9850X3D Revealedhttps://www.youtube.com/watch?v=proW65mqs8I>UPGRADE & BUILD ADVICE.Post build list or current specs including MONITOR: https://pcpartpicker.com/Provide specific use cases.State BUDGET and COUNTRY or you will NOT be helped.Building guide: https://wiki.installgentoo.com/index.php/Build_a_PC>CASEmATX: AP201, Lian Li A3, O11 Air Mini, XT M3, CH260ATX: XT PRO (ULTRA), AIR 903 Base/MAX, Lancool 217, Flux Pro, Y40, Meshify 3, 4000D FRAMEDual Chamber: Y60/70, Vision (Compact), Antec C8AVOID: NXZT, 'Silent' cases, fanless cases, Corsair 6500>CPUBudget: 9600X, 7600X, 7500FGaming: AMD X3DWorkstation: 9950X, 9900XAVOID Shittel>COOLERAIO: Best Value: Thermalright Frozen Edge, Premium: Lian Li GA II Lite, LCD: TRYX PanoramaDouble towers: Thermalright Phantom Spirit 120 SE, Noctua NH-D15 G2>MOTHERBOARDAvoid ASRock AM5 boards, CPU degradation issues still reported.>RAMGaming: DDR5 2x16GB 6000CL30Workstation/high end: consider 2x32GB>SSD (OS drive)Budget: NM790Mid range: SN7100Premium: SN850XFlagship: SN8100>GPU1080p: RX 9060 XT 16GB, Poorfag Esports Gaming: Used GPU between $100-$2501440p: RX 9070 XT, RX 9070, Budget: RX 9060 XT 16GB4K:Used RTX 4090, RX 9070 XTWorkstation: Radeon AI PRO R9700, CUDA Mandatory: Used RTX 4090 or RTX 3090 tiAVOID RTX 50 Series: False advertising "5070=4090 performance!!", unsafe power connector, broken drivers, overpriced compared to RDNA 4 in most regions, threatening reviewers, manipulating and faking reviews, evil anti-consumer company etc.ALWAYS pair AMD GPUs with AMD CPUs for SAM https://imgur.com/a/6XSNJ9b>PSUhttps://hwbusters.com/best_picks/best-atxv3-pcie5-ready-psus-picks-hardware-busters/>MONITORhttps://www.rtings.com/monitor>CASE FANSMeta: Case with good stock PWM fansMid range: Arctic P12/P14 (Max) (5-pack)High end: Noctua NF-A12x25, Noctua NF-A14x25 G2Prev: >>107777717
so, the 64GB kit I got for $299 last year is now $1500. Thank you comfyUI for convincing me to buy more memory when it was cheaper.
No point in buying the 9850X3D btw. 2.7% average uplift vs the 9800X3D and it'll cost 40% more.
nicotroony thread
poopdickschizo thread
>>107778785SHILL THREAD. AVOID
Has anyone tested dlss 4.5 on a 50 series card, is it any faster?
>>107778812You mean padoru nero? Well, I guess so, fgo is very coomery
>>107778813what about the productive site?
>>107778849Exactly the same difference but at that point you'd either go Intel or the 9950X3D(2 if it exists at all)
>>107778849You would probably want more cores for that, and pricing will likely be close to the 9900x3d with 4 less cores.Hell, for productivity, you can get a 9950x for about the same price, and that's more performant for a ton of workloads.
So mad that he created a second thread just to be sure.>>107778839
>>107778878buying amd cards can make you mentally ill
>>107778823STFU PISSBUTTFAG
>>107778872then wtf is the point of this cpu?this thing is going to cost like $150 more than the 9800x3d currently with only 3% better performance? lol
I want a hipster burger restaurant computer tower with wooden accents and incandescent light bulbs instead of RGB LEDs, /g/How do i assemble a hipster burger restaurant computer tower
>>107778982you'll want to look into getting this for your mobo then
>>107778972For the extreme autist overclockers with higher bins. Basically AMD's version of Intel's KS cpus.
Haven't built a pc in 10+ years. Is articsilver5 thermal paste still the best these days or are people now using the fancy new thermal pads?
>>107779018>>107778785
Which one is the real thread?
I wonder how lower res ultrawides really feel in person. I'm used to a 32" 3840x2160p display, though I sometimes see displays with resolutions like 3440x1440p at around 34". I feel like the worse PPI would maybe not feel so great, but I imagine the aspect ratio and added real estate are still pretty immersive. Would it just feel like stretched 2560x1440p or something? I have no fucking idea how people run the massive 5120x1440p ones at decent frame rates even with a 5090, fucking hell.
>>107779023That's old shit. The cool nu-nu is using PCM-type pads (such as the well-known PTM7950) since they're not gonna give you crazy pump-out/drying issues and require more frequent maintenance like pastes do. They're not even that expensive; I'd say it's worth it. I have it on both my CPU and GPU.
do we have any news on them making new 5800x3d? any new rumour or anything? I need to build a new PC until july and am4 x3d would be the best choice
Im too scared to build myself is this a decent deal for a prebuilt guyshttps://www.newegg.ca/gaming-desktop-pc-amd-radeon-rx-9070-xt-amd-ryzen-7-7800x3d-32gb-ddr5-2tb-ssd-azzacarmel-7800x3d9070xt/p/N82E16883989016?item=N82E16883989016
This got announced today at CESThoughts?
>>107779033Both threads are made in unreal engine
>>107779023arctic mx-7 if u still >>107779023wanna go the paste route, ptm7950 pads otherwise.
