[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip / qa] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/g/ - Technology


Thread archived.
You cannot reply anymore.


[Advertise on 4chan]


File: kirby cyberpunk rebecca.jpg (246 KB, 2060x1648)
246 KB
246 KB JPG
>UPGRADE & BUILD ADVICE
Post build list or current specs including MONITOR: https://pcpartpicker.com/
Provide specific use cases (e.g. 4K editing, high FPS gaming)
State budget and COUNTRY or you will not be helped

>12VHPWR
DO NOT USE ANGLED 12VHPWR ADAPTERS
Fully seat a 12VHPWR connector in its socket, otherwise the connection can melt.
https://youtu.be/ig2px7ofKhQ?t=1345

>Intel Raptor Lake / RPL Refresh
Instability reported with 13th and 14th gen, i5 K SKUs and above (incl non-K). Update BIOS and apply Intel Baseline settings. Avoid purchases.

>CPUS
HTPC(4K60)/Web Browsing: i3 12100
Gaming: R5 7600; budget: R5 5600; eXtreme: R7 7800X3D
Multicore: R9 7950X; budget: i5 13500
>Avoid iGPU-less CPUs
>Existing AM4 board with an old chip? Consider a Zen 3 CPU

>GRAPHICS CARDS
1080p: RTX 3060 12 GB; budget: RX 6600; ultra budget: used RTX 2060 Super
1440p: RTX 4070 Super; budget: RX 7700 XT
2160p: RTX 4080 Super (at MSRP); budget: RTX 4070 Ti Super
Amateur production: RTX 3060 12GB, RTX 4060 Ti 16GB, used RTX 3090, RTX 4090

>RAM
DDR4: Zen3/AM4 - 2x 16GB 3600 MT/s CL18; for locked "non-K" 12th/13th/LGA1700 - 2x 16GB 3200 MT/s CL16
DDR5: Zen4/AM5 - 2x 16GB 6000 MT/s CL30; for unlocked "K" 13th/14th/LGA1700 - 2x 16GB 6400MT/s CL32

>COOLERS
Standout: Peerless Assassin 120 (or variant)
https://linustechtips.com/topic/891730-cpu-cooler-performance-tier-list/

>PSUS
A new PSU should be ATX 3.0 compliant
Aim for 50-75% PSU utilization at full system load
https://hwbusters.com/best_picks/best-atxv3-pcie5-ready-psus-picks-hardware-busters/

>SSDS
WD SN850X 2TB
Avoid: outdated Samsung 970 Evo Plus, cheapest NVMe SSDs
https://ssd.borecraft.com/

>MOTHERBOARDS
Good power delivery is required for Intel CPUs like the i5 13600K and above
Don't pair a K CPU with a B mobo chipset

>GAMING MONITORS
Dell G2724D, AOC Q27G3XMN, MSI MPG 321URX
https://pcpartpicker.com/user/pcbg/saved/dP3v4D

Previous: >>102261938
>>
>>102268017
Disinformation op is back
>>
How unlikely is it that both sticks of my RAM went bad at the same time? Started seeing errors on various tests with UEFI defaults. Even getting errors on both sticks individually with memtest86+. Unfortunately I don't have any spare DDR4 laying around to rule out CPU...
>>
>>102267908
I'm a poorfag

>>102267979
https://www.youtube.com/watch?v=JosuRSY5eRc&t=92s
It's 6 screws and some basic circuit cleaning, how hard could that even be?
>>
>>102268050
Tell the class: are you an Intel or a Radeon fanboy?

>>102268057
Could be mobo as well

>>102268062
Not worth it in my opinion
>>
>>102268057
what cpu
what voltage did you run the dimms at
>>
>>102268114
5900x. 1.35v, it's CJR that I ran at its advertised speed of 3600 with tightened timings that I tested thoroughly when I finalized them.
5900x has been running on 0.9 VDDP, 0.950 CCD, 1.0 IOD, and 1.1 SOC since I got it. No PBO either.
>>
>>102268174
sounds good, cjr should not degrade at 1.35V
both sticks going bad at the same time is sussy.
try the cpu at 4 GHz 1.2V or something to rule out cpu degradation
>>
should i buy the 7900xt. i dont want to wait the 4 months, i just want to run these upcoming games well
>>
>>102268198
>7900 xt
>running upcoming games well
Pick one
>>
Played the new Warhammer 40K game today and hit 1.495v and 100C on my 14700K. Really loving my purchase.
>>
>>102268288
Now that's GAMING. What's that, over 300W to play a game?
>>
>>102268308
250W hehe. Would be higher if it wasn't sitting at 95-100c constantly. Weirdly hits harder than Cinebench? Weird game.
>>
>>102268329
Cinebench loads down all the cores leading to hitting the current limit earlier
>>
>>102268215
im not buying a 3000 dollar 4090 tard
>>
>>102268406
You don't need a 4090 but I would either wait for next gen AMD or buy a 4070TiSuper or higher from Nvidia if buying now. Buying AMD for anything other than low-end is stupid currently.
>>
File: 1385799350491.gif (832 KB, 500x283)
832 KB
832 KB GIF
why aren't there any 5k or 6k monitors for PCs?
it's like everything stopped progressing with 4k in 2014.

also, does anyone know what settings I should use to calibrate my monitor ?
I calibrated it to the "sRGB" settings but it looks sort of weird and yellowish.
gamma 2.2 is too dark so i'm guessing I would want gamma 2.0 for playing competitive games and using the monitor on the web ?
>>
>>102268420
is the 7900xt not a 4070ti super?
>>
>>102268435
No it's not, it's a 7900XT? Do I really need to explain the difference? Typical bar-graph watcher.
>>
>>102268198
>7900xt.

4080 / 4080 super is the best value right now but ofc it's going to be btfo by blackwell in Q1 2025.
>>
>>102268057
I would test the sticks on a different computer if possible to rule out your mobo / cpu being the source of the problem.
They should be easy to rma though with the error code screenshot
>>
>>102268460
>costs 2/3 more than the 4070 super but only about 1/3 faster
This is value to you?
>>
>>102268057
>>102268174
Failing PSU? CJR is fairly sensitive to voltage, the sticks could have been killed by a voltage spike.
>>
Sup/pcbg/, recently my pc has been getting disconnected from my wifi several times. Ive noticed that it ping spikes when im downloading something large or sometimes randomly when im playing a game and it causes my pc to disconnect from my wifi. This doesnt happen with any other of my devices btw. Any tips on what it could be?
>>
>>102268561
do you have i225v chip
>>
>>102268525
yes because I want playable frames with the ray tracing settings turned up and it is the only card that can do that outside of the 4090.
40 something fps is unplayable compared to being in the 60s. ofc I will tweak some settings to get higher fps but I want the RT cranked up to the max in the games that have them.

I also have a 360hz 1440p monitor for it to drive.
>>
>>102268561
See whether it still happens after xmp is disabled
>>
>>102268017
Shit OP again
>>
>>102268580
You get worse visuals not using DLSS with RT due to the lack of Ray-Reconstruction. So regardless these results are useless. The Native Res train left long ago.
>>
>>102268576
I do not.
>>102268582
I've had it disabled before and it was still happening. Im thinking my wifi card might be faulty, I think.
>>
File: Screenshot.png (796 KB, 980x980)
796 KB
796 KB PNG
Are the AM5 epycs the same shit as ye olde xeons? Same shit with differrent sticker?
>>
File: 1699832976364.png (57 KB, 560x890)
57 KB
57 KB PNG
>>102268017
>budget: RX 6600; ultra budget: used RTX 2060 Super
>1440p: RTX 4070 Super; budget: RX 7700 XT
Good luck with that.
Why are you recommending cards released in 2019?
>>
>>102268619
Yes, when you rent a VPS or dedicated server xeon and epyc CPU's compete with each other.
>>
>>102268427
>why aren't there any 5k or 6k monitors for PCs?
It's sadly too niche. Not to mention 1440p is only fairly recently become acceptable after YEARS of horseshit "facts" being spoken about it being "blurry" or "poorly supported." When all that was happening was idiots who thought 1440p entitled them to HiDPI and would do fractional scaling or they would mostly work with 1080 stuff with a blur filter. Add to that the bandwidth requirements for 2880p and most monitor manufacturers just didn't want to deal with your average PC user. And 6K is even more niche. It's why Apple can still cling on to the professional market because it's a niche of people with lots of money and the relatively small number of fixed hardware machines means compatibility can be baked in and you don't have to deal with Capt. Dell Shitbox with his crappy HDMI 2.0 chipset failing to handshake with your 6K reference monitor and causing support headaches.

