Will I be able to get a 9800X3D before the American civil war starts?

>Nvidia has broken drivers on Windows now
time to switch to AMD

Filler tile, more like glownigger tile, maybe that's where the latency really happens.

Wtf my pc just made a high pitched beep sound
Scared the shit out of me

File: image.png (669 KB, 1002x1052)

669 KB PNG

How do I smart self test an nvme in windows?
Also How do I stop buying meme shit I don't need?

File: Screenshot_2024-10-25_13-03-46.png (245 KB, 808x741)

245 KB PNG

thoughts on this AI build?

File: Lamptron HC060.jpg (42 KB, 556x556)

42 KB JPG

>>102971575
Are there any other mini monitors that fit in 2 5.25' bays? Pic related can't be bought anymore.

>>102971700
unironically get an incel chip for AI

>>102971700
overspecced meme cooler on an 80 watt chip
retarded memory, either go 6000CL30 or all the way to 8000 (pointless on X3D)

>>102971700
Base 3090 is an abortion of a card that will cook and/or fry itself, Ti or bust.

>>102971711
>incel
which gen? should i play it safe with a 12th gen?

>>102971754
thanks for the headsup

>>102971585
>before the American civil war starts?
sorry to burst your bubble but Americans are pussies that wouldn't do anything that might make them lose their depressing 9-5 slave job

What's the use case for "upgrading" to AM5 coming from AM4? What's the point? What's the advantage?

Please don't tell me it's to play some AAA unreal engine tranny schlock.

File: 1702053040997706.png (85 KB, 560x1210)

85 KB PNG

>>102971786
15th.

>>102971826
You are mentally ill and will be unhappy with either platform.

>>102971838
thanks anon, looks like the 265k will be the best value

File: 1729277928700160.jpg (46 KB, 563x398)

46 KB JPG

X3D GODS

>>102971880
it will suck at games tho, is the caveat to throw int

> by suck I mean it will be 12/11600k level

>>102971934
that's ok i stopped gaming after battlefield 2, this rig is strictly business

>>102971839
Not what I asked.

>>102971838
>90 seconds per image
lol

Is PCIE 3.0 w/ a 3700X gonna cause a performance penalty for a 4090? Hesitant to upgrade in-platform considering this.

Do i AMD or do i Intel?

>>102971826
nothing, unless you want high speed/capacity ram

File: Screenshot.png (113 KB, 500x500)

113 KB PNG

How loud are SFC psus? I had an corsair 850w in sight but it went into discount and disappeared overnight. But there is an SFX with same wattage available and it's also platinum. My estimated draw is ~400W

File: akamaihd.jpg (55 KB, 512x512)

55 KB JPG

When I was a teen I used to window-shop for computer hardware that I couldn't afford. Now I could afford anything I want but I just don't care anymore. I've lost it, anons. It's over.

>>102971575
>not using bare copper traces for interconnects
then how the fuck does intel have a fuckton more latency

>>102971804
Dont listen to him, literally the most popular cheap homebrew AI farms right now consists of multiple used 3090s

ask the AI threads

File: GLgdPFtWMAAQ20B.jpg (52 KB, 590x674)

52 KB JPG

> PC Build
- B450 AORUS Pro Wifi (AM4)
- AMD Ryzen 7 2700X
- RTX 2080
- 16GB DDR4 RAM (2933MHz)
- 2x 1080p Monitor
- Moderate budget, Canada

Stay on AM4? Perhaps 5700X3D and what, 4070 TI Super? Maybe 32GB RAM as well...?

>>102972006
This is running the inference purely on CPU and for the largest language models where you can only partially offload to GPU, that becomes useful. Most places don't know how to test AI yet so usually you have to rely on ancedotal evidence posts online to get an idea of how things perform. That being said, Stable Diffusion is pretty good at telling you how something does in other areas of computing AI except for the cutting edge heavy usecases because it's not a transformer model, at least not unless you are running v3 or above which would take way longer to benchmark on CPU.

>>102971838
>using CPU for stable diffusion
this chart is so useless... Even a 1060 would be multiple times faster than using a CPU

>>102971711
>>102971838
intelshills sure are shameless
nobody uses CPU for actual AI inferencing unless you are really desperate

>>102972253
>>102972233
you guys are the real shills, the guy asked for input on an AI ONLY BUILD NOT GAMES
and I gave him the best options, stop brand fagging

>I'm ordering a 9950x so fuck you

>>102971971
you are getting scammed if it's for AI business. CPU doesn't matter much, it's the GPU that you should invest in, the more VRAM the better.

>>102972268
CPU literally doesnt matter if he's serious with AI, he can pick any CPU he wants. Stop scamming him into picking a newly released overpriced piece of shit.

>>102972268
Holy obsessed you dont need the latest top of the line CPU for AI, AMD or Intel. You'd rather use your money into buying more VRAM

File: detailed-performance-numb(...).png (46 KB, 1080x810)

46 KB PNG

>>102972253
You do with partial offloading of a model which has to be done for big sizes like 70 billion parameter models as I said in >>102972203 but to get accuracy on that usecase, you would need to know how to use GGUFs and llama.cpp which is out of the scope of most people's repertoire and it isn't something I would expect a benchmarker to know how to run until it is easy to do which it still isn't easy to do.

>>102972278
>>102972290

>gave him an option thats cheaper than what he originally chose

WHAT ARE YOU DOING STOP SCAMMING HIM
you guys are obsessed

>>102971687
The Samsung Magician app has a SMART testing tool.

>>102972294
He's using it for business so why would he even use llama.cpp
Nobody in business would let you use CPU for inferencing

>>102972298
He literally suggested a 12th gen and you told him to pick 15th, fucking retard

Stop scamming people

>>102972178
And some of those people ended up with dead 3090s because of its inadequate power delivery design.
There's a reason why NVIDIA changed the PCB layout for the 3090Ti.
The 3090 Ti's power delivery system is closer to RTX 40 series card, you can transplant a 4090 GPU on a 3090 Ti PCB.

Link below is a guy who lost his 3090 mostly running AI loads.
https://desuarchive.org/g/thread/100967012/#100973294

>>102972298
>12th gen ok?
>noooo buy the LATEST
retard, 12th gen is perfectly fine, again, CPU doesn't matter really in AI

>>102972331
Just one person, I bet you could find another single person with a busted 3070 or 4090 over AI.

I bought mine USED probably even came from a mining rig two years ago yet it's still alive and running right now

>>102972310
You don't know that, he may be a part time business owner and would like to do that. People are already deploying llama.cpp in the field, the project has a company behind it.
https://ggml.ai
We know that there are commercial backers because a lot of improvement patches are withheld from the project as pull requests until several months later and there are custom builds out there.

About to install loonix on an old PC I have, wish me luck

>>102972350
Define old. Linux's support for anything from the late 90s to the present is really good especially with all the improvements unless your hardware is something that they have dropped support for like a 386.

>>102972343
It's still a dead end tech because GPUs are still the most cost efficient part to do your AI inferencing and training in the long run and they keep on getting better too, not like they'll be stuck while that improves.

>>102972372
i5-7500T plus Vega64 with silent bios (200w)

>>102972342
>Just one person,
Do a Google search of dead 3090s moron, there are plenty of dead Strixes, Suprim, and AORUS cards. Those are the overbuilt ones.

There is a fucking reason NVIDIA changed the power design for the 3090Ti.

>>102972377
Yeah but you can't get away with running models all in VRAM on GPUs and futureproofing it unless you massively overbuild and have H100 server(s). Having a good CPU, RAM and GPUs is essential when offloading which is inevitable.