>>107779085My response is mostly confusion. I don't see the point whatsoever in such high PPI on an IPS display that small. Comp game players will want higher refresh rates and lower latency/ghosting (so basically OLED). It hits an $800 price point which is less than the more premium 5k offerings, but people who do want 5k probably don't want a small IPS panel. Brightness is nothing special unless compared to OLEDs either. There is almost no 5k content, and it'd take a lot to run anything at peak frame rates anyway (so basically 5090), and anyone who has the kind of money for a premium GPU and related hardware isn't gonna go for this lesser visual experience. It's just weird to me overall in that it doesn't seem to suit really anyone.
>>1077790855K, huh.
Please recommend me a reasonable upgrade for my 3060TI (8gb) paired with a Ryzen 5800X so it's not bottle-necked.I only do light gaming, but could use some more VRAM for blender.Would be nice if it didn't cost an arm and a leg, but I'm sure I'll have to buy it soon or I wont get the chance...Out of my pc build I regret this CPU the most. Even undervolted it's a pain to cool T_T
>>107779131maybe its a gaymer and productivity hybrid
>>107779034i had a 38" 3840x1600 it was great. i don't personally see the point in 34" 21:9, 32" 16:9 is more surface area anyway. I think that the height of 32" 16:9 is the absolute limit, any taller is just annoying (unless you move it back in which case why bother) so the only way to grow is with a 38"+ UW. unfortunately they seem to have mostly given up on making 38", at least in oled.
>>1077791539070 xt and below
Which 5070 Ti model should I get? They're the same price
The best gaming cpu actually already existed.
>>107779182>AM5
>>107779175pny since im sure you're not using an itx pc that gigabyte's designed for
>>107779175>4 in stock>25+ in stock
>>107779173>9070 xtBut I need my cuda...
>>107779192>>107779196The Gigabyte has been selling like hotcakes though, was 25+ in stock just yesterday. Makes me think the buyers know something I don't....
>>107779198RTX 5090 or below
are there any ram purchasing hacks? just hope for the best on ebay? due to uninvited guests over xmas i have been forced into buying a full system and really don't want to pay a premium to buy the exact same ddr4 shit i had 5 years ago. i briefly toyed with looking at sodimm + chinese converters but it doesn't seem that much cheaper anyway. the prices for even a dogshit set of 2x8 to tide me over for a few months are obscene.
>>1077791985070 ti and below
>>107778785Bros my fucking LG C5 has to come from another town that's 3h drive and I don't even know if they are gonna carry it properly (not have it flat down) Wtf should i do?
>>107779219>really don't want to pay a premium to buy the exact same ddr4 shit i had 5 years agoShould've boughted 6 months ago then
>>107779226im going to knock your tv down like dominoes
>>107778785Think I'll stick with my 9800X3D
https://www.nvidia.com/en-us/geforce/news/dlss-4-5-geforce-game-ready-driver/https://www.nvidia.com/en-us/drivers/details/260405/NVIDIA WON
>>107779246guy oughta enhance his haircut wtf is that combover
>>107779246Wow, i'd kill myself if I bought a 9070 XT especially if it was a shit one like the Asus Prime
intel used to sell us 4 cores for years. now amd is selling us 8 cores for years
>>107779219>just hope for the best on ebay?You won't beat the bots camping the listings on that website 24/7.I'm a 16 gb ddr4 chud, and my plan is to just tough it out and hope this shit improves in a couple of years.>>107779257>wtf is that comboverProbably cost him $900
>>107779219Most of the chinese mini PCs haven't shot up yet. You're stuck with a laptop processor and Oculink, which is only PCIe x4, for the GPU though.
>>107779198If you go Nvidia you practically have to buy a 5090 or you are getting functionally anal raped for no reason
Intel selling Nova Lake with 16 P-cores and 288MB cache next yearAYYMDEAD stagnant and dead with 8 cores
Why are all the new 5th gen QD-OLEDs that got announced today ultra wide curved slop? Are they really not coming out with normal 27inch 1440p one till later?
>>107779168I get that in theory, but it doesn't seem to possess anything I'd identify as "productivity-focused". Doesn't have the best color coverage or proper HDR support for color work/art shit. Doesn't have ultrawide sides like for video editor scrubbing or space for modeling software shit, music production software, etc. The secondary mode could work out, but people can get higher refresh rates on panels that cost much less if they're gonna emphasize that aspect. Just odd.>>107779172Thanks for your feedback. 3840x1600 sounds like a decent compromise. Too bad all the ones I see on PCPP are some form of IPS for some reason; I don't see a single one in OLED either. Guess it's back to eyeballing 34" 2160p displays I guess...
>>107779192Just diff a bit more research and it seems that since SFF cards are smaller they also reach higher temps?In that case I'll get the PNY since it should have better temps.
>>107779085My 4090 is hardly good enough for 4k.
>>107779280Blackpill me on the event, anything new worth getting?
>>107779296smaller cooler yea.
>60hzaaaand dropped.
>>107779175The one that is in stock.
>>107779315>>107779323Thanks anons, I'll go with the PNY.
>>107779318How many times did your crackwhore mother drop you on your head when you were a baby? It's a pro monitor for graphic design work not a gaming monitor fucktard.
>>107779280all based on the same new Samsung panel
>>107779318I don't think you need 480hz for Photoshop or Lightroom anon
>>107779337what if I wanna scroll through a long wiki page and have it look nice?
>>107779337refresh rate not divisible by 24 is a problem for people that do actual professional visual work
>>107779279just 16 cores? that's cute.
>>107779310Nope, just AI this AI that, and some AMD and Intel processors nobody should care about.
>>107779352all I want is 120hz is that too much to ask for in the current year?
>>1077793186k, IPS is not a gamer display anyway.
Why are these retards making 1000hz monitorsThe human eye can't even detect that. Past 240hz everything looks the same.
>>107779369you realize one of the biggest complaints of pic related is that its 60hz right?All Acer had to do was make their monitor 120hz...
