Steve reached his dream and he makes the industry move now. He took the passing of his friend and inspiration at heart and is doing him proud.

>>106749502
buy an ad

File: AMD shills.png (96 KB, 1517x641)

96 KB PNG

>be AMD
>be cheapass chinks
>push the physical and voltage limits of CPU's beyond normal range
>the goyim won't know.dat
>it dies
>surprise pikachu face
>pays GAYMDER NEXUS to accuse someone else so we don't have to pay for RMA and brand reputation
>rinse and repeat with each generation of CPU's/GPU's

happened with the 7800X3D and it wasn't AMD's fault, it was MOBO manufacturers.
DDR5 RAM unstable (unlike Intel), must be the RAM sticks manufacturers !
new CPU's kill themselves, must be Asrock's fault even though we can't replicate the problem on their board !

remember AMD dindu nuthin'

>>106749502
>>106749666
Buy an ad.

>>106749666
Everything is shit now, the collapse is coming.

>>106749502

cleanin camera sensor with blower bulb and wetswabs dust dust not everywhere but is

>>106749834
>the collapse is coming
Care to elaborate?

>>106749666
DDR5 is unstable garbage on intel as well
https://www.youtube.com/watch?v=5R8en_FtA80

>>106749986
Sure but at least you can actually get the memory bandwidth you paid for.

>>106750002
fair enough
but CPUs don't need more memory bandwidth

>>106749666
This but unironically.
Intel needs to up their budget on YT to counter the nonstop AMD circlejerk.

>>106749986
i'm stable at 6400 CL32 2x16GB on a 13600K.

>>106750011
With iGPUs composing 4K desktops and flipping game buffers from dGPUs, DRAM-less NVMe SSDs stealing memory capacity and bandwidth for their HMBs, and just general wastage from trash programming languages constantly stressing memory allocation, maybe they do.

>>106750040
That video is for >7000, 6400 CL32 is the sweet spot for 13/14th gen.

Sure smells like LoserBenchmark ITT.

I have a 7950X3D and a GTX 4090 in my bedroom which I leave on overnight.
How have I not been burned to death yet?

>>106750060
>which I leave on overnight.
For what purpose? Why not put it into sleep mode at least?

>>106750049
well duh, more than this is overclocker territory, it's for hobbyists and has no real practical uses for professionals or gaymen.

>>106750040
happy for you but that would still be hard on many motherboards
>>106750041
>With iGPUs composing 4K desktops and flipping game buffers from dGPUs
yeah that's true, it's relevant if you use the iGPU
>and just general wastage from trash programming languages constantly stressing memory allocation
you need low memory latency for fast malloc

>>106750081
>happy for you but that would still be hard on many motherboards
it's not a particularly good MOBO, just an MSI Z790-A WIFI board, something you could buy for 200 euros.
isn't AMD unstable because they put the RAM in Gear 1 instead of Gear 2 anyway ?
chasing a mirage by doing that instead of opting for stability.

>>106750112
>it's not a particularly good MOBO, just an MSI Z790-A WIFI board
keep in mind stability means you can pass any stress test for an indefinite amount of time
not "it boots into windows and plays my favorite vidya"
>isn't AMD unstable because they put the RAM in Gear 1 instead of Gear 2 anyway ?
it defaults to gear 2 if you try to run 6400
at 6000 it runs gear 1
so 6000 > 6400 on AMD if you leave everything on auto

>>106750131
stability means it never crashes, posts instantly, no crash during OCCT tests.
i remember the 7800X3D frequently no posting at 6000 CL30 during boot, frustrated tons of people back in the days.

Redditard problems, redditnexus weren't able to replicate the supposed problem, also intel cpus are junk keep coping intelaviv lul

>posted from my 9700x

>>106749666
the problem has nothing to do with AMD and everything to do with the OEMs using their stupid linear scaling voltage based on the set frequency of the RAM. Intel had this problem too even back with SB/SB-E chips where the board was setting the VCCSA at 1.325v when you set the RAM to 2133MHz resulting in massively degraded silicon and eventually causing crashing at 100% stock settings. Early AM5 had this problem too with the SOC going to 1.3+ when using XMP and AMD had to set the max to 1.25v or some shit but guess what? with LLC and everything even applying 1.28+ at times because of LLC and other board settings. I have to set my SOC at 1.14v just to keep the below 1.2v.

