Thread 106280478

Does /g/ use i9-14900K?
Has this piece of shit been fixed since the latest BIOS update?
my only chance to buy a RTX 4090 is if I buy the shitty i9-14900K CPU
>inb4 buy 5090 anon
No, my old software isn't compatible with the 5k series.
>get AMD CPU
I would have but none are available with RTX 4090.

So, have they been fixed?

I wouldn't trust prebuilt PCs with a 14900K, you don't know when they've flashed the safer CPU microcode.
The damage may have already been done during the stability testing phase.

>>106280577
>I wouldn't trust prebuilt PCs with a 14900K, you don't know when they've flashed the safer CPU microcode.
>The damage may have already been done during the stability testing phase.

Not him but more details? All I know of the situation is the memes that get posted here.

>>106280577
Dang man
it's my only option to get a RTX 4090
what about Ultra 9 285K?

>>106280709
https://community.intel.com/t5/Mobile-and-Desktop-Processors/Intel-Core-13th-and-14th-Gen-Desktop-Instability-Root-Cause/m-p/1633442
https://community.intel.com/t5/Mobile-and-Desktop-Processors/Intel-Core-13th-and-14th-Gen-Vmin-Shift-Instabilty-Update-New/m-p/1686948

>>106280717
Which pre-built were you planning to buy? If it's using DIY motherboards, you can manually set a lower VR voltage limit in the BIOS (1.4~1.45V).
285K is safe, 265K is the better value CPU with minimal gaming performance difference.

13th and 14th gen are cursed.
Go 8th-12th or 15th.

>>106280748
I have a 13600KF but now need video encoding. 12600K looks like a good upgrade.

>>106280478 (OP)
I've had mine since march last year

>>106280478 (OP)
I have bought an i9-14900KF and an i9-14900K. I got the KF as soon as they came out and regretted it. Went through 2 RMAs before they released a real fix for the voltage issue. I bought the KF and ran it undervolted for several months before patching it. Since their second "fix" for the voltage issue, both CPUs have been fine. Overall unimpressed with their performance though, my 9th gen i7 felt at least as fast when using my computer.

>>106280743
>Which pre-built were you planning to buy?
HP OMEN 45L Gaming Desktop GT22-2000in PC
it apparently uses ArcticOC motherboard
>>106280786
what do you use it for? how often does it error if at all?

>>106280776
14600K/Fs are fine, it's the 13/14th gen i9s and i7s (much lesser extent) that are at real risk.
The voltage degradation issue arises while those i9/7s are trying to boost close to 6GHz when the thermal conditions shouldn't allow that to happen.
A 14600K is capped at 5.3GHz and will not request over 1.4V to hit that maximum boost clock.
The microcode is fixing CPUs that can ask over 1.55V, not happening on i5s.

>>106280478 (OP)
Cuda came out in 2007, you need a really fucked up old piece of tech to need 32bit.
Physx is stupid, especially for video games, the value of a 5090 over a 4090 does not justify buying a 4090 with an intel (implying the 5090 is at the same price of a 4090, the 5090 is better at AI).
OpenCL is easily emulated on the CPU (for such an old application, a CPU is fine), and you could enable IGPU multi-monitor support in your Bios so that your IGPU can be used for openCL (I think).
Intel CPU's are better than ryzen for a couple things, but if you are a gamer, the 9800x3d is the main choice for games with deveno + custom DRM + anti-cheat, such as mohun wilds.

>>106280478 (OP)
overblown issue, been using mine since last year no issues. Nobody actually read the Puget system stats that started this drama or they would have seen the 7950x3d had an even higher failure rate than raptor lake lmao.

>>106280820
I bought the K and ran it for several months*

>>106280478 (OP)
I use two on in my main machine and another in my server. Both work fine but my main machine won't sleep without hard crashing I think from something voltage related. You need a good motherboard for stability.

>>106280825
Thanks brah, you talked me off of the ledge.

>>106280821
That's a proprietary motherboard, I'm not sure if it exposes the option to set a VR voltage limit.
It's supposed to be a hidden option in all Intel boards, but DIY models like ASUS/MSI/Gigabyte/ASRock made it usable after the degradation issues were publicized.

>>106280821
>what do you use it for?
General PC use and gaming
It has never had any problems

>>106280841
This is a factory-fused 13600K V/F curve mapped out by an XOCer, your chip won't have the exact characteristics.
But I doubt the difference will be over 200mV.
i5's are safe @ stock settings, but Intel's recommending K processor owners to use the latest microcode to cover their ass.
https://skatterbencher.com/2022/11/29/skatterbencher-51-intel-core-i5-13600k-overclocked-to-6200mhz/

>>106280827
also you could plug your old GPU into a spare pcie slot to allow it to use old 32bit Cuda or Physx if you needed it.

>>106280828
PugetSystem's data doesn't represent the norm because they set a stricter 125W PL1 limit when Intel's spec for these 12/13/14th gen -K processors are PL1 = PL2 = 253W.
So those 13/14th gen processors on a 125W PL1 carebear mode are still degrading, not a good look.
https://www.pugetsystems.com/labs/articles/power-draw-and-cooling-14th-gen-intel-core-processors/

14900k user for over a year, did the ME update about 6 months or so after running it without issue. Never had any anomalies with it. 128/4070tisuper/asusmobo

>>106280827
>Intel CPU's are better than Ryzen for a couple things
Do you recall?