>>10777936116 P-cores, 32 e-cores, 4 lp e-cores for 52 cores of AYYMDEAD killing power
>>107779397Solely as a gimmickLike with basically every other aspect of your computer, monitors only have a handful of companies fueling the basic technology and everyone else is just slapping their name on shit, so to stand out as one of many you have to pull random useless shit out of your ass, like overdriving your panel to dogshit levels so you can advertise 1000hz
>>107779259KEK
I have Redditlands 4. When I get the chance I will try to test the difference between DLSS4 and 4.5 with regards to some of its thin optics reticles which does have a significant ghosting problem with the current 4.0. Also for an older game MW5 Mercs whenever there is a thin dark object over light background there's a chance it gets affected by severe FSR-like ghosting as well, so another thing that I can test and compare.
>>107779408I'm sure you'll get such a big cinebench score with all those e-cores.
https://github.com/NVIDIA/DLSS/releasesNVIDIA WON
>>107778808Can't even get a discount memory ddr4 is up 250%
you are getting the new gaming shades, right /g/?
https://www.youtube.com/watch?v=mbIIPM7Lsrg
Can't even get a CPU, a mobo, ram, a case or an nvem SSD, too broke and unemployed for this shit
>>107779496How am I supposed to use these if I already need and use normal glasses And no, lasic eye surgery is not a valid answer
>>107779504That'll be $1500 thanks
>>107779259Got em
It looks like ddr3 is still cheap
>>107779496>>107779504so i can basically use my wall as a 171" display, huh. nice
9850X3D or wait the AM6?
>>107779528if its 400 bucks or less id buy it
>>107779496>>1077795041080p garbage btw. That resolution looks pixelated and blurry even on a monitor a bit away from you. You'll be able to count all of the the pixels with it right up against your eyes like this
>>107779543faggot detected.
>>107779496>>107779504This is literally the future. Game anywhere using glasses, no hardware needed, all compute done on the cloud with you just paying a monthly fee.Can't fucking wait!
lmao
>>107779539>PCIe5 GPU on 6-core low entry CPU on a PCIe4 mobo because just fuck my shit up bottleneck my ass into oblivionYou love wasting money with missmatched hardware already why wait for AM6
>>107779562Nicotroon, care to explain this shit?
>>107779542Not likely, xreal's lineup are pretty much all much higher than that, and that's before the ROG and Bose markup
>>107779566isn't the 9070 xt 16x pcie thoughit should be fine on pcie 4.0
>>107779587That's true
>>107779587bro, your 5% FPS loss?
>>1077795399950X3D with 16 P-cores!
Are PNY cards really that bad? Does card quality vary much among 5070tis and 5080s?
>>107779604+slow CPU more like 15%
windforce or ventus for 5070ti? both are supposedly ass but no other options here sadly
>>107779621Not the Ventus.
>>107779539bless yourself
2 years since I got my 4070 and I'm still amazed at how cool and silent it staysBest 530€ ever spend (thanks old job for that 200€ Christmas credit card, I miss you)
A DLSS version that upscales from 720p to 4K but looks as good as native may not be that far away after all. that's crazy
>>107779539wait for Zen6 X3D
>>107778785>>AVOID RTX 50 Series: False advertising "5070=4090 performance!!", unsafe power connector, broken drivers, overpriced compared to RDNA 4 in most regions, threatening reviewers, manipulating and faking reviews, evil anti-consumer company etcKeep seething AMDtard
>>107779652Personally I prefer my games stretched from 480p to 8k, PS5 type shit, but to each their own
>>107779318>219 ppi200% or 225% scale?
>>1077790854k/5k for desktop and 1440p for gayming dual mode monitors seem like something I'd be interested in in the future, but I'd prefer it if the faggots just made proper 16:10 screens again.
>>107779522Contacts? They might be big enough if you got a smaller pair of glasses that can just sort of fit over them. I've done that with sunglasses before.
So is DLSS the only vaguely gaming related thing that got announced so far?
>>107779854a bunch of gaming monitors got announced
>>107778785is 100mb of cache a lot?what is the upper limit to cache size, until it's no longer even useful?
>>107779894more cache more better
Will AMD finally understand that they cannot compete with Nvdia and be happy like Intel in their lane and properly price their hardware? RX9070XT only makes sense at $400 now
>>107780030What do you mean? The 9070XT will be $700 and everyone will happily pay it because they were told it was amazing value for money when it was priced at $500.
>>107778813pict. recent 9800X3D enjoyer
>>107780030correct. redstone will flop, it really didn't have any confidence in the first place.meanwhile dlss 4.5 makes 5060 ti a pure 4k graphics card.
AMD GPU sisters... don't look...
>https://youtu.be/u4cfzFd9HzU?si=L85l9dJlRbPVxqAJ>DLSS 4.5 OUT VS NO UPGRADE OF FSRAMDsissies your response?? KEKKERS
>>107780070QRD?
the saar are literally on overdrive today
>>107780086>QRD?qrd AMDalit lost
285k bros...
>>107780070>DLSS is now a major selling point in the longevity of buying a GPU.>So much so Nvidia is limiting it to specific models of cards to encourage people to upgrade.>1080Ti is still relevant due to technology like this, and 20-Series cards are going to be relevant for years to come.>Tech reviews still won't touch it because 'muh fake frames'.
>>107780086DLSS is now even better than nativeNative sissies seething hard at DLSS chads who get more FPS and higher quality image than the developer even intended for FREEgun and shoot gamers that think frame gen is bad got btfo bigly also
>>107780086Anyone who bought a 9070XT should just try to trade it for a 5070ti asap
>>107780105>3-6% uplift for the same priceDOA
>>107780111how do I get it
>>107780111you fags been saying that the whole time and it has yet to be true
There's some random shit-stain on my new CPU. I don't have any isopropanol to try to wipe it off. Do I buy that? Or should I try qtip + water? Or do I just put it in? Fucking annoying.
fake frames, fake cores, fake resolution and you know what else is fake? yeah the msrp. enjoy your slop.
