What would you guys use for video editing, rendering, docker containers, vmware, etc? Computer will not be used for gaming.
Why are you posting a pic unrel?You're a retarded gambler go for rust and fire inside
>>106833620So that you could comment on it and bump it. Now which one is better?
>>106833614Still wouldn't trust 13th/14th gen Intel to refrain from committing sudoku. So I'd only consider 15th+, or 12th.> vmwareBig/little tends to be a headache with VMs, so 9950x.
>>106833614The one with AVX-512 and real cores
>intelaviv in current year
>>106833614285K for that
>>106833656I thought it was a motherboard/cooling issue...
>>106833660Which one is that one
>>106834389These aren't AMD Piledrivers, you can't just shove 1.5V+ into them and expect them to last. They saw that AMD did it with Zen 3 and thought they could do the same. Spoiler alert: high boost zen 3s degrade too.
>>106833614The Ultra 9 285K is the best productivity CPU, I believe. If not the best, It's at least the better value over the 9950X.
>>106833614>video editing, renderinggpu>docker containers, vmware, etcused epyc hardware on ebay
AM6 when
>>1068344109950x, dont invest money in a dead platform. am5 is still going to get zen6 with 24 real/performance cores.
>>106835032What about the i9-14900k?
>>106834389combo, bad microcode + bad motherboard settings + clock tree silicon weakness and possibly oxidation issue for some chips, if you run them at alderlake speeds 5.0-5.2ghz, risk is minimal/non-existant, But I wouldn't recommend running them higher than 1.35v, stock ,max vid for the 14900k is 1.55v , intel has warrantied them for 3 years, so they should last at least that long , but gone are the days of assuming a 10 year lifespan.
>>106835032arrowlake and raptor lake will age poorly, because they lack avx-512, more and more software is taking advantage of it and in those zen5 absolutely demolishes everything else. and zen5 is close enough to the 285k in "weak" areas , that its a no brainer, and its on a living socket, so you can expect at least 1 or 2 more upgrades. arrow-lake is too expensive frankly, if it were cheaper , then I think it would look alot more favorable in comparison.
>>106833614literally why would you ever go with 13/14th gen when ultra 9 is better in every way?
>>1068336149950x has avx512 so it wins by default
>>106835640More so, 9950x has "efficient" avx512, where the increased performance of avx512 does not increase power consumption.People dont know this small but very important fact. AVX512 is extremely efficient for games where you get additional 20-50% performance if you use them vs avx2
>>106835768>AMD's AVX512 is extremely efficient Just wanted to make this clear
>>106835627ultra 9 is slower in gaymes
>>106835768That's the 7950xAvx512 in 9950x absolutely does have a massive power draw issue but it's thankfully not as bad as it could be because of tsmc's node carrying it
>>106835852https://www.tomshardware.com/pc-components/cpus/amds-zen-5-avx-512-performance-tested-zen-5-performs-significantly-better-than-zen-4-on-linux-without-consuming-any-more-powerNope. >9950x~56% AVX512 improvement (vs off)148 w off152 w on>7950x~27% AVX512 improvement (vs off)169 w off172 w on9950x is 2X more power efficient/powerful in AVX512 vs 7950x in AVX512 and its more power efficient and draws less power than 7950X.You're ABSOLUTELY WRONG.
>>106835908Avx512 on zen5 causes massive downclocking though >Sauce: owned zen4 and 5
>>106835768>>106835852>>106835908I don't care about power draw. I just want to know which processor is best for rendering videos and stuff like davinci, dockers, and virtual machines.
>>1068360799950x wins on average especially after all the mitigations released by intel massively gimping clocks from what it used to be 12-18 months agoThe only real benefit going with intel now is quicksync which is a nothingburger since rendering is better done on the CPU for quality or on the GPU for speed.
>>106836030Thats another unsubtantiated claim.The fact of the matter is, AVX 512 gives you +56% performance increase with 1% power load difference. Thats extremely good. Downclock is a meaningless metric. The real metric is how fast your workload gets done, how much fps you get, how fast your video gets rendered, etc. 9950X is a monster with AVX512.