>I calibrated it to the "sRGB" settings
sRGB is basically the lowest of the low and only there for people who need to color correct for other users of the lowest end monitors. These days you want DCI P3.

For gamma, 2.2 is correct for sRGB and PC in general. 2.0 would be "incorrect" but if it helps you see what you're doing better and you aren't doing color accurate work, then do what makes you happy. Mac gamma was always traditionally 1.8 which would appear blown out to fuck on PC monitors with PC content, but mac software obviously was built to that standard and looks correct where it would be too dark if you used PC settings.
>>
Is redragon an ok peripherals brand? It sounds ultra chink.
>>
>>102268735
Haven't tried it myself but it seems like the thermalright of mechanical keyboards.

If you don't need mechanical, there are plenty of cheaper options both new and used.

For more detailed info, ask in >>>/g/mkg/
>>
>>102268427
>why aren't there any 5k or 6k monitors for PCs?
There are: Dell U3224KBA and Apple 5k and 6k displays
>>
>>102268619
Do you mean to say that an Epyc is just a Ryzen with a different logo like "Xeon" is just an i7 with a different badge?
More or less. Basically a Ryzen isn't certified for 24/7 usage with large memory configurations and full PCIe lane utilisation. If you put 192GB of RAM and some 100gbe NICs and a bunch of NVMe storage on a Ryzen it'll probably work, and it'll probably be fine in a server environment, but then again it might give you some headaches. The Epyc branded version is like the Xeon branded version basically a "Nintendo Seal of Quality" in that AMD/Intel has gone the extra mile to test server configs on server grade OSs and you'll have some comeback if you have problems where you'd have little to none if you complain that your ECC isn't EC'ing on a Ryzen.
>>
>>102268194
Tried bumping SOC but still got errors on each stick.
I noticed the first failing address on each is always the same. The address is different for each of course.
Also tried one of the slots in the pair I've never populated and hit the same address.
Oh, prime95 instantly fails on a thread with large fft, but didn't get any errors after 8 hours of small.
>>102268493
This is the only DDR4 computer I have access to, so I'll just grab a new kit and RMA this one I guess.
>>102268532
It's a 2 year old EVGA 850 G6. No multimeter but the values reported by the motherboard are well within ATX spec, including under load. But I do have a window AC that makes the lights dim every time it kicks on...
>>
>>102268769
>Do you mean to say that an Epyc is just a Ryzen with a different logo like "Xeon" is just an i7 with a different badge?
He means to say Epyc and Xeon are server/workstation offerings.
>>
https://files.catbox.moe/uiswz2.mp4
>>
>>102268753
I want to dip my dick into mechanical but this segment of market looks like a fucking scam.
>no numpads on every other model
>tin as fuck, built for baby hands
>often 150$+ for a piece
I need something that's like a normal kb but built well enough so it doesn't like shitty xylophone
>>
>>102268773
>Tried bumping SOC but still got errors on each stick
Yeah that's not going to help. You probably don't even need 1.1 V for 1800 FCLK, I run 1.0 V at 1866 FCLK. I think the other anon meant running the cores at low frequency and 1.2 V.
>Oh, prime95 instantly fails on a thread with large fft, but didn't get any errors after 8 hours of small
Are you running 1T GDM disabled or something? That is notoriously finicky and the best way I've found to test the stability of that is P95 Large. I can pass an hour of TM5 but fail almost instantly in P95.
>It's a 2 year old EVGA 850 G6. No multimeter but the values reported by the motherboard are well within ATX spec, including under load.
I don't think that would be the reason then, the PSU should "smooth out" external power fluctuations, but I've had a PSU in the past that killed two GPUs due to voltage spikes. B-die, DJR and Rev E can take high voltages, CJR can't.
>>
>>102268696
>sRGB is basically the lowest of the low and only there for people who need to color correct for other users of the lowest end monitors. These days you want DCI P3.
>
>For gamma, 2.2 is correct for sRGB and PC in general. 2.0 would be "incorrect" but if it helps you see what you're doing better and you aren't doing color accurate work, then do what makes you happy. Mac gamma was always traditionally 1.8 which would appear blown out to fuck on PC monitors with PC content, but mac software obviously was built to that standard and looks correct where it would be too dark if you used PC settings.

I'm really just playing pc games and watching biblical studies videos on my pc with the occasional High res picture of Jesus for viewing.
My main interest is having accurate colors without black crush on the shadows. When I calibrated to gamma 2.2 it was too dark and like this guy on reddit :
https://www.reddit.com/r/OLED_Gaming/comments/1c4t7sq/22_gamma_vs_srgb_gamma_pg32ucdm/
I set my monitor to "srgb" instead of "wide gamut" since the colors seemed over saturated on wide gamut but i really don't understand what is going on.
The main games I play are pc titles and I don't think they were made with wide gamut support since the colors looks so cartoonish.

it is all confusing to me since I am not a videographer or phtographer or colorist.
pic attached is the srgb calibration info from displaycal
>>
>>102268804
Respectfully, learn to use pcpartpicker. This link displays all the models with mechanical keys and numpads (the nerd term is that you don't want a "tenkeyless" keyboard)
https://pcpartpicker.com/products/keyboard/#k=1&T=0&sort=price&page=1
>>
have 5700x3d, at 1440p, what would /g/ choose,
both cards evga FTW, used;
3080, used for ~3 years, for 380euros or
3080ti with warranty till 2027 for 520 euros?
How likely it is for a gpu to fail? like 10% chance/year?
>>
>>102268811
I've been testing it with default settings, so the RAM is running at 2133, GDM on, and so on. I did keep GDM on with my actual settings, though.
>>
>>102268763
>Dell U3224KBA
wow, a single monitor that costs $2.5k

That res is essentially dead even though the jump from 4k to 8k is massive in terms of processing requirments.
>>
>>102268773
i've never had a bad kit of ram from the reputable companies like corsair, g skill, micron, team group or samsung.
I did get a bad laptop stick from some more obscure low end brand that would through errors when you ran tests on it. I just rma'd it and they sent me a working stick back.
>>
>>102268654
>10gb at low settings
This is absolutely a true result and can be added to the mix but, can we take a moment to acknowledge that this is retarded? The reality is most people have a 3060 and making the game unplayable on 80% of peoples computers just doesn't seem smart? Feel free to disagree.
>>
>>102268982
>Feel free to disagree.
I disagree because if we had it your way the graphics would stagnate since devs would be catering to the weak GPUs over and over again instead of pushing the bar higher gradually.
>>
>>102268933
Quit whining retard
>>
>>102268427
https://www.dell.com/en-us/shop/dell-ultrasharp-32-8k-monitor-up3218k/apd/210-alez/monitors-monitor-accessories

theres this but it costs 4k and only does 60hz
>>
>>102268933
You moved the goalposts. First, you said
>why aren't there any 5k or 6k monitors for PCs?
Now that I proved there are such displays, you moved the goalposts and complain about the price. Surprise surprise, new cutting edge technology is expensive.
>>
>>102268599
then why doesnt nvidia open source it? oh thats right their being faggots again, ass pass grass
>>
>>102268654
this games unoptimised dogshit but vram requirements are ever gonna go up especially in 2025+ when next gen comes and pro refreshes
>>
>muh praystation
>>
>>102268785
LINUX VAX TIPS
>>
>>102268876
79XTX you get 3090ti-4090d perf for 4080 money
>>
Bros, what's the best budget GPU to play Starfield?
>>
>>102269192
that depends on your budget
>>
What mainboard should I use for the 3060 and 12400f, H or B? Budget build so I dont want to spend more if it not great
>>
>>102269268
H610 is trash, B660 minimum.
H670/B760/Z690/Z790 are the next tiers in order of what you should be looking at.
>>
>>102268017
Thoughts on the crucial bx500 ssds?

Just noticed that hdd are basically the same cost as SSDs these days for 1tb storage. What the heck happened?

What's the best bang for buck storage solution for 1 or 2tb storage?