>another shill vs shill thread
>>107780105The only point of the 9850X3D is on paper higher chance of a better bin to UV/OC and maybe less chance to fry up in an assrock mobo with a new bios release specific for this new CPU? I mean it's ok for people looking to upgrade if price isn't much higher than base 9800x3d
>>107780136Yeah, you're going to need ISO for that. Do not use water. Alcohol wipes will probably be okay. At most, if you're feeling daring, you can try and spit-wipe it off with a paper towel. But, I would strongly suggest not getting moisture near your CPU.
The argument that PC building is dead is absurd.Yes ram is expensive right now, Yes it is hard to buy but its not because theres no RAM left. AI data centers didnt all just pop up in the span of one month to gouge the prices. So that tells me its just a meme FOMO right now. There's no actual real shortage of supply, just FOMO and market shifting because christmas so every parent either bought a prebuilt.Yes it sucks, Yes it costs $150 more than what it cost a year ago. But if you can't afford $150 more on a build that you were going to spend $1500 on should you really be building a PC, sounds like you need to save the money.
https://www.youtube.com/watch?v=4MnI1x35H1Q>2 autistic German guys making G-Sync even betterWhy doesn't Nvidia hire jeets?
>>107780105People buying a 285k have uses other than gaymes in mind, it doing it well is a bonus
>half life alyx runs at 4320 x 2160 pixels @ 90fps>high quality assets everywere>no shitty ai for anti aliasing>no taa>no blurry visuals>looks and runs better than 99% of all games on the market right now>uses msaa on top of thatyeah everyone who shills for dlss, fsr and ai frames is a smooth brain faggot nigger
>>107779617People who buy them say they're finePeople who don't say they're trashSo who's to say really.
>>107780169PC building is dead if you consider the only acceptable way to game is to have 64GB+ of 7000MTs CL30 RAM.Those kits are like $1K+, whereas they used to be about $300 before.But, 32GB is really fine and X3D makes speed and latency kinda moot outside a few percentages in edge cases.It sucks, but you can upgrade your RAM later when this is hopefully over, and you honestly would not notice the difference in the meantime. Now, if you miss the boat on GPUs, it is unironically fucking over for you.
>>107779522Gaming lenses.
>>107780169Hey, buddy. Get this.If I wait™, Sam "Wincest" Altman might realize he buy too many RAM that he can't use and sell them to us at below market price.Waitfags always win.
>>107779617they are the oem for nvidias professional cards.
>>107780228forgot the linkhttps://www.pny.com/professional/hardware/nvidia-professional-products
>>107780227that's sadly more likely going to happen with the GPUs companies like Microslop bought and are unable to power, and not with the unfinished RAM wafers the Gigajew put in storage
>>107780227your home machine can't use the same ram
>>107779894A normal 8-core Ryzen 9000 chip has 8MB L2 (1MB per core) + 32MB L3 (shared) = 40MB. 9800x3d / 9850x3d have triple the L3, and the 8700G APU has only 1/2 the L3 cause it needs the room for graphics.Hard to say at what cache size you'd see no further benefit in practice.
>>107780272Aren't those just normal DDR5 RAM with ECC?
>>107778885Imagine actually wanting AI upscale slop.
>>107780208I'm waiting but I dont think this is the end of PC gaming. This is just another FOMO scamdemic because theres still a massive demand for GPU parts as the amount of zoomers and gen alpha grow and the amount of genX/boomers die off.Covid made GPU's unobtanium even 750ti's were selling for $150+ on ebay.I remember when the 6700XT came out it cost 1500+. Then new after the scamdemic it came down to 500 new. Now it costs 250-300 used on ebay.Waitfags will win and I'm betting that by the time the next 9800XTX or whatever the newest flagship AMD GPU is, it will all be normalized.
Just remembered that the Nvindia jeets do this for free
Still waiting on an upgrade for my 480hz qhd oled.360hz uw is almost good enough but i really want 4k res. >>1077790855k is the future lil bro
found a new lora i like and i've been using my 5080 to generate yuki_(crunnchy) doing lewd and cute poses and also various sexual situations for approximately 2.5 hours now and i've nutted countless times throughout. This technology will be the end of our species. But more importantly what are the implications for the pc building market?
are open builds the new hip thing to do?
>>107780379Not really. Cases are very good these days, and so having contained airflow and heat sinks are popular. Basically, if you put a big heat sink on something, and then use the case to blow a load of air over the heat sink, it's way quieter and allows it to operate "Passively" in general use.