>>106836115Downclock is meaningful since you get less clock cycles to run avx2, SSE and other non avx512 instructions. That was the biggest issue with skylake and Icelake back in the early days of avx512
>>106836127Again. Wrong. The biggest issue with Intel's 512 implementation was that it would double the power load and downclock at the same time, while giving some performance boost. That made the performance metric meaningless because you're just eating up double the power wattage load. You have bad memory and you dont understand what you dont understand.
>>106836146All of your arguments are theoretical perfect world cases where the avx512 load completely fills up the pipeline in every cycleYou don't seem to understand the real world implications of downclocking at all
Definitely not the self-incinerating CPU
>>106836480intel's old implementation was just bad, amd's is completely different. doesnt run into the same problem, on old intel chips just running a few avx512 instructions would trigger the avx offset bringing the clock of the whole chip way down. amd does this purely by power and thermal limits, and each core can change its clock individually so a couple instructions here or there wont do much, newer intel chips are the same, its a non issue these days, the only time the clock will throttle hard is under heavy workloads, which are rare due the fact the zen 5 is bandwidth starved, and if that is the case, int performance isnt your bottleneck anyway. https://www.numberworld.org/blogs/2024_8_7_zen5_avx512_teardown/ zen 5's avx 512 is stupidly fast can provide a 100% speed up under ideal conditions, more than worth the clock trade off.
>>106835068>dont invest money in a dead platformI bought into X99 and 5820k back in the day because I bought into the "upgrade path" meme. Turns out nothing was released that justified replacing the 5820k, so I was stuck with the lowest end chip for like 7 years until I upgraded to Rocket Lake. My plan now is to get the top of the line DDR5 setup once DDR6 is released, so I can enjoy the flagship experience while benefitting from the "last-gen" discount. I have zero regrets buying my used and heavily discounted RX 6950XT even though it's on "dead" RDNA2. For the same money current gen, best I could do is RX 9060.
>>106834410maybe i5-11400
>>106833614I am waiting to upgrade to a 10950x to turn my pc into a server
>>106835768>AVX512 is extremely efficient for games where you get additional 20-50% performance if you use them vs avx2Which games support AVX512, besides RPCS3?
>>106840814Well a superset of RPCS3 includes hundreds of games
>>106835438>gone are the days of assuming a 10 year lifespanDo you know the last generation that has that? And is it safe to assume all generations prior are the same?>>106835466>Zen 5 is close enough to the 285K in "weak" areas, that it's a no-brainerDo you know if that applies to software Intel has been the victor in? Like Adobe?
>>106833614I’m really happy with my 14900K (after it was replaced for rusting). Performs well and consumes a proper amount of electricity at each percentile.
>>106841519Smaller nodes inherently mean shorter lived silicon.>The candle that burns......
>>106841519>Do you know the last generation that has that? And is it safe to assume all generations prior are the same?largely depends on the product, if you're buying a low-midrange chip thats locked, and runs at <4.5ghz, you can expect very long lifespans still since they are going to be running at pretty low voltage/power/temp, but when it comes to the highest end performing products, the first product from intel that sacrifices lifespan in a big way, maybe the 11900k. I think intel learned from its mistake the hard way with raptor lake, so I would assume a better lifespan from arrowlake and newer. I believe that amd still targets a 10-15 year lifespan in the stock config, the silicon FITness algorithm they use trys to keep to that afaik, however once you start using the pbo scalar , your guess is as good as mine.>Zen 5 is close enough to the 285K in "weak" areas, that it's a no-brainerDo you know if that applies to software Intel has been the victor in? Like Adobe?https://www.pugetsystems.com/labs/articles/amd-ryzen-threadripper-9000-content-creation-review/looks like its pretty close for abobe, intel no longer holds the same lead for such software that they once did, there might be some exceptions, but generally I would expect <10% difference in the worst case scenario.
>>106842889>once you start using the PBO scaler, your guess is as good as mineSo for maximum longevity on Ryzen systems, disable PBO upon the first bootup?
>>106843047You can enable pbo, you just want to make sure that the pbo scalar is set to 1x, I think by default enabling pbo sets it to 2x. enabling pbo with the 1x scalar still gives you higher power limits and the curve optimizer.