I've always used my ssds for games only and the HDD for media/digital hoarding. But looks like it's a good idea to just used SSDs?
>>
>>102269783
They are obsolete except for old builds without a M2 slot
>>
>Play Lost Judgment on a 6600xt
>Smooth enough, maybe a framedrop here and there, no stutters
>Upgrade to a 4070 super
>Constant stutters until I turn on DLSS and make shit look blurry
Waste of 600 bucks
>>
>>102268017
I bought an Asus TUF Gold 750w PSU with an intent of running a 4070 Super and 5700x3d but now I'm considering buying a used 3090.
Is it gonna handle it or am I walking a tightrope? Maybe a slight undervolt?
>>
Repost from previous thread:
>My current home PC (13700, 4070, 64GB RAM) sometimes struggles with Fusion 360 if I have big assemblies or very complicated parts with lots of features. While it's struggling to compute everything I'm not seeing I high load on any of the relevant components: max 15% on the CPU and GPU, max 2GB of RAM. Is the optimisation just shite, or is the some way I can unleash the whole system to get shit done quicker?
>As for my work laptop, it's much worse at doing the same thing. I have asked the IT guy to get me a replacement that isn't slow as balls and he recommends a laptop with an RTX A2000, 13700H and 32 GB RAM. This is all within budget - but will that A2000 fix the problems I'm currently having with the 4070? On paper it should be a slower card for gaming, so it better have some great optimisation for CAD...

>>102264101
So in that case, there might be a bit of optimisation in the GPU (A2000 vs 4070, for example), but not much? Probably could overkill it with a 4090.

>>102264116
>fusion360 is single thread performance limited
Those motherfuckers. I understand why, but god dammit that sucks big fat hairy donkey balls.
Last night I did some further testing, when I turbospaz complex models it does indeed load my GPU a bit more (up to 60-70%) as does the CPU (up to 40%), but the hangup is with computing the model itself, not rendering it - and that is stuck to single core performance? Jesus christ. Is there any way to go quicker?
>>
>>102269783
>Thoughts on the crucial bx500 ssds?
Cheap and cheerful but fine for a secondary drive. Particularly if you're out of M.2 slots.

>What the heck happened?
HDDs have a higher floor price on account of their more complex workings (moving parts etc.).

>Looks like it's a good idea to just used SSDs?
For ≤2TB capacities definitely. Well worth the extra ~$20 for no noise and fast access times. At 4TB+, HDDs become more viable depending on your budget. These days, however, 2TB is pretty scant for a movie drive if you're not a Yify enthusiast.
>>
Z790-A MAX WIFI Pro doesn't seem to align with the standoff screws (by a few mm) with the pre-installed IO shield. They gave me a standardized manual for one without an IO shield. And there's no real videos for it. Is there some technique to getting it in alignment? I've never done this before.
>>
The low key humming on my PC is driving me nuts.I don't think its my fans because I have tried turning them off and the noise is still there
Is this what they called a coil whine?
>>
>>102269857
>things that never happened
Did everyone stand up and clap afterwards?
>>
>>102269970
Could be the psu fan or gpu fan or even a noisy HDD.
>Is this what they called a coil whine?
Possibly but unlikely as coil whine is - as the name suggests - normally heard as a whine rather than hum.
>>
>>102269970
coil whine is often more high pitched
If your PSU has a zero RPM mode, enable that and check if it makes any difference
If not you could still turn it on without plugging it to anything with a clip and seeing if that's the culprit
>>
Livemixer build still serving me well.
Might build a xeon shitbox just for fun
>>
>>102269970
It may be a coil-whine of VRM or in PSU. It may even be a coil-whine in your display's power brick, anon.
>>
>>102269857
nvidia's drivers are less efficient and require a fast cpu with at least 6c/12t to run properly
>>
File: laughing-crying.gif (2.85 MB, 498x280)
2.85 MB
2.85 MB GIF
>nvidia's drivers are less efficient and require a fast cpu with at least 6c/12t to run properly
>>
>>102268580
>yes because I want playable frames
By your metrics not even the 4090 reaches my screens target refresh rate for adaptive sync, into the trash raytracing goes

Call me when cards can turn it on and lose 0 to 5% perf over raw raster.
>>
>>102269192
A second hand 6800 XT / 3080 for 1440p
>>
>>102269783
>bx500
Slow but cheap, once they exhaust their cache they write at HDD speeds. Fine if you just want some cheap storage.
>>
>>102269868
You could easily run a 4090 or 7900XTX on a 750w psu when running a 5000x3d cpu.
>>
File: dsc_bandwidths_2.png (296 KB, 717x613)
296 KB
296 KB PNG
>>102268427
>why aren't there any 5k or 6k monitors for PCs?
>it's like everything stopped progressing with 4k in 2014.
As others have said there are a few niche 8K displays, but the real reason for the halting of widespread >4K displays is down to 2 reasons:
1. Economies of scale settling on 4K for TVs creating manufacturing inertia for monitor market
2. Lack of real increase in bandwidth from stagnating display standards (full UHBR20 DP 2.1 is STILL not out). We can barely increase refresh rates as is using extreme DSC, forget trying to do the same thing at an even larger resolution with a 4:4:4 RGB signal:
https://tftcentral.co.uk/articles/when-is-displayport-2-1-going-to-be-used-on-monitors


>also, does anyone know what settings I should use to calibrate my monitor ?
>I calibrated it to the "sRGB" settings but it looks sort of weird and yellowish.
>gamma 2.2 is too dark so i'm guessing I would want gamma 2.0 for playing competitive games and using the monitor on the web ?
You are probably used to a cooler whitepoint temperature that a lot of displays come with by default since it artificially inflates the brightness of the native backlight and is supposedly better for eyesight in office enivornments. Use the Office Paper Test with Online tools to get something more neutral without buying external tools:
https://www.youtube.com/watch?v=avJTz1JhkR4
>>
>>102268580
This is pathetic. Modern rendering is a joke
>>
>>102270661
that's true of all cheap SATA SSDs and BX500s aren't all that cheap compared to other cheap SSDs, the MX500s are worth the money.
>>
>>102270661
>>102269919
What's the one to get if you want 1tb? Is it the mx500
>>
Installed PTM7950 on a 7700x idle temps are exactly the same, 40-44C'
Had 6yo noctua that I installed 2 years back in there, it was completely dry, amazing my CPU worked at 90C' full load.
Couldn't open the GPU, my screwdrivers are shit. Couldn't turn one last screw. Any advice for screwdrivers kits?
>>
>>102269095
console refreshes dont change the landscape that much if at all since devs will still target the base models
>>
>>102271131
Yeah the MX500's are better, if they're similarly priced get them instead. If you just want mass storage though the BX500s are fine if the price is right
>>
>>102268753
dont go to mkg
all they usually do is wank over 60% and chinkoid keycaps
for some odd reason theres been a lot of shilling in the last few months for those weird diy split ergo boards
>>102268804
look into keychron for a decent mechanical kb in all sizes
>>
>>102270933
another big thing is diminishing returns

4k is already good enough to not be able to see pixels at typical monitor sizes
which is not to say that more wouldn't be better, it definitely would, especially by allowing scaling by integral factors.
a 5k 27in monitor can be scaled at 200% to yield 1440p appearance, whereas a 4k panel would need to use 150%. this would improve picture quality beyond the basic ppi gain, by eliminating any artifacts from fractional scaling.
but still, while the gains are real, we are very much into diminishing returns here. 4k is already very sharp even at 32in

plus there's a couple more issues, like the fact that media content is also stuck at 4k, or more importantly the whole situation with gpu prices and game requirements getting completely out of hand.
who the fuck wants to buy a 5k-6k-8k display, when the gpu requirements to game on them will be beyond absurd and when you don't even need them for media.
really, such panels are only for techbros who are willing to drop $$$ just to make the text on their IDE look nicer.
>>
>>102268753
99% of people with mechanical keyboards do not need mechanical keyboards, which is why the K120 is recommended heavily.
>>
>>102271214
just run it at 1440p 2:1 scaling for gaymes
>>
>>102271149
Amazon, Home Depot, Lowes, Walmart all sells generic screwdriver kits. All you need is a small Philips head to open a GPU.