The takeaway from all this should be, never trust auto voltages.

>>106750199
>posted from my 9700x
it's gonna explode bro
>>106750213
>with LLC and everything even applying 1.28+ at times because of LLC and other board settings. I have to set my SOC at 1.14v just to keep the below 1.2v
what board? that's interesting, i never saw anything like that on AM4
if you have SOC LLC control in the BIOS i would try to set it to the lowest/droopiest setting possible

>>106749666
There is truth to this. A lot of the bigger companies put a lot of pressure on their board partners. Sadly the consumer space is utterly fucked with absolute dogshit. Post Alderlake debacle has me not trusting Shitel either, which is a shame because for awhile, they seem to really put out shit that mostly worked.

Yes I'm aware of the constant back and forth on TSX and other minor microcode bugs. Meltdown was extremely pathetic and a massive pockmark on Intel, but I think they mitigated it well, even then.

File: file.png (15 KB, 543x170)

15 KB PNG

>>106750213
just to add some more to this, if you want to see it in action just load the bios defaults and change nothing except the RAM frequency, boot up and look at HWINFO or something and note the SOC voltage being increased all by itself.They do this to increase the ram "stability" but its so wildly overvolted that it results in killing CPUs.

>>106750243
Gigabyte X870E AORUS ELITE w/ a 9950X3D but even my old AM4 boards did it too including a X470 Crosshair VII and X570 Crosshair VIII Dark Hero. I do currently have both LLC set to the lowest level but it made no difference, picrel as I mentioned its currently keyed in at 1.14v, I should probably drop it to 1.1v

>>106750067
So I can shit post on 4chan befire I go to sleep and again first thing in the morning without having to get out of my bed to press the power switch.

>>106750243
I don't buy asrock junk and even then I doubt it's a real problem outside of reddit, this looks like intelaviv damage control, reddit is just mutt corpo marketing central

>>106750317
Bruh, coming out of sleep doesn't take long and it can be done with a key press or mouse click. No point in wasting power overnight.

>>106750295
check the SVI3 section instead
those sensors are probably measuring voltage at the VRM, before vdroop
>I do currently have both LLC set to the lowest level but it made no difference
don't mess with CPU vcore LLC on AMD
it will lower your performance from clock stretching
>>106750319
3... 2... 1...

>>106750319
Sorry AMDrone, mommy Su's shit also stinks.

>>106750329
I run my system os entirely in ram and I put it in sleep mode every night, my uptime is like 2 weeks now, it feels like black magic or something

>>106750345
Ok shlomo but your intelaviv cpus are dying left and right, cope more

>>106750343
>check the SVI3 section instead
both are basically the same, if I just left it at 1.2v it would be around 1.28v which is nuts for just cl28 6000mhz

>>106750369
>your intelaviv cpus are dying left and right
Just like the "vaxies" right? Two more weeks, etc.

>>106750389
Lol intelaviv shlomo got zyklon'd, classic

>cpus die, intel: INTEL'S FAULT INTEL'S FAULT LALALALALA IGNORING ALL EVIDENCE POINTING TO IT BEING MOBO AND BIOS/UEFI SETTINGS
>cpus die, amd: OBVIOUSLY AMD IS INNOCENT AND IT'S THE MOBO AND BIOS/UEFI SETTINGS ALSO INTEL BAD INTEL BAD INTEL BAD STOP LOOING AT AMD

>>106750375
that's crazy but it wouldn't surprise me if gigabyte is doing a silent overvolt so they can claim to stay within AMD spec
from what i remember if it set 1.1V SOC on AM4 it would report 1.1V at idle and 1.09V under load
it never spiked above what you set in the BIOS

>>106749666
>push the physical and voltage limits of cpus
...that's what intelaviv has been doing for the past decade retard

>>106750406
Pretty much.

>>106750431
And AMD has been doing it since Ryzen's inception. Zen memory controllers have been degrading since 2017, to such an extent that it can only be by design. Intel got caught and justifiably wrecked, but for some reason AMD keeps being given more and more chances and you fanboys keep falling for it.

>>106750409
It's not just Gigabyte its literally all of them to some extent, Asus is by far the worse. I've been dealing with SOC / VCCA voltages being insane since my X79 3930k which degraded and lost the ability to be at 4.8GHz due to that shit man. My friend had one too that literally would crash at stock settings so seeing all the 14900k issues made me laugh because its dejavu to what happened to my chip all those years back.

