected to Thiel because Theil works with Jensen and Jensen works with Palantir and Palantir works with Thiel or some other reach to convince his viewers he has a good reason to spew all of this shit because it's somehow connected to consumer hardware for gaming.Crazy to think I'll live to see GN become the same laughingstock that LTT is.
>>108669319GN is a based schizo, LTT is a faggot race mixing cuck
>>108669327He promised he would go schizo on GNCA, yet he doesn't use GNCA since he knows how much profit he makes by shitting up the main channel with it. Then he has the gall to talk down viewers who are expecting him do his job of reviewing tech, where CPU coolers and cases for review are still abundant.Steve Burke is a spineless cunt that has forfeited any and all integrity he ever boasted about, just in the last 4-5 months worth of uploads.
>>108669319t. Seething israeli
>>108669343fair enough, can't excuse not doing more reviews, but on the GNCA side he's made hour long documentaries, and one dropped not even one month ago
>>108669348>anyone that dares to criticize Tech Jesus™ is an Israeli/techbro/buzzword!I've been following GN ever since they were running everything out of Steve's mom's house. I know that the average fa/g/ot has the mental capacity that's less of a basic singular transistor, but there are many long-time GN viewers that see through Steve's bullshit, and remember more than just the last video he released.He promised to keep all his consumer advocacy content on GNCA and keep the main channel to tech reviews. He barely reviewed any hardware in the recent months and instead flooded it with content he promised to go on GNCA.On GNCA he boasted about joining the class-action lawsuit on Honey. He was the only one to back out of it and never mention it happening again in hopes people will forget he was full of shit and was the only one in the CAL to not have a leg to stand on.Also he made over a million dollars on hyping up his GPU black market documentary that was insanely underwhelming compared to what he was hyping it up as.To me it's clear that Steve has abandoned his years old principles in the chase of short term monetary gain out of reactionary content.>>108669385Before I was done with his shit I was subbed to both GN and GNCA. A whole fuckton of videos that would fit GNCA went onto the main channel, so even if I unsubbed from GNCA my feed would be clogged up with Steve's obsession with Peter Thiel, and I'd have no way to filter it.Also, I keep pointing this out because I wanted to see it. FLP02 review. He promised it months ago. He said it's in works months ago. And I need to find the comment where he talked down a viewer with something along the lines of "sure, we'll review a case where you can't put anything in it!".That was the final straw for me with this idiot. His job is to review hardware, the main channel doing that is his business model. He flew too close to the sun with his ego.
>>108669423>GN exposes Palantir and co.>suddenly posts like this appearHope the 10 shillings is worth it
>>108669447Cool, please tell Steve to update his channel branding, since it doesn't say anything about being a martyr for truth and exposing the tech industry. Since apparently that's what GN is about now, and anyone wanting for him to go back to what he used to be is on Palantir's payroll.
>>108667271Did steve already apologize for his idiotic conspiracy that AMD would have blacklisted him, while not even shilling channels got it?I think for the sake of the consumers... the consumers of his channel... this would be important to clarify, right?
>>108669484>Everyone that doesn't support Steve's transition into being a white monkey chicom dramatuber works for palantirCalm xer tits.
>>108669514Oh by the way, around the time of the last GN vs LTT drama, where Louis Rossman stepped in to talk sense into Steve, he released a lengthy video explaining ethics and principles of his company and so on. He inexplicably deleted it but I have a copy of it someone screen recorded. Steve made too many fuck-ups to treat his whole "independent investigative journalist with integrity" persona seriously. And I keep bitching about him because it's a fucking shame for what happened to him, he used to be a good creator.>>108669546>suddenly plays the opposite sideHoly fuck you suck ass at this.
>>108667874>amd defense force instantly arriveskek
>>108667636>both pro consumer>They are both usefulTI is a raging narcissist. He wants his ego stroked at any cost. You're actually retarded if you think he wouldn't do a 180 the instant he thought he'd get sucked off harder for changing his views.>>108667271On topic, I don't understand why anyone is particularly surprised by this. Games like cache, and generally don't benefit from more than 6-8 cores. The production workloads like cores, but may not benefit from cache (unless it's truly gargantuan amounts).The cross CCD latency and, by extension, the latency to pull from main memory is what kills AMD at the moment. The reason why intel keep sup reasonably well is because you can get much faster main memory. That offsets the need for such a big cache, and if you have a workload that can't fit in the cache in the first place, the gap will start closing.>>108667779That's why games like factorio were much better on AMD until you got to large enough scales. In 1.1, something like a 10k science per minute base would favor AMD by enormous margins, but if you cloned that based to 50k spm intel would start to catch up. Or more accurately, AMD would experience more rapid performance degradation because you couldn't run DDR4 at more than 3800 consistently on AMD, while intel could run at 6000. It's a similar story with DDR5, albeit less pronounced, and space age bases run into other issues.Funny side note on factorio. Train bases performed slightly less worse on intel, probably again due to higher system memory specs.>cont.