Also one thing with PTM7950 you will notice is pretty much everything idles at 45C with it because that's the melting temp of the material. It only gets better as it gets hotter and more pressure is applied.
>>
>>102271149
did you fuck up the screw head trying to undo it?
is the head sticking out like a mushroom or flush with the surface?
>>
>>102271287
>PTM7950
is this the same sort of thing as that kryosheet?
>>
>>102271316
Kryosheet is graphene cut into a pad. PTM7950 is different by being a phase change material in pad form.

https://www.igorslab.de/en/overhyped-honeywell-ptm7950-in-lab-test-and-as-game-changer-for-graphics-cards/
>>
>>102271149
don't buy any merch screwdriver kits from LTT/Gamers Nexus/ifixit.
any hardware store will have a cheap $10 set of VDE screwdrivers that are most likely Wera clones that will last you forever and have a PH0, PH1 and PH2.
>>
PTM7950 seems like snake oil to me because of Linus pushing it.
>>
File: 06-Torture-Hotspot.png (50 KB, 980x766)
50 KB
50 KB PNG
>>102271401
Its used in real industrial applications. Honeywell also makes ACs you probably know someone with a Honeywell fan or chiller.
>>
>>102270418
I have a 12c/24t CPU, not the best for gaming
Setting low latency to Ultra on nvidia "fixes it" as in just dropping down ~30fps
>>
>>102271192
Mx500 is $107aud
Bx500 is $88aud

Worth just putting the little extra?
>>
>>102271297
>did you fuck up the screw head trying to undo it?
it started skipping so I stopped, head is fine
It's red devil 7900xt, all other screws aren't tiny, the retention bracket screws are small, three turned fine-ish and the last one stuck
>>102271287
it sits at 43C' for an hour now, which is better than my dried out paste already by 1.5-2C'

>>102271389
we don't have that here, seen xiaomi mijia set for $10, there is like billion of them for same price
>>
File: 1695736101864255.jpg (277 KB, 1601x911)
277 KB
277 KB JPG
Why did AMD do away withe X series by just making the non-x into X? Where is the real successor to the 7700x?
>>
>>102271379
let's pray this gives AIBs a push to use it for all the cards, it's not that much more expensive in bulk than the paste they sell it in big rolls from Honeywell directly
>>
>look guys arrows
>>
>>102271379
does he have any tests where its compared to these graphene pads?
>>
>>102271536
because everyone kept buying non x versions. personally I prefer X variant, less OC headache.
>>
>>102269014
graphics have already stagnated retard
everyone is trying to make their shit hyper realistic and it ends up looking the same as AAA slop from 5 years ago while requiring 3 times more resources to run
>>
>>102271566
there are graphene tests for CPUs, it's meh compared to paste
>>
>>102271571
id be interested to see the best looking game from 2019 vs 2024
>>
>>102271557
the bulk paste they use is dirt cheap. if they can save pennies thats huge in mass production. also its planned obsolescence on purpose to get people to buy a new gpu or buy warranty service.
>>
>>102271585
RDR2, Control
>>
>>102271594
nvidia used it for FE cards from what I heard
>>
File: 1725069115191464.jpg (127 KB, 1280x720)
127 KB
127 KB JPG
Intel won
>>
File: 1717867491359.png (62 KB, 954x855)
62 KB
62 KB PNG
>>102271566
>>
>>102271608
yeah cuz the 4090 pretty much needs it. a 4080 on paste is already hitting 100+ c after just 6 months.
>>
Which motherboard should I get for casual 1080p gaming for R5 5600x + Rtx 4060?
>>
>>102271585
make sure to compare the computing requirements relative to graphical improvement because only the latter is seeing diminishing returns
instead of making everything look like the same hyper realistic realer than real life trash they should at least try to make something different
>>
>>102271626
B550 of your choice
>>
>>102271617
huh, kryosheet looks fine though, I've seen tests where it was at the bottom
>>
>>102271631
this is on a cpu where it has a ihs. if you were to test under a gpu or direct die then results may be different. but both arent bad to use.
>>
what junction temperature should I shoot for if I want to maximize gpu life?
the default fan curve is leaving a lot of head room before I can start hearing it.
>>
>>102271627
>same hyper realistic realer than real life
that's what I like though, fuck you weeb tastes, I want my hyper realistic spess games and fantasy RPGs, I want to believe in the world I got into and realsitic graphics help to suspend disbelief more
>>
>>102271476
If it's for 1tb then yeah I'd spring for mx500
>>
>>102271617
theres less than 2 degrees between 1 and 20
theyre much of a muchness at that point
>>
>>102271639
undervolt it, what GPU? mine goes 100C' and default, with undervolt it hovers around 70C
>>
>>102271639
according to the gpu makers, anything below tjmax is "optimal operating conditions" and "not for concern for damage or warranty replacement".

>>102271648
the real difference is long term use. ptm7950 gets better with age. the rest deteriorate with age.
>>
>>102271598
better comparison would be rdr2 ultra settings vs alan wake 2 pathtraced if we're going for best looking of each year
>>
>>102271648
that's why MX-4/MX-6 is recommended because its cheap and abundant.
>>
>>102271656
dont those graphene pads last long as well?
>>
>>102271661
I got 40x80mm sheet of PTM for $8 here, It's going to peoples through wider channels with each year
>>
>>102271660
>alan wake 2
Movie with signs of gameplay. Doesn't count.
>>
>>102271693
>and control counts
it's two exactly the same games.
>>
>>102271653
rx 6800. but under volt seems unnecessary when it maxes the fan out at 25% to keep the card under 90C. and I can't hear the thing until it hits 40%
>>102271656
I don't mind not playing at max settings and don't buy every generation of card so I would prefer to keep them in top condition past the warranty period.
>>
>>102271665
they last much longer than paste and technically reusuable but if you need to reassemble for any reason (fucked up mounting on install, changing memory pads, cleaning, repair, etc) then the pad rips apart easily. ptm7950 goes back to solid at room temp and easy to remove when its not liquid.
>>
>>102271660
I really can't notice any difference in alan wake 2 pathtraced vs raster
>>
>>102271708
for rx 6800 the tjmax is literally 110c. anything below that the card will boost beyond its stock speed like ryzen.
>>
>>102271702
but only one of them fits my narrative
>>
>>102271712
same, they did a good job with lighting
2077 is the only example where PT shines, and only in some scenes, I played it last year and kept switching between normal RT GI and PT, some scenes look awful others perfect, normal RT is more consistent imo
>>
>>102271709
i reseated my phantom spirit at least 6 times in the last week trying to figure out what was wrong with the pc (turns out it was a faulty 7800x3d) and my kryosheet still fine
maybe i was lucky or others are heavy handed
>>
>>102271702
By that logic portal or HL2 aren't games either.
>>
>>102271746
how did you figure out it was the cpu? what were the symptoms
>>
>>102271746
It really depends on the batch, components, age of the pad.