>>102972082
All novelty wears off when you are an old boomer, women, hardware, everything
Either start taking hard drugs or seeking deeper meaning
Or cope or rope, idk

>>102972038
That slow ass CPU is a problem

>>102972396
Yeah it will work well. I had a Vega64 I used on Linux. Make sure to get familiar with how to control power on Linux to limit power and undervolt.

>>102972424
>to limit power and undervolt
meh, will just use the silent bios that limits it at 200 watts

>>102972190
Yeah that's the value play. Assuming you actually want an upgrade.

File: file.png (579 KB, 1200x675)

579 KB PNG

>>102971575
software updates coming in two (2) weeks to solve all the performance issues in gaming and overtake AMD

File: 1727677925286558.png (109 KB, 645x729)

109 KB PNG

>bought a used 13600k because it was dirt cheap before even hearing about these instability issues
Am I fucked? Is the chip already degraded? supposedly they do stress tests before selling and it was fine

>>102972561
you should be fine if it was from a reputable seller

>>102972561
well its dirt cheap so who cares
ask intel for a new one it causes issues

>>102972561
It's fine

File: 2514310411.jpg (18 KB, 236x354)

18 KB JPG

>>102972298
Hi Pat, are you getting ready to launch Battlemage soon you dirty white snow nigger.

>>102972634
NO go buy used AMD?NVIDIA

File: 71YwpUpu9VL._SL1500_.jpg (134 KB, 1500x1500)

134 KB JPG

>>102971575
Did Arctic update their P12 MAX again? The hub is smaller and they have curved struts again like on the original P12.

>>102972561
13600K doesn't boost high enough to ever request over 1.55V.
This is the reason why the issues mostly affected i9s.

File: 67-500x500.png (207 KB, 500x500)

207 KB PNG

is it a good idea to reuse pic related (corsair 300R) as the case for a new build today?

File: 613cvjbh5tL._SL1500_.jpg (80 KB, 1500x1500)

80 KB JPG

>>102972652
The white version seems to be still the old one. They both come with a FDB unlike the very first version that uses DBB.

>>102972561
>normie tier fretting
do you even know about the microcode update

I'll buy that chip without the compute tile for half the price

>>102972504
We know already Arrow Lake's biggest issue is the Ring Bus clock speed and der8auer proved it. It's not as good as if you can get a straight overclock but the difference in CS2 is way too much for it not to be the culprit by just raising it 400 Mhz. I'm not sure why it isn't higher out of the box given you can get that much increase out of it. To keep power consumption in line?
https://youtu.be/9xgN46TBBGg?t=1343

>>102972652
>>102972666
It's probably cost-cutting revisions at this point.

File: 51ld6RR8IrL._AC_UF1000,10(...).jpg (52 KB, 949x1000)

52 KB JPG

I've been completely checked out gaming benchmarks for a while now anybody have a clue how the 5600X is holding up in 2024?
It's my current CPU and I've been half considering an upgrade to 5700X3D and half not, because at this point it almost feels like a lateral upgrade when I really should consider moving up to AM5

>>102972660
yeah that looks fine, as long as your parts fit.

Is waiting for RTX 5000s and AMD's RX 8000s series worth it over buying now?

>>102972771
Well you already have it so you can decide for yourself if the performance is satisfactory. Whether the 5700X3D is worthwhile would depend on the games you play, your video card, and the total price i.e. less any proceeds from the sale of your 5600X.

>>102972820
I mean if you want to spend a lot of money on them

>>102972840
I have noticed a couple of games where my GPU utilization isn't 99% and I'm still dropping below 60 fps, it's hard to tell today with how shit optimization tends to be but I suspect that it may be on the cpu

>>102972771
How about the 7600x?

>>102972878
Alright you've got some bottlenecking somewhere (as always). The next step would be to verify it (Rivatuner) and look up benchmarks for those games to see if X3D gives an uplift. Like you say there's other things that can cause a "non-GPU" bind so it pays to check.

I'm so fucking pissed at NVIDIA for delaying the 5000 series until January, been wanting to buy an xx90 card for at least half a year now. I'm getting old for this crap.

>>102972923
yeah because the 3 games with good RT really need the extra fps over existing hardware

>>102972855
Anywhere between 1500 to 2000

I mean, the new series gotta have better value, right?

Just boughted new graphics card. Plugged in and got artifacts on the desktop. Replugged to a different monitor and same, but only for a few seconds.
No artifacts since.

Got 3 months guarantee. Should I just return or is it ok?
It's used rx 6600 tx and it was only 160$

>>102972966
>boughted
you should keep it

>>102972923
Unless you churn slop for a living, you'd have to be deranged to buy an njudea flagship on release.

>>102972983
I deserve to enjoy the endgame flagship performance of this tech tree before I get too old. I only buy GPUs once every few years.

At most I might go for a used 4090 if there are any good deals considering a 5090 will likely set back $3500 on launch.

>>102971826
I'm only "upgrading" to AM5 because I want to move my AM4 rig to my vacation home and hence need another PC at home. But the one point I really like with my new motherboard (Asrock X870E Nova) compared to my old one (MSI X370 M7) are the many M.2 slots, five in total. Using dedicated drives for multiple concurrent processes has been my obsession since I experienced how much faster everything gets done like this.

File: 1729894077254902355509098(...).jpg (246 KB, 1600x1600)

246 KB JPG

Oh no! You took off your CPU cooler to access a very hard to reach fan header and you realized you're all out of thermal paste!

You look around the house for a substitute and pick...?

>>102973003
Oh, it's for a midlife crisis build.
That's somewhat understandable.

>>102973014
used condom

Best price to perf GPUs for prompting stuff in SD right now?

I love amazon
>buy 7950x because i wanna tinker
>9800x3d comes out in 2 weeks
>return 7950x for full refund
>get 9800x3d from newegg with even longer return window

I WON

>>102972960
>I mean, the new series gotta have better value, right?
lmao what? no

older hardware will always have a better *value* and people who really value performance will spend a fortune to get 5% faster

File: file.png (195 KB, 1117x637)

195 KB PNG

>>102972038
you will lose less than 3 fps
the bottleneck is your cpu, its very slow, you wont notice as many problems at 4k but at some point you should upgrade to an am4 8 core x3d, keeping your motherboard and ram, sell your 3700x

>>102972190
> B450 AORUS Pro Wifi
Just one M.2 slot? This would be the only thing bothering me.

>>102972920
That would require a new motherboard

How will they price the 9800X3D?
Not a lot of room left between the 7800X3D and the 7950X3D at this point.

>>102972771
I went from a 5600x to a 5800X3D a few years back. The x3d is clearly better of course, but meh. not a huge upgrade.

File: sparkys a boy dog.png (2.22 MB, 1920x1440)

2.22 MB PNG

>>102972082
anon, there comes a time in a young anons life when things change ....

>it's just another machine

>>102972082
sounds like you have some other issues if this is affecting your other interests as well

>>102973175
iunno as you age, things just aren't as exciting. My first build I was bursting with excitement for my shit to show up, and I bought on black friday too so it took 5eva for everything to get there. Could NOT stop thinking about it as I waited weeks for the damn case and mobo to show up so I could finally build.

>>102973200
i still feel like that will all of my hobbies
coming up on ~ 13 years of building pcs/restoring cars
I enjoy everybit of my life that I can anon

>>102972652
I bought a 5-pack two weeks ago and mine still have straight struts
Could just be older stock though

>>102973026
lol, i guess you saw the episode of ligma tech tips where he cummed on a heatsink and said it was better than arctic silver?

>>102973127
I mean you did say you were considering an am5 cpu

File: stats riva.jpg (58 KB, 415x212)

58 KB JPG

>>102972922
anything that sticks out here?

File: 513sOwKiA2L._AC_UF894,100(...).jpg (49 KB, 823x1000)

49 KB JPG

Name some cheap pressure optimized fans

They can't be Arctic because I've already confirmed they are basically incompatible with my case, very loud whine when mounted to the front panel.