>>106750476
>zen memory controllers have been degrading since 2017
...source?

>>106749666
What am I supposed to do? Buy an Intel CPU? This isn't Skylake anymore. Intel CPUs arent assuredly bulletproof like they once were surely?

>>106749666
>only happens with one board manufactures
>AMDS FAULT THIS IS AMDS FAULT ONLY AMD
Retard

>>106750494
Is this nigger serious?

>>106750483
>It's not just Gigabyte its literally all of them to some extent
i wasn't denying that
>seeing all the 14900k issues made me laugh because its dejavu to what happened to my chip all those years back
14th gen was 100% intel's fault
they've been using the same boost system forever and it wasn't properly equipped to handle a 24 core CPU safely

>>106750431
obviously, and this is why they stopped doing this with the 15th gen because the 13th and 14th gen almost killed their company.
meanwhile AMD still doesn't understand that the old monolitic CPU architecture has no future because electrons will destroy everything at such high voltage.
and i still don't get why they want to push such powerful CPU's with 8 gazillions core and a toaster just to be able to run games (that barely use half of the CPU capacity in most cases).
the average consumer doesn't need monsters like that, it's for the pros.

>>106750505
the 13th/14th gen problem only impacted the i7's and i9's, if you're gayming then an i5 is sufficient.
it's cheaper and works just the same (without the crazy voltage and RAM problems).
i don't care what you buy but if you're not doing pro work related stuff on your PC, you most likely don't need all that CPU power.

>>106750538
Is this what we have come to?

>>106750631
>they've been using the same boost system forever and it wasn't properly equipped to handle a 24 core CPU safely
The cores are not the issue. They changed VID requests for at least Alder Lake onwards and it's absolutely insane, it actively tries to predict power draw and requests an overvolt ahead of time to attempt to compensate for droop... and because Intel engineers really can't see the future, the fucking things get slapped with the full voltage (in some cases like ~1.7v) when the load doesn't come, so the ring is slowly destroyed. ADL never pushed the voltages high enough to damage the ring, but RPL did. And the worst part is they fucking knew the risks but still went for a V/F increase with 14th gen.

>>106750749
>it actively tries to predict power draw and requests an overvolt ahead of time to attempt to compensate for droop
yes, i know how it works, i have a 14700K
it always did this, even back in the skylake days (but you only had 4 cores on those so it wasn't unsafe)
the prediction is not smart at all, it always compensates for the worst case scenario, if a core can pull max 50A it will compensate for 50A even if the core is sleeping in the C1 state
>ADL never pushed the voltages high enough to damage the ring
12900KS does but no one bought those

File: newkang.jpg (114 KB, 1280x720)

114 KB JPG

>>106749502
Does this mean we're about to see the 9950x3d's come down in price? I don't care if they can blow up because I run a portable air conditioner pointed directly at my server rack so everything always runs cool.

why would anybody ever buy an asrock mobo for a high end cpu?

>>106750794
>come down in price
lol

File: 12842174.jpg (216 KB, 1920x1080)

216 KB JPG

so what the tldr of the video?

File: file.png (1.47 MB, 1358x900)

1.47 MB PNG

>>106750804

ASRock's Evolving Claims
The video outlines the timeline of ASRock's public claims versus the observed changes in BIOS:

PBO and Current Limits: ASRock initially claimed the failures were related to high Precision Boost Overdrive (PBO) settings, specifically the Thermal Design Current (TDC) and Electrical Design Current (EDC) limits [04:40].

Testing of the "fixed" BIOS 3.25 showed that the actual operating TDC and EDC values remained functionally the same under load, directly contradicting ASRock's public statement [13:22].

The most significant change was an irrelevant reduction of the Package Power Tracking (PPT) limit from 1000W to 270W, leading the host to conclude that the PBO explanation was a "smoke screen" defense [15:54], [17:12].

VSOC Voltage: ASRock initially denied that VSOC voltage was an issue [05:10]. However, BIOS 3.25 changed VSOC behavior, and the later BIOS 3.40 explicitly included an option to adjust VSOC voltage parameters, indicating a silent shift in ASRock's focus toward this voltage rail [18:09].