For example MBA AMD cards used graphene pads and they pretty much rip apart once the cooler is removed.
>>
>>102271759
i took the cpu back to the shop and they swapped it. pc is now being stress tested but it wouldnt even boot yesterday
>>
>>102271709
Can you use pm7950 on cpu's?
>>
>>102271610
>benchmarks
nor remove the estrogen cores from the equation and show the real world data in cpu heavy games since that's the only thing X800x3d excels at
>>
>>102271850
yes but they're meant for direct die more than ihs.
>>
>>102271863
i hope amd doesnt go down the same route and put those 'c' cores in future zen desktop processors
>>
>>102271892
they don't have to, intel can't into proper chiplets yet that's why they have to use e-cores to save on yields while keeping multicore performance competitive
AMD "infinity fabric" was a revolutionary protocol others have to catch up
>>
>>102271455
your cpu works harder with the faster graphics card and stresses the vrm to the point of throttling
how shit is your motherboard?
>>
File: 941Dqjfx8PVe4RUX.jpg (238 KB, 2133x1200)
238 KB
238 KB JPG
>>102271892
C cores aren't like Intel's cuck cores, they're the same architecture as the main core made in the smallest footprint possible with a sacrifice made to cache and due to its density, clock speeds.
Zen6 is going to be on N2 and will probably be 12 core CCDs like Zen4C but full cache, but also the packaging on Zen6 is changing so that's going to be interesting as well. but we're getting ahead of ourselves, Zen5X3D has a bunch of changes and I think it will be a significant uplift over the 7800X3D.
>>
>>102271990
It's a MSI MPG B550 gaming plus and the CPU is a 5900x
That game's not even that demanding, what the fuck are you smoking
>>
have 5700x3d, at 1440p, what would /g/ choose,
both cards evga FTW, used;
3080, used for ~3 years, for 380euros or
3080ti with warranty till 2027 for 520 euros?
How important warranty for gpu is, or fail rate is low enough not to care about it spending extra 25%?
>>
>>102272135
How much is peace of mind worth to you? Does you know how the 3080 was used? Mined 24/7 or peak gaming loads? Was it kept in a well ventilated dust free case?
>>
File: mpgb550gamingplus.jpg (49 KB, 554x320)
49 KB
49 KB JPG
>>102272115
turn off 6 cores and see if you still get the stuttering. alternatively see if your xmp profile is enabled
>>
>>102269857
Lost Judgement loves AMD because it was optimized for consoles first. Nip devs don't care for Nvidia unless Nvidia pays them via Gameworks (Sega PSO2)
>>
>>102272135
I would get neither and get a 7800XT for 495 euros
>>
>>102272162
Quite a lot, thats why im asking. If it was RAM or CPU I wouldnt bother thinking about warranty. Seems nicely cleaned so no idea about dust-free environment.
Its non lhr version, so there is chance it was mined with. In benchmark reaches 75 core temp.
Guy says it was used only for gaming, but who knows.
>>
File: 1722401336517406.gif (151 KB, 400x267)
151 KB
151 KB GIF
I've been stuck for hours thanks to a seemingly completely normal 3-pin fan connector refusing to go in the SYS_FAN header. Why can't even the simplest things just work?
>>
>>102272239
Why? Its around 3080 performance in native resolution and doest have dlss. 45 euros extra gives GRE, which goes around 3080ti.
>>
>>102272195
No change, XMP is enabled
>>
>>102272276
>Quite a lot, thats why im asking
I think you know the answer then
>>
>>102272195
What makes a change is turning off PCIe resizeable bar on the BIOS
It lowers FPS just like setting low latency to ultra but it get rids of the stutters I guess
>>
should I buy lexarnm790 with integrated heatsink (same price as without heatsink) or use the heatsink on my motherboard?
>>
File: yin-yang.png (619 KB, 1600x1600)
619 KB
619 KB PNG
is +300€ to go from a 7900xt to an xtx worth it?
>>
>>102272276
miners usually sell multiple cards at once
>>
>>102272479
buy the one with heatsink so you don't have to mess around with thermal pads.
store the motherboards heatsink in the box so you don't lose it
>>
>>102272319
If it jams, force it. If it breaks, it needed replacing anyway.
>>
>me before ordering the pc
yea i can for sure assemble it
>me after ordering the pc
fuck am I sure I can do it
its so over
>>
>>102272598
watch a youtube video tutorial
>>
File: s-l400.jpg (38 KB, 400x374)
38 KB
38 KB JPG
Why couldn't they just have the entire shroud cover the entire board?
Just finished my AM5 build and I know it doesnt matter too much if some paste squeezes on to the small bits but it's fucking annoying. I like tidiness.
>>
>>102272619
i bought the contact frame just for that reason
not sure if its going to void my warranty
>>
>>102272611
i watched the linux tech group one
are there any others good ones?
also watched the one that had my cpu/motherboard/case/gpu in it
>>
>>102272598
put the cpu in the board
put the nvme drive on
put the cooler on
put the ram in
put the psu in the case
put standoffs in the correct holes
screw the board into the case
connect cpu 8 pin and 24 pin
connect power button to motherboard
connect fans
put graphics card in
connect pcie power
connect displayport
connect mouse and keyboard
connect psu power cable
turn the switch on
press power button
>>
>>102272522
so if i put a motherboard heatsink on top of the nvme do i need thermalpad too?
>>
>>102272669
it's probably already on there just remove the clear plastic protector
>>
>>102272659
Just have it saved in oyur phone or something. Also your mobo should have a pic with all connectors named in it. Just stick shit where it belongs.
>>
bros my pc is going to be so good
im finally going to be happy
>>
>>102272711
oh anon
post assembly depression is a real thing
>>
>>102272718
>building my first pc that wasn't random secondhand shit
>took me 4+ hours
>drained as fuck
>only checked if it boots and went to sleep
>>
>>102272677
im getting a asscock livemixer mobo cant find it whatever it has it or not
maybe thsi anon >>102270131 can confirm or deny that livemixer's nvme heatsink has thermal pad
>>
since op is apparently unreliable, is the 90 model necessary for 4k? or will a 4080 or new gen equivalent be just fine?
>>
>>102272735
yeah i made the same mistake with my first one too

the second build i just took off an entire day, did everything step by step, eating dinner, walking my dog in between and just said fuck it, gonna finish it tomorrow
0 mistakes on that one, absolutely smooth sailing
>>
>>102272790
if 4k gaming is huge priority for you, id wait a bit for 5th gen to drop
4080 can be a bit hit or miss in 4k gaming
>>
>>102272790
>is the 90 model necessary for 4k?
yes and no, you can do 4k with 4080 or 7900xt, hell even with 7800xt/4070ti/7900GRE but it's not a very comfortable 4k, running it close to that 60fps lower limit
we practically know nothing about next gen, it could be another dud or revolution we are waiting for a few gens now
>>
>>102272619
to put the capacitors on the front side
>>
>>102272660
you forgot
>connect the psu to the power grid and keep tapping it to discharge static electricity
>turn the hot water on for half an hour in the bath room so all the dust sticks to the walls while you are assembling your shit
>put ram in before the cooler
>flash the bios before putting in the cpu
>apply paste to the cpu before putting the cooler on
>remove the film from the cooler
>>
>>102272619
do not add that much paste, duh

>>102272498
no, are you crazy? 7900XT is perfectly fine for couple more years even at 4k

>>102271850
>>102271868
well, I ran a few tests with my recent 7950 application, and it runs pretty well on a 7700x and D15 cooler. averages 50C' at 50% load for aout an hour, used to be 60C on my old thermal paste, it should get better with time in theory
I have a weird phobia now, I know it shouldn't happen, but can it leak out under the cooler in liquid state if its vertically mounted?
>>
>D15
Ishygddt
>>
File: 1718363814559565.jpg (102 KB, 660x462)
102 KB
102 KB JPG
Bros can someone tell me the difference between "Full", "Minimal" and "Driver Only" installation types for AMD GPU drivers?
>>
>>102272959
driver only means you only get the driver, not the software suite with useful features
I usually click full install, it doesn't have faggy bloatware like geforce experience
>>
>>102272959
driver only you get only the driver, no tools, no settings
minimal they cut recording and screenshot functionality fromw what I remember, I think OC and monitoring functions are in
full you get recording and screenshots
I never installed anythign but full, in driver screenshot tool is too comfy to use
>>
>>102272979
I get the driver only option but what does minimal leave out compared to full?
>>
>>102272955
I got it, soon to be, 8 years back, I think it's worth it for $100 to use a thing 8 years and it will last me another 8
>>
>>102272996
You don't get undervolting with minimal.
keeps it nice and quiet and you get more performance
>>
a bit out of distribution for this thread, but any good USB wifi adapters? my tp-link is starting to fail me. I need 5G but nothing else really, I'm near the modem.
>>
>>102272735
the wisdom of the ancients says "never format after midnight"
>>
1440p doesn't seem worth it to me. Having to spend $600+ on a gpu sucks, even if budget 1440p monitors are hardly more expensive than 1080p these days.
>>
>>102272898
>but can it leak out under the cooler in liquid state if its vertically mounted?
This was my burning question as well
>>
>>102272955
i'd buy the tard tower in a heartbeat if my 19-year-old scythe ninja ever stops working
>>
>>102273218
Lower settings: Ultra is basically Medium/High and don't play AAA slop.
>>
>>102269014
Shoving a higher texture resolution has zero impact on visuals past a certain point so not really sure how graphics would "stagnate" RDR2 is still one of the best looking rasterized games and doesn't use more than 6.5GB at 4K.
>>
>>102273232
from the open tests I've seen in it's liquid form it is the same consistency as a thermal paste, so it shouldn't?
>>
>>102273438
>>102273232
its resistant to pump out over rapid heating and cooling cycles, that's the whole point of the thing. it can still squish out if you exert too much mounting pressure and even then who cares its non conductive and easy to clean.
>>
Damn, what happened to userbenchmark?
Last time I was shopping for parts ~5 years ago it seemed like a somewhat legit site, now their AMD reviews read like they were written by a deranged stroke victim whose wife eloped with an AMD employee and who lost all his life savings on Intel calls.
>>
File: 1724698983085477.png (157 KB, 332x966)
157 KB
157 KB PNG
How profitable is pc reselling / flipping ? Like building computers off shit i buy on taobao / used parts and flipping them. I need to make some bank because i only make 80% of the median salary in my cunt
>>
I give up on shit
Any games where ReBAR actually makes a difference on nvidia GPUs? I'm finding it does more harm than good, this nonsense
>>
>>102273692
smart access provides a small but real benefit in most games
rebar is placebo or hinderance
that was the mantra from 2 years ago and I don't think anything changed since then
>>
File: 1696116576114296.jpg (53 KB, 640x480)
53 KB
53 KB JPG
Is there a big downside to buying a one fan card over a dual fan? Is it going to be super loud or overheat way more?
>>
>>102273992
depends on the card is it a 4060?
>>
>>102274023
3060
>>
Anons, I currently have a 5600g and plan on getting a GPU first then buy a better CPU after a month or two

Which GPU would you recommend on $300 range and a CPU to pair it with later on
>>
>>102274023
nta but would the zotac single fan 4060 be bad? I saw a video it performs well, not sure about the heat
>>
>>102273992
I think they usually run silent-type vbios on cards like these to keep the temps down a bit, so they're a tad slower than dual fan cards
>>
>>102274075
4060 is an extremely low power part so yeah I reckon a single fan card is fine but if you can find a dual fan card just get that for a similar price unless your case is highly space constrained
>>
File: GF3o5__bEAA2yxn.jpg (1.02 MB, 3000x3000)
1.02 MB
1.02 MB JPG
Day 520 of waiting for next-gen GPUs.
>>
>watching gpu unboxings from 3 years ago
>people paying $700-$800 for a 3060
jeez surely nobody selling any of this now can be making any profit on them
>>
File: 53815_3.jpg (91 KB, 626x531)
91 KB
91 KB JPG
>>102274096
This is my current setup with a single fan 1660s
If I get a dual fan 4060 im not sure if theres any cool air that would reach the tower cooler.