>CPU: Superconducting Flux Core (Reversible Josephson Array 128k-core, 100 THz equivalent)>Motherboard: Photonic-Superconductor Hybrid Backplane (Hollow-core optical traces, cryo-integrated power delivery)>RAM/Storage: Unified Universal MRAM — 4 PB addressable non-volatile memory (Exabytes/sec internal bus)>GPU/Optical Engine: Holographic Light-Field Renderer (16K × 9K per eye @ 10,000 Hz, spectral-locus laser primaries)>SSD: Crystalline Optical Archive (petabyte-scale, photon-addressable, zero-latency access)>PSU/Cryo: Active Cryogenic Power Module (closed-loop liquid helium refrigeration + superconducting power delivery)>CPU Cooler: Cryostat with Quantum Refrigeration Loop (sub-4 K continuous cooling, vibration-damped)>Case: Vacuum-sealed Faraday Cryo Chassis (EM shielding, hollow-core fiber ports, acoustic isolation)>Mouse: Interferometric Laser Tracking + Optical-Analog Actuation (nanosecond latency, interrupt-driven)>Keyboard: Optical-Analog Switch Array (photonically sensed actuation, no mechanical debounce)>Display: 32" Holographic Light Field Panel — 15360×8640 (16K), 10,000 Hz, 0.01 ms response, 0–10,000 nits local peak, 100% spectral-locus gamut, 16-bit per channel color depth>Audio: Wave-Field Synthesis Array with Plasma/Graphene Electrostatic Drivers (room-scale holographic sound)>Network: Hollow-Core Fiber Interface to Local IXP (latency = distance / c, quantum-encrypted photonic link)>Interconnects: Photon-native PCIe-equivalent Bus (64 Tb/s per lane, optical backplane)>OS/Runtime: Reversible-logic aware microkernel with photonic instruction dispatch (zero-copy memory model)>Accessories: Vacuum optical docking station, cryo-safe UPS, vibration isolation platformFuture Anon here. R8 my build /g/
>>107780451mid
>>107780451>no quantum entangled peripherals for sub-instantaneous inputsngmi
amds desktop kit. good or bad idea? for a diy steam machine.
>already confirmed DLSS 4.5 is worse than DLSS 4 and barely a visual improvement at a much higher compute costlol
Do PNY GPUs really explode or is that just Reddit MUD?
>making up nonsenselol
>>107780501case by case basisanyone that ever tells you a generalized 'fact' about a brand is not someone worth listening to
>>107780501i ones had a gtx 780 and it broke. i then switched to a rx 480 which later also broke. now i am on a b580. nvidia and amd only make trash gpus that suck.
>>107780501no, here's some useful data. It is worth remembering PNY is who nvidia uses to make their all-important enterprise cards, so they are obviously competent.
>>107780170so only the acer model is 4k?the wait continues
I can't tell the difference between preset k and m my fps is the same too. Wonder if I'm doing anything wrong or if my eyes are filtering me
So did waitfags get BTFO or not?
>>107780476Bad. PCIE slot is bandwith limited and steam machine has a better cpu.
>>107780528Sapphirebros....ZombiEVGAfriends...not like this.............
>>107780476>>107780768So disabled iGPU and PCI 3.0 x8? I am not sure about the specs of these kits. I think the CPU is around a 4000-series and will bottleneck any modern GPU anyway, even a 9060XT might be limited. But depending on your goal, something like a cheap second hand 6700XT or similar might be a good fit
>>107780086Everyone who bought a 9070xt got scammed because they were LUSTING for 200$ discount kekypow
So Nvidia announced dlss4.5, and showed off zen6 epyc and intel announced mobile panther lakeAnything else of note today?
>>107780797They didn't get scammed. They got exactly what was promised, nothing more nothing less.
>>107780797I don't feel scammed, the card was a big upgrade from my 4070, especially for unoptimized older games and Stalker GAMMA. Just brute forcing its way through, its a delight
>>107778982Fractal
Did they already announce new OLED monitors? There's literally not a single good 4k OLED and I'm not a jeet going for QHD.
>>107779260For gayming its more than enough
>>107779507You're too late to the party
>>107779507Spent less on gacha then you dumb gooner
>>107779621Windforce. Ventus is dog shit chinesium
>>107780086Amd stocks will go down like 5/10%
someone post the beat sam altman picI need to vent my frustration
>>107780129Get an Nvidia card newer than rtx 2060
>>107780288No. Hbm is completely different
>>107779360good thing refresh rates aren't fixed in the year of the lord 2026. you can just run it at 48Hz if that's so important for you.
>>107780476Depends on cpu and pci e lane. Also if it's ddr3 don't even bother. If Ddr4 better put 32-64 gig minimum
>>107778901>Report and ignore this troll threadAll PCBG threads are troll threads and full of branding flame wars you cunts are fucking retarded
>>107780528No idea if this is still the case as I haven't had to RMA a product in 15+ years (Amex chargeback meta). However, back in the day PNY was the only partner whose RMA centre was actually in Europe and you didn't have to pay out of pocket to ship the card off to chinkland.
>>107780757Yes. Anally. The prices of current cards will only go up now with the dlss 4.5 and no supers coming
>>107778785amazon finally delivered the ssd i panic boughted on BF
>>107781024Based hold on to it for 5-6 months and you can sell it for double the price
>>107780943a super refresh would have been a paper launch anyway because GDDR7 isn't availableI am guessing RDNA5 will be delayed by 6 months to a year, I bet Sony is sweating right now because they were likely banking on launching the PS6 in November.makes me wonder if they'll do a run of PS6 mainboards with slots for sodimm.
>>107780105>counter strike 2legit who gives a shit, this game runs on a potatoe>t. 30 yo boomer
>>107781057Ps6 won't come till next year. Even steam machine who has all the Sodimm for the first batch is hesitating on launching officially
>>107781092yeah and anything under 200fps feels like dogshit in this game. not because 200fps is bad but because 500fps is much better. counter strike 2 is also going to get updates over the years which are going to cause frames to drop. it's one of the most played games on steam and here to stay. like all the other games after a few years nobody is going to play them anymore appart from r6 myb. so who gives a shit about those.
>>107781129>it's one of the most played games on steam and here to stayand how many of these play on a 600$ cpu lmaoif the graph showed actual numbers instead of relative gains, we could have a good laugh
>>107781095yeah they're not getting modules for the next 6 months, but if they've already ordered them by now then they will get some in Q3 but they should have already been pumping them out all year so Q3 will be late.which makes me think they could do a sodimm model which would just require dimms to be placed in at Q4 to meet bulk demand.this is all assuming that AI companies don't buy up all the ram again in 2026
I asked AI (grok) and it said Its time to upgrade. I will listen to our overlord
https://rog.asus.com/articles/gaming-monitors/rgb-stripe-pixel-oled-tech-debuts-in-the-rog-swift-pg34wcdn-xg34wcdms-pg27ucwm/Any oledfags out there? Should I wait for the pg27ucwm if I was just about to buy a pg27ucdm? Allegedly this new model should launch in q1, possibly Feb for around the same price.