User Data and Conclusion
Failure Data: Analysis of over 280 user-reported failures showed that 81% of the reports were associated with ASRock motherboards [39:25]. Unofficial sources within the industry also suggested ASRock has a higher failure rate when normalized for market share [39:50].

Next Steps: The investigation was unable to replicate the failure, meaning the root cause remains unknown. The video suggests the problem is likely related to something abnormal during extended low-power states on ASRock boards, with the VDDIO_MEM voltage rail warranting further scrutiny [48:35]. The hosts advise users who own ASRock boards and are concerned to consider disabling sleep and hibernate states, though no action is guaranteed to prevent the issue [48:49].

File: Boogienator.jpg (1.04 MB, 3456x2592)

1.04 MB JPG

>>106750536
>be AMD (gay blasians)
>impose specific voltages and shit to MOBO manufacturers
>do it wrong (like always)
>MOBO manufacturer doesn't know that
>he thinks he still has enough overhead (silly goy)
>MOBO manufacturer does a slight OC to get better sales with his better performing MOBO
>fries entire CPU lineups
>MOBO manufacturer goes into frenzy and does 1 quintillion BIOS updates, still doesn't know why it's happening
>brand thrown in the mud, reputation tarnished, etc...
>AMD knows but won't tell coz it would hurt their sales (and OY we don't do RMA here)
>better them than me!
>AMD : see it's not me goy, it's the MOBO DEVIL
>pays techniggers to tell cattle (you) that AMD didn't nuffin'

and this method is a timeless classic, has been used since forever (especially in the GPU department) because AMD fans are happy to eat shit, they do it for free (and no, techniggers don't actually use AMD products, they get paid for that)
nothing works but not our fault
drivers don't work on Windows (the basic bitch OS) but it's Windows fault (even though Intel stabilised brand new GPU lines faster than us), even if NVIDIA and AMD have been around for decades.
and people STILL defend this shit.

INTEL does bad thing = HEHE bad rep for IntelAVIV, billions lost

NVIDIA does bad thing = NVIDIOTS drivers sucks, they're expensive, yada yada, billions not lost because it's actually a good product anyway.

AMD does bad thing = it's the other people faults,must be the RAM, must be the Intel CPU drivers (bad mix with AMD GPU's gwailo), must be the OS, must be the PSU, must be the MOBO, must be the wrong god workshipped (not hindu), AMD could never betray and lie to the ca-consumer now would they ?

>>106750826
he doesn't know what the problem is
nobody knows what the problem is
but there is a problem because dead cpus
but Asrock can't find or fix it
and AMD can't find or fix it
and GN can't find or fix it

File: screenshot_20231105192017.jpg (48 KB, 500x505)

48 KB JPG

>>106749502
THANKS STEVE FOR KILLING LTT THEIR VIEW COUNTS NEVER RECOVERED LMAO

>>106750908
intel fired this lady :(

File: 1759025901816221.png (31 KB, 414x391)

31 KB PNG

I sit on a Asrock x870 rs with Ryzen 9900x (not x3d)

am i in danger?

File: The American.jpg (119 KB, 825x791)

119 KB JPG

>>106750826
ASUS (ASrock) can't replicate the problem in their millions dollars labs
GN can't replicate the problem with a MOBO that already killed a CPU

nah must be the MOBO surely, it can't be that AMD sells defective units, haha that would be silly

File: 1749764978853743.png (563 KB, 983x1042)

563 KB PNG

why is nobody considering that the chips may have been bad

So basically the moral of the story is people are overclocking their cpus and not providing enough cooling and frying their cpus and somehow this is AMD's fault.

>>106750884
>>106750961
yikes. yeah i think the AMD CPU is faulty to begin with. i remember some earlier ryzen dying after doing some specific linux workload.

File: 2025-09-30T23:28:43,36038(...).png (185 KB, 971x467)

185 KB PNG

>>106750961
But the thing is that other motherboard vendors aren't affected as much. Only ASS ROCK nukes CPUs that much

File: 1754802882301955.png (97 KB, 802x625)

97 KB PNG

They are never going to fix this problem btw they do not know what is wrong and they never will

>>106750986
https://www.phoronix.com/review/new-ryzen-fixed

early ryzen chips had segfault issues

>>106750982
Because they're not bad. People are overclocking their unlocked chips and not providing enough cooling to their boxes and frying it and now they are out of $600 and want a free replacement for their fuck ups. This is why we can't have nice things. I'm glad AMD is rejecting their warranty. They shouldn't offer any warranty on their unlocked chips if people are going to do retarded shit with them and not use and build them properly

>>106750999
Brilliant software AMD. OCs chips then reports they are not OC'd. Wonderful stuff.