Im also considering the Gigabyte 4060 lowprofile , 3 fans but a whole lot slimmer along air to reach the cooler
>>
>>102271214
I'm salty because i have some 5k bible study videos and the only monitor i see to view them on amazon for windows desktop pc is a $900 samsung
>>
>>102274463
why
>>
>>102274463
the bible study industry has always been at the forefront of new tech
still, with a half decent downscaler you'll still get excellent image quality on a 4k panel.
>>
File: Projected build v2.png (61 KB, 901x818)
61 KB
61 KB PNG
>>102268017
Omeletanon from a couple of threads ago, here

Building a new computer for gaming and plugging in my TV to watch YT in the living room.
I'd like not to change too many parts over the next ten years.
Budget is under 3K€, France.

Here's the updated list:
https://fr.pcpartpicker.com/list/vtVbWt

I'd love some opinions, especially regarding the RAM (should I go for 2x32, for future-proofing?) the case (how do I make it as noiseless as possible?) and the motherboard (will I have compatibility issues/how do I compare with other mobos out there?)
>>
File: arrowlakeprice.png (79 KB, 1166x706)
79 KB
79 KB PNG
>>
>>102274282
dual fan would still be ok in this config since again the 4060 is so low power the air reaching the cpu cooler should not be that hot buy a single fan gpu will still be fine
>>
>>102274463
what are you doing just get the apple pro display xdr
>>
>>102274687
Yes go for 2x32gb
You can compare motherboards using the am5 motherboard spreadsheet (first reddit link on Google). That board is fine though but the 3rd nvme slot is a x2 link which is really weird
>>
>>102274687
you literally dont need the noctua
peerless or even phantom spirit will do the trick with 7800x3d, its not intel
invest a bit more and get yourself a 4080 super
mobo is fine I guess, you're not gonna change the cpu until far down the line, its gonna be peak for years still
32 gb is ok for now, you can get those now and in a couple months you get other 32 and presto you fixed your shit
and eventually get a 4tb ssd
thats it
>>
>>102273676
judging by the shitty deals i see on Facebook marketplace and some auctioning sites it appears to be fairly lucrative to sell to the ignorant.
>>
File: 45346467756.png (711 KB, 1700x1770)
711 KB
711 KB PNG
>>102275226
>you literally dont need the noctua
cooler? Noctua
case? Fractal
PSU? Seasonic Titanium
yup, it's PC building time
>>
>>102275226
>Get 32 now
>Add 32 later
That's not how ddr5 works
>>
90% of 4090 owners don't actually know anything, they're wallet gamers obsessed with buying the most expensive thing. you can pick them out because they always have $600-700 motherboards or a 420 watercooling for their 75W gaming TDP X3D chip.
>>
File: jpeOBzKh2YU.jpg (40 KB, 326x416)
40 KB
40 KB JPG
Are we ever gonna get an SSD capacity breakthrough with 10TB+ sticks?
>>
>>102275362
500 layer nand needs to be produced for that to happen
>>
>>102275362
>Samsung's roadmap calls for a 1000 plus layer V-NAND product by 2030
maybe?
>>
>>102275424
>>102275430
Does nobody do non-silicon based tech yet?
>>
>>102275430
That will make 20tb+ m2s possible
Were already at 8tb with just 112-layer bics5 and 276-layer ymtc nand
>>
>>102275451
you mean biotech? yeah.
>>
>>102275451
too far way, fun times ahead when 16A aka 2nm+++ becomes a common thing
>>
File: 719VKGbV1iL._AC_SL1500_.jpg (202 KB, 1450x1500)
202 KB
202 KB JPG
>>102272619
Use the DeepCool copper frame or ThermalGrizzly graphene pads.
>>
File: hmmm-3420533807.jpg (28 KB, 600x497)
28 KB
28 KB JPG
how many generations till we see vram kits?
>>
>>102275984
There’s a lot of incentive for them to exist. We’ll see, but I’m fairly certain Nvidia wouldn’t want them around.
>>
Help me understand /pcbg/
I keep seeing everywhere people claiming that if you're going to use thermal paste on a GPU die/laptop you must spread a thin layer. It is imperative that you do it this way because apparently too much TIM is bad but so is not covering the entire die. Like you're supposed to spread it thinly, evenly, perfectly. Isn't the GPU cooler supposed to push out the excess anyways? Am I missing something?
I was thinking about how with an older GPU I usually had less problems when I just used the X method instead of spreading it. It was a bit of a mess afterwards because some of these pastes are a bit extra viscous, but it did the job. What's this all about?
>>
>>102276195
GPU is a naked die. You need to cover everything.
>>
>>102273676
Decently just have in your terms that you won't offer tech support. Only basic troubleshooting 30 days from purchase.
>>
>>102275326
>90% of 4090 owners don't actually know anything
I'm not sure if it's actually that high, but generally I would agree. The 4090 for gaming is fine if you know exactly why you want it (path tracing), or just want da best, as you're ((((((only)))))) paying paying a $300 price premium for the performance, and against the total cost of the build, probably not a big difference. Still, I'm against consumerism.

>>102272790
>is the 90 model necessary for 4k? or will a 4080 or new gen equivalent be just fine?
If you can wait, I would wait for the 5080, launching in early Jan. It'll be cheaper than the 4090 while having similar performance. Also it might have better RT performance or some new feature (DLSS 4?). It's important to understand that no current graphics card can run 4K native reliably, even without RT. From this perspective you'll often be using DLSS Quality, which is a 1440p input resolution. If you look at the 4080 Super's performance at 1440p, that's a good proxy generally speaking for its performance at 4K with DLSS. The only games the 4080 Super isn't great for are the path traced ones, like Cyberpunk 2077, Alan Wake 2, or Outlaws. If you really want to play those games at the highest possible quality settings, you'll definitely want a 4090. Otherwise the 4080 Super will get the job done.
https://youtu.be/-HQPkXGvmA0?si=98vTXLiIdkW7FYbT&t=1679

>>102272849
Radeon is a non starter unless you play a very specific game or have a specific use case. You're forced to use FSR for 4K, which is garbage. You don't save yourself $50 to lose DLSS. And of course there's all the other compounding problems with Radeon
https://www.youtube.com/watch?v=HfI9rmTnxe8&t=331s
>>
>>102276079
i mean, if intel and amd implemented the feature on their GPUs, nvidia might have to follow suit right?
>>
>>102276218
I understand that, but the entire die gets covered with enough pressure and enough paste. What I don't get is why the excess of paste is so bad in this situation. Or why "spreading it thick is bad".
>>
>>102276270
>$300
Have you looked at 4090 pricing? You can't get one below $1700.
>>
>>102275984
>muh vram
Generally speaking the performance of the GPU is the limiting factor, not the VRAM. A good example of this is the recently released Space Marine 2, where the 4060 has no problems with VRAM but it needs variable resolution even at 1080p with optimized settings to maintain 60 FPS.
https://youtu.be/T9CwH7f1l1o?si=5d0jflyzT3JU3-4c&t=982
Space Marine 2 isn't nearly as demanding as UE5 games, where the 4060 needs reduced settings and DLSS Quality at 1080p to get a playable FPS.

>>102276325
OK, so a $400 premium. My point still stands. The 4090 is about 30% faster in PT workloads, so a linear price/perf increase from the 4080 Super would be a $1300 price point. The premium is what you pay over that. And of course the absolute performance of the 4090 (as opposed to its relative performance) makes PT playable while reconstructing to 4K, so its extra worth is limited to its relative performance increase over the 4080 Super.