>>108669670>contAnyways, if they can crank the shit out of the infinity fabric for zen 6, things will be significantly better. If they can have a unified cahe block for both CCDs, that will be huge. Might not help games that much due to most games still being heavily single thread bound, but there's a number of production workloads that'd love that because it'd mitigate having to load stuff into both caches or move it back and forth.TL;DR the chip is very very good, but it's not worth it unless you need/want the absolute top of both production and gaming in one chip. It's kind of neat that that's a thing that actually exists, but for the overwhelming majority of people it's just a waste of money. That's what I expected, and I'm genuinely confused that anyone expected otherwise. I'd like to see how it chews on factorio because that game's a bit of an outlier, but I'm certainly not going to replace my 9950x.
>>108669447Steve is the one and only person on this world who criticizes Palantir and who is fighting against them by saying things on a big terch video sharing platform! He is such a danger to the elite, that Palantir invests billions into shiling against him on 4chan!
>>108669670>>108669678In short, X3D2 was kinda pointless on Zen 5, AMD knew it was kinda pointless, people wanted it anyways, they delivered, it turned out to be pointless, people are mad. Wait for Zen 6 where doubling the X3D cache will bring some actual benefits.
The Palantir shills in this thread just make me resist even harder!I just bought a gaming dice, with a cat picture on it, from the GamersNexus store!Together we will defeat the big corporations, we just need to give Steve more money and he will make things happen!
>>108669776>willI wouldn't say will. It might. I'm not entirely convinced that it will. There's limitations with the physical architecture at this point. We might not (read: probably won't) see enormous gains while we're on the AM5 socket.Zen 6 will probably be a moderate performance bump, maybe allowing for more than 600-6400 memory speeds, or maybe some tweaks that make running with the different ratios and memory in the mid 9000 to 10000 range perform notably better. Right now 9k memory is barely stable, and it barely performs better for some workloads. Fix a few timing issues with that, and make it stable and that'll be a moderate performance bump without needing significant IPC or clockrate boosts.
>>108667636>threat interactive
>>108667271Buy an ad Steve.
>>108667954>4x faster than 9950X3D which in itself is slower than 9950XI see moronix is at it again
>>108669423People calling him "Tech Jesus" really got to him huh? There is nothing of integrity left, just doom and gloom drama ragebait.
>>108670236Yeah the parasocial trend of calling creators with long hair and a beard "Jesuses" like calling Ian McCollum "Gun Jesus" or Brandon Herrera "AK Jesus" really needs to stop. Though it's a good way to get yourself some sycophants, grow your hair out, grow a beard, now you're "Jesus".Which is blasphemy btw
>>108669670>On topic, I don't understand why anyone is particularly surprised by this.I'm not surprised by the charts. I'm also not surprised that benchmarkers don't "get" it. They do one thing, measure it and declare success or failure. That's easy for them, but that's not how ALL workers work. The days I'm hitting GO on a long running task and walking away are rare. More often than not I'm starting a compile and getting on with some other task, or encoding a video and setting up a VM or fucking around in blender and chrome back and forth and this is where having two full cache CCDs will likely help.
>>108669319nice try, jensen/lisa you fucking slimy snake, go back to the hole you slithered out
>>108670116Can you explain what i am supposed to be outraged about here?Like this pull request is very obviously not LLM written , he just changes some parameters after autistic testing, which we know he does.
>>108667383>>108667461>>108667954Then they should have X3Dmaxxed the already X3D-equipped CCD by sandwiching it in two X3D dies despite thermal concerns or stacking X3D dies, this way the CCD that actually matters for gaming can spend even more time doing actual stuff instead of waiting for memory, there's some performance to gain with even more cache as even on X3D CPUs some performance is gained by tuning/overclocking RAM.>>108667902Yeah me too
>>108669790So true, slava Rex Donaldvs I!
you fand destroy any argument steve burke tries to make simply by pointing out his greasy, unkempt appearance and flabby mantits.