>>106750999
Just assume its bad Ryzen samples. it's probably already fixed on the production line.

>>106750993
asrock and asus are the most common brands between prebuilts and usually anything goes wrong with em. they use the worst ram/psu ever.

>>106750999
reminder that radeon GPU drivers also install the ryzen master drivers by default
it's how they give you CPU OC controls in adrenalin

>>106750993
because they probably don't know what the fuck is going on inside the CPU (proprietary) and therefore can't tell what is wrong with their settings (OC and whatnot).
it happened before with the 7800X3D until AMD finally gave them new guidelines.

just look at the Intel problems on the 13th/14th gen, MOBO manufacturer would have NEVER known it was coming from the AI core, how would they know this ? (and btw, ALL techtubers were WRONG about the real cause)

look at this :
https://community.intel.com/t5/Blogs/Tech-Innovation/Client/Intel-Core-13th-and-14th-Gen-Desktop-Instability-Root-Cause/post/1633239
who else but the CPU manufacturer would know this ??
they make the CPU, they should be able to know what's going on and fix it.

The 3d cache tile has a much lower breakdown voltage to the rest of the chip, only way they could suddenly pop like that.

>>106751029
they shouldn't unlock their chips if they don't want people OCing them.

>>106749666
>>106749986
>>106750169
>>106750213
Will people ever admit that all this memory OCing and boosting shit is possibly not meant for a general use PC?

I never not once even enabled XMP after my PC got unstable once, I understood what it was and decided that it's not worth it for me. If I had an AMD CPU I probably wouldn't enable PBO

Nowadays people expect to run everything to the hilt and still somehow expect stability.
Where does the entitlement come from?
Do people refuse to admit memory vendors and mobo vendors are selling you a bill of goods? I mean I get it, you buy good memory and a good board and expect the best out the CPU but at some point you have to admit you're being sold a dream.

>>106751193
>boosting shit is possibly not meant for a general use PC?
No, DIY is never going to ever do that
If you don't want to do that buy a prebuilt

>>106751193
>Will people ever admit that all this memory OCing and boosting shit is possibly not meant for a general use PC?
no
because i know what i'm doing and i know how to run stress tests manually
the issue is people building PCs that shouldn't be building PCs
reddit sold those people a dream by telling them it was as easy as building LEGOs

>>106751125
>You shouldn't peel out in your 2025 mustang gt v8 if you don't know how to handle it properly and them demand a refund from Ford when you crash and total your car.

>>106751193
>Will people ever admit that all this memory OCing and boosting shit is possibly not meant for a general use PC?
it's free performance if you don't go full retard.
for DDR4 it was 3200 CL16
for DDR5 Intel it's 6400 CL32 and AMD 6000 CL30
you won't kill your RAM by going to those speeds, more than that is not worth it because it's way more expensive for no real perf gain (like 1 or 2 FPS in a game lol)
RAM has a sweet spot, no matter if it's DDR4/5 or 6 in a few years.
what's baffling is that low end CPU's are not allowed to use DDR4 still, even tho laptops don't actually need DDR5 (locked at 4800 with horrible timings, which is worse than their old DDR4 3200 they were all using).

>>106751193
literally all ryzen cpus can do 6000cl30 with zero issue.

>>106751283
The problem is mobo vendors are already doing alot just to stabilize 6000mhz cl30.
The "sweet spot" is already untenable to a degree but no one wants to admit that.

>>106751193
memory OC is harmless as long as you stay within the safespot (3200 CL16 is for DDR4, 6400 CL32 for DDR5 Intel and 6000 CL30 for DDR5 AMD) more than this is expensive and retarded (with no real gain outside benchmarks)
laptops with DDR4 were at 3200 at the end of the DDR4 gen.

>>106751252
okay nigger, i guess that explains why AMD refuses to RMA for RAM related instability if you dare saying you run with EXPO (literally can't kill a CPU)

>>106751338
>literally can't kill a CPU
Except the higher voltages the mobo gives when EXPO is enabled

>>106751308
it's only unstable on AMD because they refuse to set the RAM to gear 2 instead of gear 1 because of muh performances (at the cost of stability) unlike Intel that stays at gear 2.
it's retarded but that sells more CPU because of muh performances (almost 0 difference in games).