>>102274687
>ten years
You're not getting 10 years out of any PC in my opinion. You'd be using a GTX 770 as the equivalent if you built a 10 year PC 10 years ago, and that would be complete garbage. You should expect to drop in an end of the line AM5 CPU (hopefully) and upgrade the graphics card in about two generation, so four to five years.

>to watch YT
The above is assuming you're gaming. If you're just watching youtube, you don't need a dedicated graphics card at all.
>>
>>102276523
>so its extra worth is limited to its relative performance increase
is not limited*
>>
>>102268288
you didn't follow the first two rules of intel. 1 Disabling all ecores 2. manually setting volts

>>102268427
if you use your monitor for gaymin only, calibration is for faggots, use wide gamut and as high gamma/blackeq + saturation as you can cope with
simple as

>>102268561
I was gonna say your router is shitty but if it doesn't happen with other devices....

>>102268735
that's a hyperchink brand
if you aren't buying hall sensor switches in current year you are fucking UP, mechanicals are DOA
Wooting 80HE, Razer huntsman v3 pro and so on

>>102268876
3 years means it has 0 warranty left
buying used gpus is like fucking non virgin women, do you really wanna do that?

>>102269268
b760m riptide is the only good one because a 12400f that is a 200 dollar bundle and ocs to 5.2ghz because the external clock gen on the mobo along with some 6000-6400 ddr5 ram and you are as good as you are gonna get for that and gives you free 12900k like performance

>>102272790
first rule of 4k, you are now always locked into buying the **90-series from Jensen, forever, since AMD does not compete
but you should honestly hzmaxx instead because we got the 480hz 1440p which unironically super seeds the 4k 240hz's

>>102273218
you know you can have a higher pixel panel and not run it at native or use upscaling too
if you are an ultra tryhard faggot you can use integerscaled 720p at 1440p if you want 1000fps in your esports

>>102273676
enjoy dealing with subhumans daily

>>102275326
>0% of 4090 owners don't actually know anything
maybe they spent all their time earning money instead of researching the best value per fps card
what is a better allocation of time?

>>102276195
first rule of thermal paste application, the universal correct application is a full spread, due to different sized chips underneath a ihs, viscosity of the paste and so on
>>
>>102274057
You don't need a better CPU if you're only buying a $300 GPU. If you're buying new, your only options are the 4060 and the 3060 12 GB. Which one you buy depends on what your priorities are. A problem with your setup is that you only have PCIe 3.0, which can cause issues with the 4060. A good option might be picking up a used 3060 Ti or 3070. Maybe a 3080 would be in your budget if you know you don't need a new CPU

>>102273692
>Any games where ReBAR actually makes a difference on nvidia GPUs?
Yes, IIRC the Dead Space remake was one.

>>102273218
1440p DLSS Performance looks better than 1080p DLSS Quality, and the performance requirements are the same. Always better to buy a 1440p over 1080p if you're having to buy a new monitor. The monitor is one of the longest lasting parts of your system; you can easily grow into it. And of course older games will run no problem even on relatively weak hardware.
>>
>>102272898
>7900XT is perfectly fine for couple more years even at 4k
It's far too weak, but the XTX he's considering upgrading to is also too weak. Clearly he's thinking about upgrading because the XT he has isn't good enough

>>102272135
3080. The 3080 Ti is a huge power hog

>>102271626
A cheap B550, preferably with Wifi/BT for wireless controllers or whatever
https://pcpartpicker.com/products/motherboard/#xcx=0&c=145&V=9000&sort=price&X=0,14451

>>102269783
Not worth buying SATA SSDs new. Use a PCIe expansion card for NVMe SSDs if possible.
>>
oh look more games nobody heard of
>>
>>102276523
>>muh vram
if only nvidia didn't cheap out on it you wouldn't have to repeat this multiple times per thread
why is green vram so much more expensive than vram?
>>
File: airflow.png (208 KB, 436x468)
208 KB
208 KB PNG
My GPU takes a shit if my case fans are just a little bit misconfigured. I don't know how it works but basically I have to go high RPM with these fuckers to avoid listening to some annoying humming at certain RPM (in some of them at 900RPM, some of them at 1100). Happens with all of them (even the noctua one) a bit inconsistently. No silence because if these fans drop just a bit below that RPM it's not good enough for the GPU and fans ramp up a lot quicker. I have a noctua cooler (NH-D15S), 2 Arctic P14 PWM intakes (front) and 1 Arctic P12 PWM exhaust (rear). Pic related. There's one slot for a 120mm (and I think 140mm?) fan on the top but I don't know if that'll end up adding even more noise as an exhaust.

Any ideas? Could I fix that humming somehow?
>>
>>102269192
DLSS is massively better than TAA and FSR in Starfield, so some Geforce GPU. With settings optimization you might be able to get away with a used 3070 if you don't care about image quality too much, otherwise a used 3080.

>>102269115
The 4080 is better than the 3090 Ti, retard.

>>102268654
That type of VRAM chart is not an argument
https://www.youtube.com/watch?v=foXX41XkT_E&t=7333s
>>
File: 1428183251790.jpg (37 KB, 269x259)
37 KB
37 KB JPG
>soifield
>>
should i buy a 4080 super? and should i just get a 7800X3D?
>>
>>102276769
>repeat this multiple times
Because AMDrones infest this this thread. /pcbg/ used to be an AMD safe space, and old trends die hard. VRAM is the only number where Radeon has a (generally useless) bigger number

>why is green vram so much more expensive than vram?
Who fucking cares? VRAM != performance. In workloads that actually need VRAM like rendering or AI, a 3060 is better than Radeon's best in terms of performance

>cheap out
I agree that Nvidia is somewhat stingy with VRAM, but if you look at e.g., the 4060 performance tier (historically the 2080 Super, 3060 Ti, and now the 4060), they've all got 8 GB. It's good to give developers a stable target, and it's also good not to hang your previous customers out to dry. Nvidia could have given that class 12 GB in my opinion (an extension from the 1080 Ti's 11 GB, which is in that performance class as well), but the cards generally don't need more VRAM considering their performance. Additionally, if Nvidia had given those cards 12 GB, you'd be whining about 16 GB anyway, and developers would be just as bad at optimization while filling whatever VRAM cache is given to them.
>>
the harder you try the more I am likely to buy amd just so you get a stroke
>>
>>102276523
>local retardautist shillbot
didn't read
>>
>>102276956
I wish you would. It would serve you right
>>
>>102276523
>My point still stands
Your employer's point
>>
>>102276916
It's so funny seeing every 6000 series AMD card be useless despite 16GB of "Muh VRAM"
>>
>>102276844
answer
>>
>>102276916
Yeah 16 is mostly a meme on the 4060 but there is no reason for cards as powerful as the 4070 is still on 12GB AND most of the 50 series will be as well according to leaks. Post the SW Outlaws VRAM usage.
If the 4070 had 16GB AMD would be forced out of the midrange market too.
>>
>>102276523
>>102276554
>>102276653
>>102276751
>>102276800
>>102276916
>>102276967
>>102277113
the first thing this guy does when he wakes up is not take a shower, not brush his teeth, not even wipe his ass, he just comes here to posts these walls of texts combined with microsoft paint scribble images.
>>
File: Projection.png (5 KB, 198x168)
5 KB
5 KB PNG
>>102277207
Sorry people disagree with you but try not to project so hard.
>>
SFFchads, are there any ATX3.x SFX PSUs to even consider other than the 2024 version of the Corsair SF750?
I try not to give Corsair any money if I can avoid it but their SFX PSUs never had much competition.
>>
>>102277152
>Post the SW Outlaws VRAM usage.
Literally the only game
>>
>>102277152
NTA, and I completely agree with you. The trends clearly show VRAM utilization increasing over time in modern AAA titles. Nvidia has made it clear that if you want to game at 4K, you’ll need to buy their highest range cards such as a 4090.

They will continue to cripple each new generation with inadequate VRAM. The 3000 series would have been perfect if not for their greed. Nvidia is doing everything they can to avoid another Pascal situation.
>>
Remember, if it is just one example in favour on nvidia, it is irrefutable proof of the future. If it is not in favour of nivida it is a hack dev job not worth discussing as it is completely irrelevant. Once anons understand this things will be a lot smoother around here.
>>
>>102277284
The only game so far
>>
File: 1720189715652830.jpg (21 KB, 612x444)
21 KB
21 KB JPG
Is this a good build?
>>
>>102277231
4chan is the easiest website to inspect element
>>
File: FxjJhsPXsAAd6G_.jpg (599 KB, 3840x2160)
599 KB
599 KB JPG
>>102277323
>if it is just one example in favour on nvidia, it is irrefutable proof of the future
Quite the contrary, there are many more examples in favor of Geforce than Radeon, as every new game needs DLSS, whereas very few need VRAM.