>>107781149you can check yourselfhttps://prosettings.net/lists/cs2/https://steamcommunity.com/sharedfiles/filedetails/?id=3240880604&searchtext=fps
>>107780451>implying human eye can see more than 5000FPSkek
>>107781192in future we will be able to put nvidia chips directly in our eyes to improve graphic fidelity
>>107781151>this is all assuming that AI companies don't buy up all the ram again in 2026the OpenAI are running until 2028, aren't they?
>>107781203you can forego your natural eyes for a cyborg implant like in cp and it will have full spectrum vision, 1000x zoom, thermal vision for use in heavy snowstorms, nightsight, all for a low cost of $49.99/month (and an always on mossad 7g modem and remote self destruct feature)
>>107780451>current year>not using a psychic interface cyberdeck to jack directly into the mainframe
>>107780451Future anon hereDon't fall for his boasting, this guy is a legit poor fag in the future
>>107781178https://press.asus.com/news/press-releases/rog-rgb-oled-ces-2026/>The ROG Swift OLED PG27UCWM is a 26.5-inch Tandem RGB Stripe OLED monitor with TrueBlack Glossy™ coating.vs PG27UCDM QD-OLED matte
>>107780170He just turned off motion blur.
>>107781033It could sell for more than that later on as a collector's item from before the Sandisk product rebranding.
>>107778808could have bought 2x64gb 6000c42 for 390€ in stock and shipping from amazon before stock ran out and 4xcould have bought 128gb in summer with dual 2x32 for 32x2 6400c32 for 100€ a kitcould have boughted 6600c32 2x48 titaniums for 460€ shipping some time summer (and two more kits like this)Collection of could haves
>>107778813you already knew this clock speed is not the lowest hanging fruit on poozen you didn't buy the course
So how much better will the 9950x3d be compared to 7800x3d? Do we have any cute graphs yet?
>>107781381>9950x3dthe graphs for that are a year old now
>>107781381>>107781383meant the new 9850x3d desu
>>1077813819850X3D is 3-6% uplift from the 9800X3D
>>107781192>Human.There's your problem, meatbag.
>>107781393I asked about the 7800x3d, chud
>>107781305Guess I should wait, especially because they're doing refreshes of ucdm models anyway.
>>107781406yes, and you can extrapolate the uplift from the 7800X3D to 9800X3D
>>107781417So no cute graphs yet?
>>107779280PG32UCDM3 has Near RGBstripe-V stripe subpixels>>107779310xg27ucwm for tandem oled (yes not woled, no white subpixel) rgb stripe tandem subpixels OR pg32ucdm3 for v stripe qdoled subpixel>>107779822200 for 3k res if u want perfect scaling bruh>>107780169>But if you can't afford $150 more64gb went up from 90$ to 900$ m'am>>107780837I just listed two up top>>107781178yeah you should wait(tm)even 140ppi qdoled vstripe is sharper than 166ppi old qdoledthat is if you get eye strain or not
After opting into the beta and getting DLSS 4.5 Steam FINALLY decided to give me the survey popup box. Guess it was right to wait this long since I only got my OLED recently; maybe the algorithm tries to determine if you are at "risk" of updating your hardware in the near future and waits for it (which implies it regularly scans for your hardware).That's another 9800x3d / 5070 Ti to the results tally.
>>107781438>5070 Ti to the results tallyMy man
>>107781438actually it detects if you have amd and then it does an additional roll of 1 to 10 and only show you the survey if you get a 10
the human eye can only see $3500
>>107781421this is the only chart of the 9850x3d in the world thats out
>>107781438>he want to be part of the npc registry>he want to be part of the botnetngmi
>>107780050oh get outta here with that naked SpongeBob , you goofy goober!!!
>>107781464fuck you>t. 7900xtx owner
How to make moneys with rtx 5090???
the new dlss presets look like oversharpened garbage...just like dlss 3 presets. they went backwards. why?
>>10777881310800x3d doko
>>107781514its 4.5 so its better than 4.0
>>107781487stay mad, stay average, stay registered. i’ll be here in the "0/0 results" tier. literally "the golden rabbit" while you’re just "the lab rat". my gpu is the rarest of them all. i am the "0/0 results" anomaly you are peak botnet tier, a literal headcount for the steam surveillance survey.
>>107781514you need new eyes, now there is 0 ghosting on motion and it fix all the issues around volumetrics and particles, M preset is supposed to be used with performance DLSS and L is for ultra performance, if you are using anything above that you should use the updated pre-set K
can I use dlss4.5 on my 3080ti?
>>107781546yeah
>>107781541preset K was updated?
>>107781546yes
>>107781554nobody knows
>>107781554I don't think it was because it doesn't look oversharpened like preset L and M do.
>>107781554Yes, its much better at handling ghosting, shimering and particles
>>107781552>>107781555>>107781556apparently 4.5 actually tanks your performance on a 3000 card
>>107781601Not my problem? Works on my 5080 just fine
>>107781601both preset L and M are quite heavier as they are meant to use alongside perrformance and ultra performance (L being heavier) K should be just as performant as dlss 4.0 while being better at dealing certain issues
>>107781601the previous gen nerfing to sell rtx6000 has already begun...