>>102973447
Fractal case?

File: {1094355B-FDAA-49EE-9F9B-(...).png (112 KB, 1500x594)

112 KB PNG

>>102972300
First, it requires creating a partition first, for absolutely no reason.
Second, it only works with samsung nvme driver. The same nvme driver that is known to cause data corruption if installed on nvme system drive and should only be sideloaded during clean windows install or if samsung nvme drive is secondary.
Guess I should just run some live linux and

nvme-cli

it.

File: unnamed.jpg (65 KB, 512x512)

65 KB JPG

>>102973458
Anidees raider XL
Wanted to add a 360mm radiator with push pull

>4090
>7800X3D
>32GB 6400CL32
>*boots GTA Vice City*
Yep yep

>>102973544
>>5090 (RADEON HD 6950)
>>7950x
>>48GB 8200CL38 (RIP ARL)
>> boots MGS4 in RPCS3
>Yep yep

>>102973544
fake because dx9 shit is broken on modern nvidia gpus. wheres your dxvk

>>102973142
7800x3d probs won't be available for much longer. By all accounts, recent production has been quite low, to the point of demand outstripping supply enough to cause its price to go up. Once 9800x3d is out, there will be even less reason for AMD to keep producing 7800x3d.

>>102973617
dx8*

>>102973617
NTA but I can boot even dx8 games without issues on my ampere gpu.

Maybe not the right thread but does anyone have ideas on how to clamp a monitor arm onto a desk where the bottom side of the desk is not an even surface

>>102973447
> cheap, silent, effective
you can only pick two of these

>>102973638
it will boot but not work correctly unless nvidia applied specific fixes to it like sonic adventure.

File: 1729898214681630344749711(...).jpg (106 KB, 720x720)

106 KB JPG

>>102973447
big phat phan

File: where the hell am i.jpg (35 KB, 896x188)

35 KB JPG

>>102973660
yes but a picture of the problem might further help anons visualize suggested solutions

>if only there was a magical place where you could post such a picture...

>>102971700
>CPU
Your solid, could do better in the foreseeable future but now the best bang for your buck, and what a buck

>Storage
The Samsung 990 has noted degradation issues and is also needlessly more expensive

also I don't any other storage listed

Have an SSD for a boot drive and essentials, separate SSD for work, and HDD for mass storage

>Memory
If you want speed, single rank ram tops out at 24 GB's per DIMM, make sure you can use the most of the available channel speeds on your board by doing some math

>Board
Meme, you can do better for way less and look cooler to all the other kids on the block

>Power
Good and future proof from a solid source

>Cooling
Water cooling is best with your projected work, but air cooling works to an extent, you will need more fans

>>102973704
high air volume != high air pressure

File: 1000hours.png (11 KB, 1152x648)

11 KB PNG

>>102973720
An artist's approximation of the desk and the clamp

File: really now.jpg (36 KB, 1152x648)

36 KB JPG

>>102973816
the technical term for it is "use a stick of wood" or something

>done with my CAD for the day

>>102973853
But where would I go about finding said stick of wood of the exact thickness

>>102973884
Get some wood and a saw from some DIY store.

>>102973884
Take some measurements of the thickness and width and go to your nearest Dick's Wood and Lumber store
Or steal some lumber lying around a construction site and cut and sand it down

>>102973019
>a midlife crisis build
Probably no better way to describe it.

File: thiswillhurtyou.jpg (2.3 MB, 2560x1920)

2.3 MB JPG

>>102973884
i personally use pic related when i need a precise /plank/, don't everyboddy haz one of these or maybe has a fren who haz one?

>measure it
>can be a hunk of solid plastic, metal, an old hard drive, mini cutting board, paperback book, crusty socks...
>you can even laminate solid cardboard to make a block
>pls no eating the glue

whatever can bear the pressure of the clamp

File: overtime.jpg (22 KB, 612x465)

22 KB JPG

>>102973660
>>102973853

also fren if the molding is "eased" (curved) on the inside where the clamp hits you might have to have two bits of whatever material, one to spread the load for the clamps movable jaw

>CAD is a lot of work

>>102972983
Does he really?
4090 got more expensive over time.
3090 price got inflated by crypto & scalpers, starting very shortly after release.
2080 Ti had a theoretical $999 MSRP, but models below $1,200 weren't available for long.

File: 11-994-093-V01.jpg (113 KB, 1280x1757)

113 KB JPG

Anyone got one of these?
My motherboard has a USB 3.2 Gen2x2 header. All the documentation for this says 10 Gbps but the cable shouldn't determine the speed right?

File: Screenshot 2024-10-08 at (...).png (109 KB, 503x249)

109 KB PNG

>>102972056
prease lespond

>>102974075
the wrong cable could affect the speed, so make sure the one you get is specc'd for the requirement.

>>102974078
go fanless if noise is the concern?

https://www.quietpc.com/powersupplies?sort=price_desc&cooling=fanless

>>102973014
Cooking oil

>>102973014
copper based anti-seize worked for me in the PIII days

File: [SubsPlease] Highspeed Et(...).jpg (487 KB, 1920x1080)

487 KB JPG

>>102971591
I have a 3080 my brother has 6800XT. Drivers are about on par for the last two years. AMD is amazing in games, almost zero crashing only issues were with AC Valhalla and Dead Space remake but I'm going to blame this on Ubishit and EA not AMD. Nvidia is a bit better outside games but they tend to have minor issues recently. Maybe it's worse on 4000 series.

None of them are really bad.

>>102972056
Louder than a beQuiet Pure Power 12 M 850W.

>>102974136
>Dead Space remake
I played it a little late but had no issues on 6700xt

>>102973513
>>102972300
never mind smartmontools on windows supports nvme

My brother is running a 1650 ti and I'm considering upgrading him to a 6650xt for about 200usd. for gaming purposes is he going to be noticing big differences?
I could also go up to a 7600 for a extra 50usd. His cpu is a Intel i5 10400. Idk if that will bottleneck him at all

>>102974136
I'm triggered by 3000 series. I had a used 3080 and it was hot & loud as fuck plus I had to flash the bios for rebar support. One of the worst gpus I ever owned user experience wise. The last AMD card I owned was a MSI 270x and it was quiet and performed well but the card started showing artifacts in the BIOS so its dead jim.

>>102974085
dammit, and Fractal doesn't make a faster one I forgot a link in my first post: https://www.fractal-design.com/products/accessories/connectivity/usb-c-10gbps-cable-model-d/black/

Anyone know alternatives?

File: 1424792332026.png (299 KB, 606x401)

299 KB PNG

>>102974098
>450€

>>102973723
>Meme, you can do better for way less and look cooler to all the other kids on the block
the fat guy from level1tech has highly positive things to say about ram performance
>fastest aida64 score he's seen
https://youtu.be/FgRp_u_6DWI?t=660
also seems fairly well constructed with an 8 layer pcb

>>102974238
>My brother is running a 1650 ti and I'm considering upgrading him to a 6650xt for about 200usd. for gaming purposes is he going to be noticing big differences?
It's basically double the performance. Should be a nice upgrade.
>I could also go up to a 7600 for a extra 50usd. His cpu is a Intel i5 10400. Idk if that will bottleneck him at all
I don't think a 6650XT is a strong enough card to be bottlenecked by the 6-core 10400. But you can use something like the GPUBusy stat in CapFrameX to evaluate if he is CPU bound after upgrading his GPU and if so by how much. See this:
https://www.youtube.com/watch?v=n6qAoDquTgo

>>102974470
Whoops wrong video link, see this:
https://www.youtube.com/watch?v=4QJglCSGEt4

>>102974454
not that anon, but at 7800x3d for AI/workstation is wrong anyway

So after having a mini-heartattack after my GPU started acting up after I upgraded my old monitor to 1440p/160Hz, I learned that Freesync apparently doesn't like Windowless mode in some games. Is there a fix for this, and is there any other shit I should be aware of that might cause issues now that I have a display from this decade?