>it is a hack dev job
>amd sponsored flopspoken
Yes it is. An ugly, technical abomination
>>
>>102276844
answer
>>
>>102277433
Nvidia GPUs are huge scam and full of drivers issues. Watch someone reply to me mad at this fact and say shit like "moronic liar" like a stupid bot
>>
>>102277376
For what purpose?
>>
>>102277456
a good, a circlejerk post. im glad i waited 50 minutes for this answer
>>
>>102277507
You waited 50 minutes for a bot reply?
>>
>>102277422
Any idea why that happened?
Obviously that's not just a lack of VRAM.
>>
>>102277422
To be fair, I just bought an nvidia GPU with 12GB of VRAM and the game included with it was SW Outlaws. It's kinda funny. I have yet to redeem that piece of shit.
>>
>>102276844
If you absolutely need a PC now, yeah looks good if your main goal is gaming. I'd wait for 9800X3D and next-gen Nvidia coming out at the beginning of next year.
>>
is radeon really as bad as anons itt make it look? I feel like it's impossible to get the same longevity and value out of nvidia. A 6750 beats a 4060 every day for less money, new.

But oh no, no DLSS or AV1 encoding! (Does literally anyone actually care? Is anyone actually ok with smudged AI frames and a content slop creator?)

The only reason I have nvidia right now is CAD software but radeon cards have basically caught up, and for the 5% performance delta between them I get better linux support and don't have to create an account to log into the software that allows me to use all the features of my gpu.
>>
RTX 4070 TI Super here. All my games look blurry. How to fix?
>>
>>102277772
stop cumming on your monitor
>>
>>102277772
change monitor
>>
>>102277777
funny joke but seriously how do I fix? im doing max settings in cyberpunk and game still looks blurry compared to wow.
>>
>>102277772
Use native instead of DLSS, FSR or XeSS
>>
>>102277772
Are you playing games with forced TAA by chance? Use DLDSR at x2.25 + DLSS Balanced.
>>
>>102277795
turn off motion blur, use different antialiasing, don't generate sloppy frames, add a sharpening filter via nvidia overlay, make sure your monitor is running at its highest refresh rate and resolution
>>
>>102269079
>5k
>new cutting edge technology
HA HA HA
>>102269026
>noooo you cant criticize an obsolete product like that
>>
>>102277790
its already a 27 inch 1440p. do I really need 4k for these new games?

>>102277820
not using upscaling and its still blurry.

>>102277823
what is dldsr?

>>102277872
ok
>>
>>102277704
>Gaming
Allegedly AMDippin in competive titles (DOA), Problems in older gaymes and jank, good for triple-a singleplayer slop
>Software
DOA for content creators, nobody buys them, good for loonix, I guess?!
>don't have to create an account to log into the software that allows me to use all the features of my gpu.
the new nvidia app does not have login
>A 6750 beats a 4060 every day for less money, new.
USA market? Radeon in my jewropean country has historically been uncompetitive every single time I check prices, it's real bad here almost always, nothing like murica, really don't understand anything tqbh

nvidia has specific jank too like the DSC alt tab jank bug which doesn't exist on AYYMD cards right now

tldr ymmv faafo
>>
>>102277704
>don't have to create an account to log into the software that allows me to use all the features of my gpu
Just nuke that shit and use nvcleaninstall
>>
>>102277704
>I feel like it's impossible to get the same longevity and value out of nvidia.
Quite the contrary. 3070 buyers still have relatively good image quality, while 6700 XT buyers get fucked over by FSR in nearly every game
>>
File: TaaVSNativeRIP.webm (217 KB, 864x1000)
217 KB
217 KB WEBM
>>102277878
> not using upscaling and its still blurry.
Disable TAA and use native instead
>>
redpill me on the 7800X3D, i havent looked at pcs in a while. why is it the ultimate cpu?
>>
ah yes the mid-range card with 14900 and ultra xdddd
>>
>>102277951
i disable taa in cyberpunk and now everything is fizzling like crazy like lights having a seizure
>>
>>102277878
>DLDSR
It basically does what DLSS does but in reverse. Downscales from a higher resolution. So using DLDSR 2.25x and then using DLSS balanced mode brings it back down to native from a higher resolution basically being a form of super-sampling. and removing a lot of TAA nasties when you can't disable it.
>>
>>102277975
Fast as fuck
>>
>>102278015
Use another form of AA now like DLAA, DLSS, or SMAA. You can't just disable AA, you'll get jaggies.
>>
>>102277975
>why is it the ultimate cpu?
Competitors exploded
>>
I could use some help. My pc has been crashing while playing HuntShowdown but when i do stress tests/benchmarks i get no crashes or issues. PC crashes but lights on my mouse, keyboard and gpu stays on, same goes for the fans which keep spinning. Temps seem fine to me while playing, the gpu is in the low 80s and the cpu is around the low to mid 60s. Drivers are all up to date other than the Bios which i havent updated yet.

specs:
>ryzen 7 7800x3d
>xfx 6950xt
>tuf gaming b650 plus
>seasonic focus gx 1000w gold
>corsair vengeance 32gb ddr5 running at 5200mhz
>>
>>102278065
Get a real GPU (4070S/4080S/4090)
>>
>>102278078
>4070S
>real
rofl
>>
>>102278078
this
>>
>>102277975
>why is it the ultimate cpu?
built and designed for youtube benchmarkers and retards (aka consoomers)
in short it requires no effort and tuning, you can run ass shit ram settings, cpu settings, everything and it will just be hard carried by the cache amount on the cpu and "just work"
non x3d or intel requires more effort to get fast
so if you are not gonna sit in a bios screen, it will be faster out of the box
>>
>>102278065
did you mess with anything in the bios. Undervolt?

>>102278078
is it nvidia shill payday? This thread is borderline an ad for rtx
>>
Hey faggots are there any good monitor arms/desk mounts that you can recommend?

I'm not spending $150 on a fucking stick of metal to hold a screen
>>
>>102278194
VIVO STAND-V100H
>>
>>102278114
>built and designed for youtube benchmarkers and retards
That is one of the most retarded posts in this thread and a colossal misunderstanding of what the cpu (well, v cache) actually is.
>>
>>102278194
ARCTIC X1-3D
>>
>>102278291
>>102278291
>>102278291
>>
>>102278247
>>102278229
I just said I'm not spending 150 on a fucking arm
>>
>>102278309
>>102278309
>>102278309
>>
Your thread is late, delete your thread
>>
>>102268017
What intel CPUs were fucked and getting damaged each time they crashed?
>>
File: 4k native vs taa.webm (3.74 MB, 1920x2022)
3.74 MB
3.74 MB WEBM
>>102277820
>>102277951
>native
Complete garbage.

>>102277878
>what is dldsr
A setting you can enable in the Nvidia Control Panel that downsamples from a higher resolution to your monitor. It can help clear up some of the blurriness from TAA. It'll provide clarity improvements at a commensurate performance cost. Often it can be used in conjunction with DLSS for extremely cheap supersampling. One note about DLDSR is that it has to be use with exclusive fullscreen, not borderless. Also obviously use DLAA when available
>>
>>102278174
Didnt do anything in the bios other than set my ram to 5200mhz.

Been doing OCCT tests last few hours and no crashes
>>
>>102275226
Thanks for the input! What's the problem with Noctua? Or are you saying that a cheaper option will already make my computer silent?

>>102276523
I'd like to have as little upgrading to do over the next ten years, if I only have to upgrade the graphics card and maybe add some ram that's fine.

>>102275119
Thanks! What's the matter with the x2 link slot? I don't know enough to understand if that's bad or what.
>>
>>102279639
>What's the problem with Noctua?
Waaaaaaaaaay too expensive when you can get the same performance (including noise) from coolers less than half the cost.
>>
>>102278065
Try holding shift+control+alt+B when it happens to reset the graphics driver
You could also try AMD's AFMF2 technical preview driver which is unironically more stable for me on my 6800XT than the main branch drivers
>>
>>102278174
The threads for like the past week that I've noticed have had heavily shilling for the 4070 series by dlss-kun, who is very likely a nvidia employee.
>>
>>102272736
They come with pads already on the heatsinks



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.