>>107781601literally no performance difference on my 5090 thoughever
>>107781614so what preset do I use on a 3000? K? I really don't get the difference to using the ingame DLSS option in cyberpunk
>>107781609tranny gpu with the bone structure of a 4070
>>107781512ebay
sometimes i think i fap too much
Preset L and M are the only ones with the updated dlss overlay information, therefore it only makes sense to assume they are the only ones that support dlss 4.5 and everything else is still dlss 4.
New DLSS solves the issues with volumetrics, foliage and ghosting that previous models hadhttps://imgsli.com/NDM4OTA5But the performance hit is quite noticeable on my 3060, although for the first time ever 1080p/DLSS performance doesn't give me eye cancer
>>107781625More like a 2070 or 1070 really. The 4070 is quite a bit smaller relative to its flagship than the 5080 is. Smaller die and much smaller memory bus too. But you're nicotroon, you're too retarded to think.
>>107781644all it took was for dlss to not have a performance improvement over native, gee, thanks nvidia
>>107781679I don't use DLAA, but on previous models, DLSS quality/balanced was much much better than Native with TAA enabled, TAA is absolutely horrible man
I played Cyberpunk 2077 for two hours this morning to test out the new DLSS Model. The improvements are so significant that I've decided I'm going to be using frame generation going forward if DLSS Quality isn't enough to get me to 116fps with my 5090. With 4.0 my biggest gripe was with the UI artifacts and ghosting, it was awful -- while 4.5 hasn't fully resolved this issue, it has been mitigated at least 95%. As far as I can tell, the only artifacting remaining comes from stuff on the edges of the screen. For example, if a wall is lined with lights, there will be a very brief continuous light streak that will persist (very briefly) since it assumes (I think) the lights will continue. Besides the frame gen improvements, I haven't noticed any major improvements to image quality with the transformer model. This might be because I play at 3840x2160 (where DLSS excels at), so fine detail issues have never really been a thing for me. Anyways, pretty solid update, because I was a huge frame-gen hater prior to this.
>>107781713>--this post was literally written by chatgptwere being astroturfed by nvidia marketing agents
>>107781728always have been
>>107781713>get me to 116fps with my 5090
Okay im getting a spare 5070tiPrices vary a lot though, anything I need to avoid?
what case do you guys all use?
>>107781746its so blatant that its obvious they arent even trying to hide it anymore
slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within slop within
>>107781781Musetex k2
the TLDR is use Preset M for everything (it's the new default for a reason) and maybe preset L if you're using ultra performancePreset K does not have the new image quality or stability improvements
>>107781431I dislike QHD and wide monitors so I'll just get the PG27UCWM. I can't get anything better in Q1 right? I want the best 4k OLED at 27 or 32 inch
>>107781764I have an older LG OLED 4K Panel that caps out at 120hz. Newer Panels are 144hz, but I'm not fucking with EDID again to get my matrix working on both my panels. >>107781728Fuck off retard, chat bots uses em dashes not double hyphens. ---Anyways, you can see that there is some artifcating in the quest log, specifically with the text "Jackie." It must have stabilized really fast though, because I did not notice or perceive at all in motion.
>>107781850I too am going to get it even though I have the pg27ucdm already. The back of it looks really nice too.With how close I play I don't think I will like the 32 inch. Will probably never switch to the dual mode but cool to have.
Did I fuck up not buying a pristine RTX 3090 for $670 last month?
is over https://www.youtube.com/watch?v=9rnjawJ_QDc&t=122s
>>107781630I'm an obsessive perfectionist and yet I never actually went all the way to get that drawing tablet or to incorporate advanced workflows for inpainting, controlnets, and so forth on my xx60 shitbox
DLSS 4.5 working all the way back to the rtx 2000 series effectively kils ayyymd.
>>107782020Yes. That's a really good price
>>107781844'm using K with new drivers and DLSS and I'm seeing improvements in image quality and no ghosting of particle effects.
>>107782036We are the future. Vertically integrated ai-powered sexuality. Sam Altman's dream come true.Going from a 3060 to a 5080 was horribly unhealthy for me, because now i can make an image in like 4 seconds instead of having to wait like half a minute. The degeneracy accelerates.
>>107781852still blurry shit i see
>>107781631https://forums.guru3d.com/threads/nvidia-geforce-591-74-whql-driver.458947/page-7Multiple people can confirm preset K is also using an updated transformer model that fix ghosting and motion clarity at no performance cost
>>107782079placebothere are no changes to preset k
>>107782079theres a huge performance cost for old cardshttps://forums.guru3d.com/threads/nvidia-geforce-591-74-whql-driver.458947/post-6387595
people are killing their gpus with sag brackets cause it just makes the pcb twist... even if you use them right you only delay the inevitable vs unsupported gpu. The only sane orientation is vertical. Buy a riser or simply place your computer case on its side if you have a big modern (and expensive) gpu. Not worth doing any less.
>>107780111what is the point of fake frames? you are trading slower real graphics for faster slop
>>107782120the gpus sags because people install their gpu wronghttps://youtu.be/liChy76nQJ4?t=599
>>107782112No they tested specific cases which compared pattern grid, ghosting and volumetrics, dissoclusion is much better on preset k 4.5 than preset k on 4.0
>>107782186say it with me nowplacebo
>>107782186they only added two presets, everything else is the same.
>>107782223>>107782210cope
The amount of misinformation in the pc community after an update drops is wild
Fuck this, I'm just going to buy some Sennheiser HD 560S for $160 while they're still that cheap.
>>107782243It's literally on the nvidia notes > Preset K: Default preset for DLAA/Balanced/Quality modes. Less expensive performance wise compared to Preset L.K is stll the default preset for anything above performance, both M and L are specific for those presets and K still has the new transformer model improvements
>>107782265Correct, and to add to this, none of the new presets work with DLSS-RR which is still using v310.4.0, so if you use ray reconstruction in a game, you are not using DLSS 4.5.
https://www.overclock.net/threads/increasing-rdna3-rdna4-desktop-class-power-limits-and-adding-vid-offsets.1816083/?post_id=29473161&nested_view=1#post-29473161based
>>107782066I probably should have saved a ton of time by making a jump to a 40 or 50 series GPU or something with a lot of VRAM. I keep thinking nah it's not that important, but I've been prompting all sorts of things since 2023.