File: mpv-shot0003.jpg (442 KB, 1290x718)

442 KB JPG

new buildzoid video just dropped
https://www.youtube.com/watch?v=Xcn_nvWGj7U
now you can watch this instead of asking about optimal AM5 RAM speeds for the millionth time

>>102974734
finally. i was wondering whether 2:3 fclk:mclk really mattered or not (it does slightly)

File: 1971891079932.jpg (239 KB, 1555x814)

239 KB JPG

intel bros I thought Arrow Lake would be the most powerful cpu ever made but...

>>102974673
>Freesync apparently doesn't like Windowless mode in some games
install windows 11 and enable "optimizations for windowed games" : this forces dxgi flip model giving your GPU control over the frame presentation and bypassing the desktop compositor (DWM) that is forcing vsync and fucking up with VRR
inject Special K (https://wiki.special-k.info/en/SpecialK/Tools) into your games and verify with its UI if it is indeed running an independent flip mode as well as VRR

on another note, win10 seems incredibly cucked performance wise compared to 24H2 and all the modern game-related updates MS made

>>102974779
it's not intel's fault that cdpr can't code for shit

>>102974779
How does this compare with the 9950x? That is it's competition.

People are quick to trash it, but the gaming performance seems similar to the 9950x, and it has slightly better mt. And slightly better power.

There is a weird level of hypocrisy from amd fans. It is as if they expect Intel to be top in both gaming and applications.

Certainly the performance will evolve with microcode and software patches. I see it as a necessary evil, they went to a new node using tsmc. If it is within a few % of amd 9950x, they are functionally equivalent processors and decision will come down to other factors.

I am guessing their strategy is to not announce their socket roadmap so that more people will buy this, yet any reasonable person wants to know if there will be more generations on this socket. It is reasonable that next generation will show further improvements. And it may be a quick generation.

>>102974821
so the new socket is another dead socket?

>>102974784
Honestly I don't really have that big a problem with Win 11 since I use it on my laptop but I don't want to deal with the hassle of upgrading unless it's really worth it, is that the only option?

>>102974889
the Special K tool by itself should be able to upgrade games to independent flip mode by itself even in Win10
try to inject it in some of your problematic games and see if it fixes your issue

I'd still suggest upgrading to windows 11 at some point, it's just better for gaming purposes now.

>>102974856
early longterm socket support announcement
>no ones buys 15th gen intel
Late longterm socket support announcement
> people buy 15th gen Intel and 16th and so on

> Intel design dept/overhead did not account for the terrible job of the marketing department


Core Ultra marking as game CPUs is a massive fail

Core Ultra should've been marketed as a work-only CPU /w overclocking for enthusiasts

All this chips are still sold out though
I will admit the 285k was a pure paper launch though

>>102974779
>5800x3D still that high
I might seriously wait for Zen 6 to drop at this point.

>>102974812
RED Engine is one of the best multithreaded game engines in the industry, it leaves Unreal Stutter 5 on the dust when it comes to multithreading.

>>102974956
RED is great but most of its multithreading "formula" is based on pre-existing systems

Core Ultra using P and E core without hyperthreading prob spooked it

>>102974929
>All this chips are still sold out though
>I will admit the 285k was a pure paper launch though
https://pcpartpicker.com/product/xNXV3C/intel-core-ultra-5-245k-42-ghz-14-core-processor-bx80768245k
https://pcpartpicker.com/product/mBsV3C/intel-core-ultra-7-265k-39-ghz-20-core-processor-bx80768265k
Yes, the 245K & 265K, which are supposed to be the volume chips are still in stock. I really can't see who would even buy the 285K though, there is almost nothing that it is the best option for over Intel 13/14 Gen or Zen 4/5. Frankly at the prices they set for the whole lineup and the new platform cost I don't see who Intel think is going to buy any of these chips.

>>102974982
i would say if you want a better IO its a better choice
Ranking by IO
1. Thread Ripper
2. Core Ultra
3. Ryzen

>z890 platform is great. Intel would be retarded to not put a "fixed" CPU series on it

>>102974946
>5800X3D
I'm unironically waiting for DDR6

>>102974821
The reality is that gaming is becoming its own workload divorced from a common computing workload. Gaming workloads, while they can be parallelized, for the most part are still single threaded latency sensitive applications. You solve that by using high clockspeeds balanced by good enough IPC, large caches and having lower latency in the chips where a Ryzen 5800X3D hits 2 of the 3 and the 9000X3D will finally allow overclocking and solve the clockspeed issue. General purpose is now sticking in a bunch of ASICs where possible to save time and power, more IPC and less clocks to save power, and as small of a die area as possible. That's why Apple Silicon has been successful for everything but gaming.

Currently have a 7900 XTX that I have had for over a year now. It started great but I get more and more driver timeout errors now. I've tried everything and nothing fixes it. I fixed it in fucking WoW by turning everything down to low but I didn't spend $1k on a GPU for fucking low settings in a 20 year old fucking game. And the driver crashes are getting worse with time. At first it would happen once a month. Now it happens if I have a game and a video on the other monitor at the same time. I reboot after the crash and my hardware acceleration is fucked and Windows is unusable without me booting into safe mode, wiping my drivers with DDU, and reinstalling them only to have the same problem next month. This time it was like 2 weeks between the drivers not just crashing but completely fucking themselves. Should I bother waiting for the 5000 series or just bit the bullet on a 4000 series and not deal with this retarded bullshit anymore?

File: 1729332842106646.jpg (278 KB, 1920x1080)

278 KB JPG

>>102974981
>PP-EEEEEEEE-PPPP-EEEEEEEE-PP
Nah it's no about hyperthreading
Intel was asking for scheduling issues by reordering their p and e cores in the most cursed way possible

>>102975142
do you DDU and install the drivers with the internet disconnected?

>still no 5500x3d news
budgetbros is it over?

>>102975157
I sure do. Googling the issue helped me disocver that for the games I play this is a known and long running issue with AMD. In WoW for example, this has been happening for years on DX12. If I want tom play WoW on my 7900 XTX I have to turn all settings to lowest and run in DX11 or it will regularly crash the drivers. I've had to do more driver uninstalls and reinstalls with DDU for this one GPU than I've done total for all computers for the last 20 years I've been building them. It's making me seriously consider giving Nvidia $1k right this minute for a 4080 S.

>>102975143
Scheduling isn't the main problem. Sure, there is more they can do since there is a 10k MIPs difference in 7zip compression on Linux vs Windows, for example, which is the difference between a 7900x and 9900x on Windows. But that isn't the main issue as far as gaming is concerned. It's the latency that is killing it with the ring bus, see >>102972734.

>>102975095
>The reality is that gaming is becoming its own workload divorced from a common computing workload. Gaming workloads, while they can be parallelized, for the most part are still single threaded latency sensitive applications. You solve that by using high clockspeeds balanced by good enough IPC, large caches and having lower latency in the chips where a Ryzen 5800X3D hits 2 of the 3 and the 9000X3D will finally allow overclocking and solve the clockspeed issue
This is true. The real fly in the ointment is that the sort of tasks that end users deal in which are general compute-heavy and benefit from high MT performance are "Walk-Away" scenarios where the improvements are not really appreciated. If a code compile project or video encode takes 1h 20m instead of 1h 28m I don't care because I'm still AFK making a sandwich during that time or have the process minimized in the background while I'm funposting on /g/. Gaming is about the only type of workload where performance gains are of a real-time feedback nature that you can immediately see and FEEL the difference. They are more mission-critical if I could use that phrase. The only other workload sorta like that might be audio production and editing where you can hear latency manifest if bad enough.