>>108667617does he always swear in his gay ass videos?
>>108671716when it's necessary yes
>>108669343> Steve Burke is a spineless cuntShut the fuck up Linus!
It doesn't matter what people think.People will still flock to go pick this up simply because its the top chip in the am5 lineup. Just look at the hoards if reddit complaints from people who bought the 9950x3d because they thought the chip with the biggest number designation would be the best for streaming fortnite.This has always been the case and the cpu manufacturers have always leaned into it for both the consumer market and the business market. There will always be more fools to fleece with whatever high end halo product they can shit out of the lithograph. The general consumer doesn't know the first thing about compute architecture or processes, so whatever is on the top of the stack of the top shelf will always sell like hotcakes. Amd could literally sell a wet turd and as long as it has the top designation it will sell out.
>>108667636Throat Interactive is a gigantic faggot bitchboy.
>>108667271I still want the 9950X3D2. I just don't care about real world use, or gaming. I want my CPU to have all the bells and whistles.
>>108671626Found the Vaginus fan. You prefer hairless leaf twinks?
I don't get why fags are so upset about the price. It's only 10% higher than the 7950X on release.
>>108669579>taking ethics lessons from a narcissist like Linus Sebastian.How dumb are you?
>>108673871I thought Steve was all about transparency? Why delete an over hour long explanation of how you operate just because you got talked into doing so by a two-faced conman? Then again, backing out of that Honey lawsuit and memory holing it without making an official explanation or an apology is more than telling that Steve's no better than Lienus.
>>108673662nobody is going to buy a chip 80% more expensive for 5% more performance
How much do you get paid per eceleb thread, jannies?
>>108676177100 rupees
>>108675550it's the best CPU on the market, so people will buy it
>>108674285he took donations for that lawsuit and everybody who gave him money for the lawsuit got scammer :^)But consumer protection is not that important if its about the consumers of GN
>>108674285I suspect that the honey question lead Steve to a jewish maffia. Everything around this reeks of it.
>>108675550>5% more performanceLooking at the charts there are many areas it falls behind in performance. It's not worth it in any scenario.
>>108676423I mean the extra cache would help your computer feel more snappy I presume. I feel that it would be an excellent upgrade to my 7950X.
>>108676418Everyone else who was in the class-action lawsuit didn't dip out. The only party that did was GN. Likely because they would get steamrolled in it due to not having shit to bring up as it was just a publicity stunt. If Steve made a genuine mistake of biting off more than he could chew then he would've made an official statement about it, but instead he went radio silent about it in hopes people'll forget.Well I have a bone to pick with this caveman and I have a very good memory.
>>108675550Pot met kettleThis is exactly the mentality that drives the modern PC culture.But I guess since it's only 50% more costly for 5% more performance (even less outside of 1080p) it's permissible.
>>108676514Others did drop out as well. But GN did it with prejudice, while none of the others did.So basically, Steve formally signed a paper stating:>i drop out because i was wrong and don't have a casemeanwhile others did a:>i drop out because there are already a dozen other parties in that lawsuit and they don't need me for thisIt is an important difference, because it means that Steve ceased the right to sue them again. He can never fill that lawsuit again. It also harms the people who stayed in the lawsuit. You have a legal formal statement stating that there is no case, signed by one of the suing parties.Also Steve said later that he dropped out because he wants to report on the case, and if he is a party, he needs to stay silent about it.... he never reported about it ever again.
>>108676441>$479 to get a 9850x3d which gives you a 70% increase in performance>an additional $420 to get a 9950x3d2 which gives you an additional 7% in performance
>>108669343Okay, but realistically, how many people actually care about reviews of cases or CPU coolers?
>>108675550>nobody is going to buy a chip 80% more expensive for 5% more performanceyou'd be surprised how many retarded shopping addicts there are
>>108677149NTA but I would but I understand why Steve doesn't.The entire PC market discourse is reddit tier obsession with optimalism so there ends up being little room for anything that isn't the best or cost effective and everyone has already made up their minds.I will probably never, ever get a proper review of a case of I want because it isn't a mid-tower. 10 years ago I would have gotten multiple for a giant autism box.