L and M are the only ones using the new model but maybe they also also retrained the older ones.
>>107781601Well, yeah, it uses your card's hardware.It's only going to work on cards and games where you've got headroom to use the tensor cores to generate frames.The reason why stuff is locked to the 50-series is not (Entirely) to motivate people to buy new cards.When rendering frames, rasterising is done by the CUDA cores, and then ray tracing is done by Tensor cores. Ray tracing is taxing, but with high-end cards, frame rendering is far more often bottlenecked by CUDA cores rather than tensor cores. This means you have the headroom to generate extra frames without a performance hit.With 30XX and 40XX cards, there's fewer tensor cores, and they're occupied by ray tracing, so there's less headroom to generate frames, and in the case of 30XX, that frame generation negatively impacts ray tracing and bottlenecks frame rendering. It's not just a software layer, it uses the actual hardware of the GPU.
>>107782289tl;dr
>>107782298You need tensor cores to generate frames and trace rays, choose one.
M is the default preset for literally everything once you update, however if you force a quality mode in the NV App, it will force a preset.e.g:DLAA - Balanced - Preset K usedPerf = Preset M usedUltra Perf - Preset L usedIf you leave everything default, Preset M is used.
>>107782265wrong this is from official nvidia documentation that says L and M are for performance and ultra performance while K got updated and is the default for anythinb above performance mode https://github.com/NVIDIA/DLSS/blob/main/doc/DLSS_Programming_Guide_Release.pdf
>>107782276i already posted about this months agoyou can use it to undervolt the VRAM, didn't really do anything for high hynix temps thoughbut how is this based>Update Dec 2025: make sure your linux-firmware-amdgpu package is not older than 20251111 version. Starting with 20251125 version AMD took away SMU I2C access. So downgrade to at least 20251111 version (and reboot) before using the scripttypical AMD shit, they've been progressively locking down their GPUs since RDNA 2this isn't how you build a loyal audience when your products look objectively worse in comparisons
>it's another episode of indians and various other third worlders being unable to understand englishThey did not mention any changes to preset K.
kek
>>107782324They did and they were tested retarded jeet, You can literally launch a game with DLSS tweaker and see 4.5 preset K having no issue with artifact dissoclusion and volumetric fog.
>>107782334>>107782334>>107782334
>>107782318firmware driver is independent from mesa driver. you rarely need to update gpu firmware unless your gpu is broken. you still get all the dev fixes from mesa.
>>107782342‘’With today’s NVIDIA app beta update, all GeForce RTX users can experience the new DLSS 4.5 Super Resolution model in over 400 titles. Simply select “DLSS Override feature - Model Presets” in the Graphics tab, and choose “Latest”. This sets our new Model M across all modes. You can also try out Model L under “Custom”, which is optimized for Ultra Performance Mode at 4K. Ultimately, you decide which balance of image quality and performance is right for your PC.’’
>>107782357i know how AMDGPU firmware workseventually there's gonna be some bug fix or performance improvement you're gonna want from the new firmwaretypical AMD move in any casei could do almost anything with MPT on my 6800 XT, it only went into safe mode if you tried to override the core voltagethis 9070 XT is so locked down in comparison it could be an apple product
>>107782387At least they didn't remove hotspot like njudea because they didn't want everyone to know all the rtx 50 series gpus are running over 90 degrees on it.
>>107782406>b-but what about nvidiai don't give a fucki bought a 9070 XT, not an RTX 50 whateverif radeon ever gains any signfiicant marketshare it's gonna get even worse, they clearly hate overclocking and the enthusiast audiencethey want to sell loads of GPUs to clueless normalfags with prebuilts like big brother NVIDIA gets to do
https://youtu.be/SXwXGYhi8_Igsync pulsar just dropped
>>107781850>I dislike QHD and wide monitorsthe pg32ucdm3 is 4k 32" 16.9 bruhbut yes that or this will be the best 4k monitors for nowwill see if the 32 4k OLED from LG is true RGB no WOLED
>>107779219There's actually a trick: buy an used prebuid on Ebay. The ram alone cost as much as the whole machine. You won't get gaymerz DDR5 9000+ but you will get the ram, which is what you want, right?DDR4 machines are readily available so you will get one quickly with 32+gig. DDR5 ram will take more time but it is doable.
>>107782585Oh snap. So what's better then? The tandem OLED or QD OLED gen3?
>>107782442>they clearly hate overclocking and the enthusiast audienceGoodReliability is a higher priority than allowing AIBs to sell overcooked bullshit.Tuners can BIOS and hardware mod if they want to OC so badly.
>>107782784depends how good the lg oled isbetter contrast ratio in bright rooms and better coating lg but more color most likely on qd oledboth have RGB stripe soitdepends(tm)and one is 27 and other is 32unless the new lg 32 is true rgb oled tooworst thing on qdoled was the subpixels so now they fixed it
>>107782585>>107783482>>107782784Just ordered a C5 42''. What will be my expirience?
>>107781033>sellYOU MEAN FUCKING USE?? Holy fuck I hate you rats.
>>107780169You are buying a worse RAM stick for more than I bought in september>CL46 vs CL36>5600MHz vs 6000MHz>260$ vs 110$By your deranged logic, what does it matter that a 5090 will rise to 5K or that modern mobos ask upwards of 400$ for basic functionality like a code indicator. If you cannot afford the price hikes, maybe you shouldn't be buying anything at all