File: 36c6a03defbc758531376b713(...).jpg (66 KB, 720x734)

66 KB JPG

I should have buyed a card the past years, if I buy now I will regret it even more

>>102975202
5700x3d costs only 2x 5600x

>>102975238
I agree but the improvements are appreciated and companies notice, I have seen companies reduce their rack sizes due to Eypc or even condense down to a single server. I think you are looking for "user-visible" as far as workload scenarios are concerned, you can somewhat see the workload complete faster but the feedback is minimal to the end user. Gaming is one of the scenarios as with video editing/audio where the feedback is immediate and you can tell when things improve. The others would be any GUI tasks like your browser and etc.

How do I undo the thermal paste to slot in a new CPU

>>102972734
An Asus engineer said the weird layout >>102975143 limits the 285k to 4.2ghz ring and the 245k to 4.4ghz ring
I'd assume there are potato CPUs out there that wouldn't even do 4.0ghz stably so that's why intel ships them at 3.8ghz

>>102972734
>>102975237
Upping the Ring & D2D Clocks don't fully compensate the performance regression in ARL. der8auer himself mentioned even mentioned that after maxing them out to alleviate some of the latency issues he could not get ARL to beat or PAR with the 14900K in games.

>>102975142
this is not a driver issue
why do you think it's fixed if you turn everything down to low? AMD sold you defective hardware

>>102975274
isopropyl alcohol

>>102975272
> think you are looking for "user-visible" as far as workload scenarios are concerned, you can somewhat see the workload complete faster but the feedback is minimal to the end user. Gaming is one of the scenarios as with video editing/audio where the feedback is immediate and you can tell when things improve. The others would be any GUI tasks like your browser and etc.
>user-visible
That's it, thanks.

>>102974498
yes thank you, that topic has already been litigated by others. but i'm wondering if the longevity of the AM5 socket is compelling compared to the lifespan of the LGA1851 if i choose to upgrade in the future

>>102975274
To actually take the cooler off the CPU? I've never done it but my understanding from these threads is you want to get the CPU warm before you power off and then twist gently while you pull the cooler off.

>>102975288
A layout shouldn't really do that if the ring bus was fine but what do I know? I feel like there is going to be interest in a 285KS part where they can bin a top line CPU where the ring bus can be cranked higher. Intel should really improve upon that hopefully for Panther Lake.
>>102975294
Sure but it is the primary culprit given over 50% of the performance gap in the games he tested was clawed back with just the ring bus overclock and 80-90% with faster ring speeds on other components. I don't doubt again there are gains that can be had if they optimized their scheduling as you said though but latency is the primary issue for Arrow Lake in gaming.

>>102975330
>yes thank you, that topic has already been litigated by others. but i'm wondering if the longevity of the AM5 socket is compelling compared to the lifespan of the LGA1851 if i choose to upgrade in the future
If we have more "generations" in the future like Zen 4 to Zen 5 or worse RPL to ARL despite the new platform, then I'm going to start questioning the functional utility of a longer life sockets.

File: dNYvc2i5uE0.jpg (344 KB, 1200x1137)

344 KB JPG

Why are psus 80+ when the efficiency is 90%+?

>>102975208
well, that's unfortunate.
probably best to spend that 1k, send your card in for warranty and sell it on when you get it back.

>>102975366
There are serious intercore latency issues on windows but they seem fine on linux. Its like intel's engineers didnt even bother testing on windows at all

>>102975311
If I turn on AA with everything set to low it still crashes in WoW. This behavior is not replicated in any other game I have played. Driver crashes DO occur in just about every other game I play but they don't happen with this frequency. It's possible the GPU is just faulty but I don't think that's it. My theory is that the Adrenaline drivers flip out and crash when they detect any system instability, including XMP/EXPO. My driver issues have been there since I got the card but they started getting worse once I upgraded to faster RAM. I ran mem test and everything is okay with the sticks so I'm starting to think that AMD GPU drivers just don't play nice with XMP/EXPO. Because I am not the only person experiencing this exact issue with their AMD GPU in these same games. I also noticed that if I go into BIOS and disable XMP/EXPO that the crashes become much less frequent. That's nice and all but I didn't buy this hardware just to fucking gimp it.

>ordered everything from different places
>cooler turns up first
thanks amazong

>>102975371
im gonna be an adopter of LGA1851 via 265k

>will update if it makes me blow my brains out

>>102975412
I'd RMA it and request my money back.

>>102975392
It's not a hardware defect so it's not covered under warranty. It's a driver defect. I've checked and I'm definitely not covered.

File: Screenshot 2024-10-26 033143.png (6 KB, 300x108)

6 KB PNG

>>102975436
>unironically buying intel
blow your brains out
pic related

>>102975445
its a cool new architype with room for OC and lots of tinkering anon
not everything is about muh gaymes

bros i love occt

>>102975454
if you're an overclocker for fun then why are you getting a 265k? 285k is the most OCable you if you got a decent mobo and good CUDIMMs you should be over 1k

File: 121231312312.png (207 KB, 379x376)

207 KB PNG

>>102974779
I still don't understand how they managed to go fucking backwards.

Has this ever happened in the history of microchips? An actual decrease from the previous gen?

Its not even a refresh either. Its a totally new architecture, with a totally new socket.

Like I could maybe understand putting out a refresh so that everyone who bought this trash can at least get a working CPU into their system without having to rebuild the entire thing. But this isn't that. This is their next generation product.

I think we are witnessing history right now bros. This has never been done before. Intel is making history right now and I feel like nobody is acknowledging it.

>>102975412
>It's possible the GPU is just faulty but I don't think that's it
the GPU architecture is faulty
it's not just your machine, every 7900 XTX behaves like this
go buy an NVIDIA card (or 6900 XT) and you'll see there's nothing wrong with your setup
>I ran mem test and everything is okay with the sticks so I'm starting to think that AMD GPU drivers just don't play nice with XMP/EXPO
i'm sorry but this isn't how computers work
if you've verified the EXPO profile is stable through extensive memtests (6+ hours, not just a quick 30 minute test) the RAM is not to blame
>I also noticed that if I go into BIOS and disable XMP/EXPO that the crashes become much less frequent
because your CPU is performing worse and decreasing the load on the GPU

>>102975470
285k is impossible to get and I don't feel like waiting anymore

>>102975470
285k has no oc headroom anon
the ring clocks better on the u5 and the u5 has lower factory clocks making it more OCable

>>102975470
looking at the stock for stock
>265k somehow ~=~ 7950x

>>102973014
Yeah I've got tubes of that just lying around the house

>>102975239
New gen coming early next year

>>102975488
so just buy AMD
>>102975491
285k has lots of headroom you mong
>>102975501
somehow the chip on the 265k that is 2y newer and on a newer process beats a CPU in some benchmarks oh no

>>102973014
toothpaste

love the new captchas

Don't know why people care so much about CPUs theyre so boring compared to GPUs

File: 1591892994228.jpg (63 KB, 640x480)

63 KB JPG

>>102971575
How good or bad idea it's to build a brand new AM4 Build today?
> B550
> R7 5700X
> 32GB RAM

File: 41696 (1).png (31 KB, 550x420)

31 KB PNG

>>102975473
It's probably before your time but:
https://www.anandtech.com/show/4955/the-bulldozer-review-amd-fx8150-tested
To give you an idea of the context, the FX-8150 was an 8-core 3.6Ghz Stock/4.2Ghz Boost Part built on AMD's brand new years-in-the-making uarch competing against the out-going Phenom II X6 that was a 6-Core 3.3Ghz/3.7Ghz Boost SKU. This was the result and would be the foundation for every uarch revision for several years until Zen came out.