>>106280478 (OP)
Read Puget
https://www.tomshardware.com/pc-components/cpus/puget-says-its-intel-chips-failures-are-lower-than-ryzen-failures-retailer-releases-failure-rate-data-cites-conservative-power-settings

>>106280748
Had 13700k for years before we knew about the microcode issues, smooth as butter. Running 14700k now, updated microcode, no issues. I always underclock for better temps, so there's that.

>>106281536
uh depends on the CPU you are comparing against.
the 14900k is maybe better than the 9800x3d for productivity because it has more cores (e cores), and maybe 1 or 2 games might have a higher 0.1% low (but the 9800x3d is much better at AAA games due to the 3d cache in like 3 games that matter, such as monhun wilds / warhammer space marines).
But the 9950x exists with 2x more cores than the 9800x3d and it's power efficient in perf/watt or something if you enable an option in the bios (but it lacks the x3d cache, the 9950x3d exists, but the 9800x3d is supposed to be a lot cheaper with the same gaming performance).

>>106283288
Any point in undervolting a 12700 (non-K)?
If so, any decent guides on how to do it?
I have a gigabyte z690 board.

>>106280478 (OP)
It can't be fixed, you have to RMA that shit. It's not a software issue.

>>106280478 (OP)
They have allegedly been fixed, yes, but keep in mind that this is "allegedly" and we don't know if some sort of slower degradation is still happening. Maybe after some more time passes we see them start dropping like flies again, basically I would not trust them.

Beyond that by the way you're phrasing things I take it you're buying some prebuilt? You need to be extra careful if so. If you're buying a used system from somebody else, then the CPU is used and potentially degraded to some level, you don't know if it's been updated and if it has you still don't know how long it was used before the fix was applied. If you're buying a new prebuilt, then the problem there is that some of them may come with custom or OEM mobos that aren't normally sold directly to the consumer. You don't know whether the BIOS on the board is updated sufficiently to contain the fixes and if it's not a standard retail mobo meant for direct sale to customers then you also don't know whether actually finding a BIOS for it will be easy or no. Standard retail mobos just have pages on the manufacturer's site where they provide updates and such, but if it's some OEM thing that may not be the case and you may have to rely on the OEM to supply them.

If you're buying one of those CPUs nowadays, you really need to do your homework to make sure you know exactly what you're getting and that it's either already updated or that you can easily apply the update yourself. I think it's a bad idea to buy 13th or 14th gen, you should just build your own PC with a 4090 and a CPU that doesn't suffer from this shit.

>>106285064
I think undervolting is solving the problem the wrong way. The chip pulls too much power targeting high clocks, or rather, it predicts the voltage ramp incorrectly for those high clocks, and so undervolting really only power starves it. You shave something off the highest clocks and you get the same result with less time with the CPU spent in an "error, no power" state.

>>106280828
Yeah bro it's totally overblown, that's why Intel shit its pants when it got out and then spilled their spaghetti while spitting out a million "fix" updates one after the other which all failed the fix the issue until much later.

>>106280478 (OP)
I benchmarked a 14900K and 12400T, literally 1.4fps difference in Cyberpunk 2077 maxxed out.

>>106280828
>>106282041
based

>>106282041
>falling for clickbait
litterally post the puget systems article you insufferable faggot
https://www.pugetsystems.com/blog/2024/08/02/puget-systems-perspective-on-intel-cpu-instability-issues/

>>106289016
cringe

>>106280577

Even if that’s the case I believe you can get Intel to replace it. This should be a recall.

>>106280478 (OP)

I got a replacement processor with the cash refund from Intel in Nov 2024 and has not experienced a single blue screen. The PC temps are also lower with the same hardware and software install of 11. I am insured to late 2029 per Intel. I think they more or less fixed it but time will tell.

>>106289020
cope

>>106280478 (OP)
I have learned quickly that any Intel CPU must be an F variant.

>>106289020
>>106289495
samefag

Use case for Intel over AMD in current year?

>>106280821
You can use intel extreme tuner to undervolt the cpu and the create a .ps script to automate it every time windows boots up.
If you are still concerned you can shave off from the max. frequency in the power settings. I have an older hp desktop and while normal boost clock is over 4k mhz I limit it to 3,500. Nice and quiet.
Won't help nu games anyway because they are full of unoptimized code...

>>106280478 (OP)
Intel is a sinking ship

>>106280827
>Intel CPU's are better than ryzen for a couple things
Like what exactly? A second faster rendering in final cut? Some obscure instruction set no one actually uses? Synthetic benchmarks only faggots care about?
I always hear this but its always some dumb shit no one really cares about then years latter its revealed that Intel fudged the numbers in the reviews. When are you fuck faces going to get it, Intel is a jew company and everyone knows how honest jews are about the products they sell.
>pic related a jew and the product they shill

>>106280478 (OP)
I had the 13900ks which was basically a better binned 14900k
it was not a nice cpu to use because it unironically draws like 2.5x to 3x the power draw of a comparable x3d ryzen in games

>>106280861
I can tell by slight irregularity of some of the pixels in this image your CPU has degraded a fair margin already