>>102975545
dumb

>>102975545
bad idea, am5 is better value for perf
>>102975549
bulldozer was a fucking dozer

>>102975475
This all makes perfect sense! Thank you anon!

>>102975142
>>102975412
In wows case it has nothing to do with graphics settings. The game will cause a driver timeout on all chiplet GPUs because some DirectX log file gets full. The fix is to switch to DXVK or stop playing Nu WoW.

>>102975545
If you're building an extremely minmaxed build for gaming at high resolutions and targeting 60fps... $100 ryzen 5600 can't be beat for price/perf.
Everything else is a bit awkward as it gets close to a cheap b650 + ryzen 7600 build, and then you can upgrade to 10800x3d when it's ~$250.

>>102975536
its also ~$100 cheaper

>>102975545
Not a good idea unless it's entirely for gaming and you want a cheap system with a 5700X3D off AliExpress.

>>102975589
sorry but if you're actually chasing performance from ARL-S you'll be buying the OCF and CUDIMMs so 100$ shouldn't really be a concern

>>102975545
If you want cheap why not just buy some used parts

>>102975568
good
now start a class action lawsuit against AMD for selling faulty GPUs yet again
>>102975580
lol retard
7900 XTX is broken on all platforms (because it is literally a fucking broken product, stop looking for software workarounds to a hardware problem)
https://gitlab.freedesktop.org/drm/amd/-/issues/3131

>>102975475
Do you think I should just replace it now or deal with it until the 5000 series?

>>102975613
You are wrong and retarded

>>102975615
>>102975568
See
>>102975580

Driver timeouts are related to DirectX problems aka Windows problems not driver problems. You compain about bad drivers when you should be sending your bug reports to Microsoft to fix their WDDM like how MPOs are broken on Nvidia on Windows.

>>102975371
DIY is pretty much the only segment where AMD has parity with intel, i'm sure the oems appreciate the consistency

>>102975615
you should demand a refund and replace it now
i'm going to quote you:
>I didn't spend $1k on a GPU for fucking low settings in a 20 year old fucking game
>>102975624
i use a 6800 XT on linux
i'm not touching RDNA 3 with a ten foot pole, i want my computer to just work

>>102975545
>>102975556
>>102975559
>>102975585
>>102975591
>>102975604
>>102975604

I'm not sure if it's a good idea to buy and build brand new AM4 system as described + something like a RX 7600 and a 750W PSU

or to buy something like this:
> B650 mobo
> R7 8700G
> 32GB Ram
> 750W psu
> no GPU for this build

My budget gonna be something like ~750USD

I'm on a i7 3770 / RX 6500XT, so anything aM4/AM5 may be an upgrade to this I guess, and the 8700G iGPU has similar performance to my GPU (which is runing on PCIe 3.0 btw)

>>102975678
what country are you in for PCPP

>>102975545
bad idea because ded platform
the only scenario i can think of where it makes sense is some ultra poorfag a320 from the trash with a 5700x3d

>>102975690
Tacoland.
An we don't have the MSRP prices here because taxes and shit.

>>102975697
what the value in taco dollars

>>102975602
CUDIMMS arent needed. I already have DDR5 8200CL38 hynix M die. parts of my raptor lake build that I sold

>>102975706
$15K MX

>>102975709
CUDIMMs outperform UDIMMS if you OC the 285k you mong

>>102975678
what's the closest border state to you? because a microcenter deal + RX6600 is probably your best bet.

>>102975722
thats if you have shit DIMMS YOU MONG

I've been following the 285/265k ocing before they officially released.

>>102975544
Recently, yes. They might become more interesting if Intel's tile strategy plays out. Same reason AMD bought Xilinx

>While Xilinx’s core business is driven by FPGA (Field Programmable Gate Array)-based technologies that are employed as adaptable accelerator solutions in embedded industrial, automotive, data center AI, wireless and wired networking, aerospace, video broadcast/streaming and more. You can think of FPGA technologies as a third type of processing resource – CPUs, GPUs and FPGAs – though FPGAs have a superpower of sorts due to their adaptability. With the ability to be reprogrammed on the fly, in-band and without dropping a packet, FPGA-based accelerator solutions are great for applications that have changing and evolving workloads, like AI inferencing, SmartNICs, 5G cellular base stations, even algorithmic trading and analytics, in addition to many others.

>>102975652
>demand a refund and replace it now
waste of time
gpu works perfectly but its bottlenecked by directx

>>102975730
Texas and New Mexico

>>102975645
Then why did this anon >>102975613
link an entire toubleshooting thread where they discovered that it's not a Windows problem since it also occurs the same on Linux? It looks to me like RDNA3 is just defective and I got scammed. I am willing to accept getting duped but I am not willing to delude myself. Also, I tried Linux with the 7900 XTX as well and was still experiencing crashes related to the GPU.

>>102975645
It's not the MPO issue, that is solved by a registry edit.

>>102975731
clearly not because you should know the 285k can scale beyond 9400 on CUDIMM

>>102975536
spouting untruths doesnt change objective reality you neanderthal

>>102975549
Phenom was really good though.

the 13 and 14 gens were literal dumpster fires.

File: 1.png (213 KB, 872x721)

213 KB PNG

>>102975747
he posted some random gitlab issue that was resolved with a new kernel update.

>>102975756
MPO is broken on Windows you stupid fuck
Nvidia tells you to disable it because Microsoft and Nvidia don't want to fix it
https://nvidia.custhelp.com/app/answers/detail/a_id/5157/~/after-updating-to-nvidia-game-ready-driver-461.09-or-newer%2C-some-desktop-apps

>>102975734
Then why do I have the exact same crashes on Linux? Why do Nvidia GPUs not experience this issue nearly as commonly as AMD GPUs?

>>102975774
neanderthals were actually more intelligent than homo sapiens

>>102975774
as much as I have been shitting on intel
>he doesn't know about how good memory OC is on ARL

>>102975779
And I have turned off MPO and I still get the same crashes.

>>102975782
User Error? Why don't Wendel from Level1techs crash with his 7900XTX? Why dont Gamers Nexus crash with his 7900XTX? Why don't Linus crash with this 7900XTX in his system?
You're a stupid cunt.

>>102975734
>it works perfectly
>trust the science
>>102975747
>It looks to me like RDNA3 is just defective and I got scammed
if you went through the 5700 XT saga 5 years ago you would know this is sadly nothing new for AMD
honestly i haven't seen much evidence for the 7800 XT and below being defective so i don't know how many people are actually affected but yeah, i believe you got scammed

>>102975792
>and I have turned off MPO
why?
are you on nvidia?
do you have hw acceleration disabled too?
or are you too stupid to use a computer god damn also nvidia has this same issue

>>102975800
You're a stupid fuck.

>>102975737
yeah going to Houston for a might be the best bet

https://www.microcenter.com/product/5006808/amd-ryzen-5-7600x,-asus-b650m-a-prime-ax-ii,-gskill-flare-24gb-ddr5-5200,-computer-build-bundle
then pay $10 to get the 48gb kit instead

and then
https://www.microcenter.com/product/680053/xfx-amd-radeon-rx-6650-xt-speedster-swft210-dual-fan-8gb-gddr6-pcie-40-graphics-card

thats only $560 and you can get a PSU and maybe a new SSD and still have enough to pay for gas and lunch

>>102975792
>>102975782
you have outdated kernel and wrong WDDM

>>102975779
>he posted some random gitlab issue that was resolved with a new kernel update
it's still open
now answer me why helldivers 2 was crashing on 7900 XTXs on both windows and linux, despite having completely different driver stacks
>>102975795
no techtuber uses that card in their daily systems
they test it for 10 minutes at launch to get some FPS numbers
>>102975803
>You're a stupid fuck
post your card
>OpenGL renderer string: AMD Radeon RX 6800 XT (radeonsi, navi21, ACO, DRM 3.59, 6.11.5-lqx1-1-lqx)
>OpenGL core profile version string: 4.6 (Core Profile) Mesa 24.3.0-devel (git-65f7fb29f5)

>>102975803
Because disabling MPO solved a lot of the crashes. It applies to AMD cards too. It didn't resolve the other crashes though. You seem really angry for no fucking reason though.

File: 1617216288206.jpg (216 KB, 1146x1154)

216 KB JPG

>hear about all the good gaming performance improvements in 24h2
>learn it installs Recall into your machine
>yeah you can "turn it off", but I want to delete it
>deleting the files borks File Explorer
>only options are to "turn it off" with it possibly still able to run again, or use a third-part explorer

fuck cunt shit ass dick faggot bitch-

>>102975851
>using W11

File: 1685700409735822.png (146 KB, 384x390)

146 KB PNG

What's the go-to case these days if I want something that has good build quality, is easy to work with, and has at least passable airflow?

File: file.png (1.97 MB, 3440x1440)

1.97 MB PNG

>>102972424
Done, shit feels smooth as fuck even though it is running on a 7 year old 35w CPU.

>>102976029
Montech 903 (either the Base or Max version) or for a bit more either the Lian Li LANCOOL 207 & 216, Fractal Pop Air (and it's many variants), and the Antec FLUX.

File: 1728844128172859.jpg (153 KB, 720x634)

153 KB JPG

is there a recommendation for UPS? My fucking power keeps dropping for a second or so randomly which is super annoying

https://www.youtube.com/watch?v=6QoCCFXD0xc
>MSI shipped reviewers boards with default BIOS that only recopgnized Engineering Sample CPUs
I mean it's good that they weren't going to retailers and it was fixed within days of launch, but come on Intel give your AIBs some slack as this launch reeks of crunch time BS.

>>102976164
>is there a recommendation for UPS? My fucking power keeps dropping for a second or so randomly which is super annoying
I'll just tell you what I've been using:
https://pcpartpicker.com/product/qvVBD3/cyberpower-cp1500pfclcd-ups-cp1500pfclcd
>Sine Wave Output
>Plenty of VA/Wattage capacity
>Cheaper than other options

I'm completely rebuilding a secondary PC from 2012~2013 that I'm going to use for gaming and media. Is Intel currently 'avoid no matter what' tier garbage or is it still acceptable depending on how good of a deal I can get?
Current AM5 prices aren't so bad where I live and I can currently get a good deal on a full size B650 board (Asrock Pro RS) and 7600 + 6000MT CL30 RAM. Just wondering if the Intel offerings are worth considering because of those degradation issues.

>>102976249
>Is Intel currently 'avoid no matter what' tier garbage
yes

>>102976249
>Is Intel currently 'avoid no matter what' tier garbage or is it still acceptable depending on how good of a deal I can get?
>Current AM5 prices aren't so bad where I live and I can currently get a good deal on a full size B650 board (Asrock Pro RS) and 7600 + 6000MT CL30 RAM. Just wondering if the Intel offerings are worth considering because of those degradation issues.
If you get a good deal it's fine since Intel root caused the problem and issued their final microcode patch that's been folded into BIOS updates since mid-September:
https://community.intel.com/t5/Blogs/Tech-Innovation/Client/Intel-Core-13th-and-14th-Gen-Desktop-Instability-Root-Cause/post/1633239
Cross shop some 13/14th Gen i5/i7 against the 7600 and see which is a better buy locally for you. Also since this is a gaming build consider an AM4 5700X3D off AliExpress (Global Shipping) for REALLY cheap w/DDR4 RAM and but the savings back into your GPU or pocket them.

>>102976268
It's fine if he's buying a fresh non-degraded chip and slaps it into a new build with Day 1 BIOS Update,

>>102976249
Intel is Bulldozer tier now.

>>102976029
Fractal Pop Air

>>102976268
That makes the choice easier at least. Thanks.
Can I still go either Nvidia or AMD with GPUs depending on whichever is cheaper? I plan to keep the 1060 in this thing for the time being and save up for at least a 70 series Nvidia or 800 series AMD card.

>>102976283
>Cross shop some 13/14th Gen i5/i7 against the 7600 and see which is a better buy locally for you.
They're about the same. Only the 12th gen Intel stuff are significantly cheaper than 13th and 14th gen.
>Also since this is a gaming build consider an AM4 5700X3D off AliExpress (Global Shipping) for REALLY cheap w/DDR4 RAM and but the savings back into your GPU or pocket them.
I've considered this path too. But there's still some risk of ordering stuff off of Aliexpress and the possibility that I'd need to straighten bent CPU pins in case of any shipping oopsies. Boxed 5700X3D do go on sale here but nowhere as cheap as AE prices so it's not worth it.

>>102976356
in general, if you want features like DLSS or plan on doing AI/something that needs CUDA go nvidia, if you just want the best raw price/perf go AMD, both of the next-gen cards should be announced in january

File: minimum-fps-1920-1080.png (121 KB, 560x2050)

121 KB PNG

>>102976356
>They're about the same. Only the 12th gen Intel stuff are significantly cheaper than 13th and 14th gen.
12th Gen is very viable for a gaming build and still has alright MT perf to boot. The i7 is basically a 7600 in games for example (see chart). Moreover, unlike the Ryzen 7000 series you can still leverage cheaper DDR4 RAM with it also.

>I've considered this path too. But there's still some risk of ordering stuff off of Aliexpress and the possibility that I'd need to straighten bent CPU pins in case of any shipping oopsies. Boxed 5700X3D do go on sale here but nowhere as cheap as AE prices so it's not worth it.
Honestly, no more risk than ordering on Amazon these days. The chip is both really small and very lightweight so risk of shipping damage is low.

File: 121231312312.png (128 KB, 1270x424)

128 KB PNG

>>102976231
This is what I have.

We get brown outs here frequently in the summer so this thing is well worth it to me.

The key is to check what the maximum wattage output of the battery is. The amp hours is fucking useless on these cheaper units. You just need enough time to shut down your computer off as soon as the power goes out. If you actually want to run, you need a whole new tier of battery.

And you have to measure the maximum wattage your WHOLE SYSTEM pulls. That includes your MONITOR because you will need that to turn your computer off in the event the power doesn't come right back on.

If you are pulling 500 watts, and your UPC is only rated for 300 watts, and the power goes out, your system will turn off. Bad things will happen.

>>102974078
I recognize that Ayano

>>102975405
It's on Linux too, the scheduling is just better and mitigating some of the issues but Intel and AMD are concentrating on making server the best because the margins there are the juiciest and it doesn't matter to them that highly that gaming either improves very slightly or regresses in Intel's case when it is one of the many workloads they are putting in lower priority. That is why the 285K is at the top for things like Selenium and Pybench which are all single threaded but since they don't have that much latency sensitivity, they don't suffer like gaming does.

File: 8tbssd.png (1.25 MB, 1153x1158)

1.25 MB PNG

Welp, this is it

File: selenium-jetstream-2-goog(...).png (234 KB, 900x1467)

234 KB PNG

>>102976517
>That is why the 285K is at the top for things like Selenium and Pybench which are all single threaded but since they don't have that much latency sensitivity
not sure where you got this from
browsers are very latency sensitive, javascript is not a friendly language to branch predictors or prefetchers

>>102975545
Hello Peter!

File: file.png (294 KB, 1260x2052)

294 KB PNG

>>102976562
Should've specified, it was in WASM.

>>102976605
those benchmarks are as synthetic as it gets
but yeah, if you have a small workload that fits entirely in cache you can run it really fast over and over again (this is what happens with those factorio benchmarks on X3D chips, they lose on bigger maps)

>>102976550
ill buy this in ~12 months. that should be ample time to let them fix any issues that may arise.

>>102976367
I might wait it out then to see if current gen prices drop.

>>102976407
>The i7 is basically a 7600 in games for example (see chart).
Does this still apply to the non-K 12th gen i7? The K versions are still expensive here.

>Honestly, no more risk than ordering on Amazon these days. The chip is both really small and very lightweight so risk of shipping damage is low.
They don't have a good return and refund policy, so unlike local online shopping platforms here that offer hassle-free returns within 2 weeks in case of any defects on arrival, it is more risky.
The impression I get when buying stuff on Aliexpress is that one should be prepared to burn the money in case it's DOA, so I make a note never to spend more than $30~50 there.

>>102976407
>as good as Bezos
good luck with you aliexpress RMA

>>102976550
does it lose to samsung again?

File: 1080p_Average.png (70 KB, 1395x1565)

70 KB PNG

>>102976680
>Does this still apply to the non-K 12th gen i7? The K versions are still expensive here.
https://www.techspot.com/review/2391-intel-core-i7-12700/
Pretty much little difference.

>>102976721
they dont have an 8tb nvme consumer drive so it wins be default

>>102976550
hmm I'll take the teamgroup drive for $10 less

>>102976721
samsung wins in speedrunning drive failure desu
>>102976762
that's up to you. this wd drive uses newer bics6 nand and has a controller with dram cache while teamgroup has chinknesium nand and doesn't have dram cache

>>102976780
Sounds cool, but what do the new bits actually do anon

>>102976805
better random i/o & sequential performance making it perform at parity with older 1-4tb drives instead of performance regression like other 8tb drives and much better power efficiency

>>102976830
so probably not worth $10, that said if you are spending ~$800 it probably doesn't matter to you

Anyone know a good UPS? I am in Europe and I need at least 900W, since I have a 850W PSU.

>>102976879
apc

>>102976879
not apc

>>102976844
its $10 more for the highest performing 8tb drive so I don't see why you wouldn't get one if you were shopping for it
this drive has obsoleted all phison e18 8tb drives that was the previous performance standard for this capacity but also costs >$900 and were slower than their 4tb counterparts

I havent built a new pc for 10 years at least. What AM5 CPU air cooler do you recommend for a low profile ITX build? I was looking at a Noctua NH-L12S but I dont really know who else is making good cooling these days.

Will probably be going with 7800x3d cpu.

>>102976988
that'll be fine

>>102976988
9800x3d is launching on 7 november so i'd suggest going with that if you're able to get one on day 1

>>102973425
Obviously you can't infer too much from a freeze frame but it does seem like a CPU+RAM issue (assuming you have 16GB and are using 15.5 here). It's not thrashing so shitty game optimization can't be ruled out.

>>102976943
>>102976953
Any reasons as to why? I'm very inexperienced with UPS.

New builzoid Vid for ddr5 AM5
https://www.youtube.com/watch?v=Xcn_nvWGj7U

File: 1707538611249935.jpg (27 KB, 222x361)

27 KB JPG

Isn't this one of the best CPUs for gaming rn? How is it so cheap?

>>102977330
aging platform that noone except poorfags care about anymore

File: {5AC09DAA-DC8D-4970-B21D-(...).png (200 KB, 781x337)

200 KB PNG

>>102977330
>210
>cheap

>>102977330
unless you already got ddr4ram/AM4motherboard it not worth buying

>>102977365
remove another ~$40 or so from the 11.11 sale soon

>>102977252
nta but it strikes me as a barbell market. On the left side you've got big standard units which are all built to cost and basically interchangeable. Lead-acid, line-interactive, simulated sine wave, etc. - nothing special but affordable at least. Then on the other side there's the ones you'd actually want with all the bells and whistles like double conversion but they're like $10k starting. So IMO the takeaway is that, if you're not willing to spend the big bucks, just shop by price and check for recalls/defects. It's like shopping for a washing machine - quite hard to buy a /bad/ one these days but it's never going to last 30 years like your mother's.

>>102977370
There's already one for 155 that's somehow not at the top when sorted by price, have to set upper price limit to find it faster.

>OCing 4070S memory
>get performance regression compared to stock even at +100mhz
wat
and no i'm not power limited

are the raptor lake chips fucked should I get amd instead

>>102977365
I mean, the 7800X3D in here is almost 500 and the 5800X3D is 341...
What can i say, my country is a shithole.

>>102977396
They're great value and are nominally "fixed" with the new microcode but degradation takes time to set in so it'll be a few months before we know if the patch has taken. Unfortunately, because of this, few are buying them so we don't have nearly as much data. Low signal to noise.

>>102977396
Not if you buy new and patch the motherboard bios immediately after assembly.

>>102977402
Nah it's just aliexpress.
If you don't trust them or cant wait, does computeruniverse ship to your area? It's 180 there but only for OEM tray version.

>>102976732
Well, after looking around and adding up everything the difference between a 12th gen Intel system with the i7 12700/F and an AM5 one with the 7600/7500F is about $60 here. That difference of course comes from the motherboard and RAM, but the AM5 one is a full sized B650.

>>102977451
>>102977458
Thank you

>>102971575
>i7-7700
>16GB ram
>7900 GRE
I've had a problem with my PC freezing whenever I try playing something that would even be mildly "resource intensive", low spec games run fine (Corekeeper, Legends of Runeterra etc.) but even running Tekken 8 now will cause the system to freeze up when i could play it 4k60 just fine a month ago. I already tried formatting my PC but the problem persists.

I've noticed the CPU getting wild 70-90% usage spikes even when I'm just watching videos so I repasted but the problem persists.

What the hell is going on

>>102977483
The mobo being full sized doesn't really matter unless it has extra features you think you'll use. Basically comes down to how much the $60 savings is worth to you. AM5 will have an upgrade path to whatever cheaper X3D chip is available years later like the 5700X3D is on AM4 right now.

>>102977503
>radeon
found your problem

>>102977525
>The mobo being full sized doesn't really matter unless it has extra features you think you'll use.
It's the extra USB ports that I'd like to have, and two easily accessible extra m.2 slots for storage expansion.

>AM5 will have an upgrade path to whatever cheaper X3D chip is available years later like the 5700X3D is on AM4 right now.
So something like a 7500//7600/7700X3D? I think I'll go for that then. Thanks.

>>102977544
So it's not the CPU that randomly spikes and it's the GPU that's been working fine till now?

>>102977375
sage advice

>the in-between is to obtain enterprise grade stuff that needs batteries replaced
>... or is it just me that does that?

>>102977503
have you updated any drivers or anything else recently?

>rollback

>>102977563
>It's the extra USB ports that I'd like to have, and two easily accessible extra m.2 slots for storage expansion.
Sound good then.

>So something like a 7500//7600/7700X3D? I think I'll go for that then. Thanks.
Yes, and no problem.

Are seasonic's branded psus better than corsair? a prime 750 goes for 300 while corsairs 100-200, but both are made by seasonic. Will seasonic fry my pc if i refuse to pay the 200 shekel tax?

>>102977931
No, whether it's under their own brand name or as a white label for other companies Seasonic never makes bad PSUs. There are very few truly bad PSUs on the market these days.

>>102977119
>9800x3d is launching on 7 november so i'd suggest going with that if you're able to get one on day 1

thanks. That sort of complicates things. I just found a local store has R9 7900x3d on sale for less than the 7800x3d. Now i'm not sure if i should wait for the 9800x3d or just get the 7900x3d instead. what do?