>>106749502
>Asrock
>shitty motherboards
I mean... This was true 3 decades ago. Do people never learn?

My 9800x3d is running fine. It wouldn't boot on the asrock board I initially got on release day, but it's been fine on the replacement asus ever since.

>>106751193
I kinda agree but at the same time if the CPU vendor sets a specific RAM frequency, and you set that speed, it should just work. there is no OCing to achieve that.

>>106751378
honestly I was inclined to choose an Asrock X870E but once I finally had the money to buy all the parts at once I opted for Gigabyte solely because of how badly Asrock communication was regarding this whole situation, and that was months ago. I won't support behavior and I certainly won't forget it either.

>>106751359
mine are 2x16GB, 6400 CL32, Gskill Ripjaws S5 XMP, at 1.35V, that's not killing anything.
my old Corsair DDR4 3200 CL16 was also running at 1.35V.
idk what AMD does but it's not supposed to be above that.

>>106749666
> Cheapass chinks
AMD is American

thanks Steve!

>>106751378
they used to make good boards in 2016, i used one with my i7 7700K and it had everything i needed for cheap (and still works to this day), it also had two M2 SSD slots and loads of expansions PCie expansion slots. (even a red led to indicate if electricity was still passing through).

>>106751420
It's not the voltage of the memory but the voltages to the CPU like the SoC
Motherboard vendors have been having to increase voltages just so pleb RAM can boot and run reliably.
It's becoming clearer however that even with "sweet spot" ram the stability still isn't great.

>>106749965
14 days

>>106751517
when I enabled expo on my asus board it wanted to 1.28vsoc by default. imo thats retarded high for 6000mhz. i set it down to 1.2v and its solid.

I'll keep saying it DDR5 was a mistake

>>106751378
Asrock boards give the most features for the best (lowest) prices. Tinkerers that know what they're doing - it's a dream board. Retard gamers that do not - it's a cheap board that will kill their CPUs.

>>106751737
works fine on GPU's (GDDR7)

>>106751737
It's fine
The mistake is thinking standards where made to please min-max tards.
There are DDR5 standards meant to improve speeds like CUDIMM but since daddy AMD doesn't support it, it doesn't exist and DDR5 is the problem and not the compulsive liars like AMD and Taiwan hardware venders

>>106751737
I mean, yeah, letting DRAM stick manufacturers be responsible for onboard voltage regulation instead of the far more experienced motherboard manufacturers was always a bit stupid.

>>106751791
>CUDIMM
btw, can you mix those with regular UDIMMs in a dual channel 4-stick setup? assuming compatible frequency/timings ofc

>>106751737
the problem is there's no need for it, if you have 32 cores that need to be fed with lots of data you should be on a quad channel platform anyway
>>106751791
>The mistake is thinking standards where made to please min-max tards.
micron used to manufacture memory explicitly for overclockers (crucial ballistix)
but then they faceplanted so hard with DDR5 they had to discontinue the brand
>>106751796
separate VDD/VDDQ is cool

>>106749502
>amd cpu explodes
>i swear it's asrock fault not the cpu!!!!!
lmao

>>106749666
>post with nose
shalom!

File: 1652636478380.jpg (131 KB, 1280x877)

131 KB JPG

>>106749502
>asscock
You're getting what you're paying for.

>>106752247
you niggers have a funny name for every brand under the sun.

>>106752726
This element is the most despicable of all of this, the double mortality from kiketubers about hardware issues, how they pretend to be neutral but they do show preferencial treatment to one but not to the other(s). In this case with Raptor Lake, most of them were quick to point Intel as guilty (and Intel was guilty, no one denies that). Now on this ongoing issue, it apparently it does not exist as it is rarely mentioned, and in the small coverage given, AMD apparently is free of guilt, as if they don't specify guidelines to motherboard vendors on the voltages that they must provide to the chips. The excuse used to hide AMD of guilt, that Ass rock is the only one with this issue, is absolutely wrong as Asus boards too face this, proving that it is not a motherboard issue, but the chip itself or the guidelines provided for them.
But here something doesn't fit, as this has been largely ignored by those same kiketubers they pretend to be "consumer friendly", and they don't lay blame on AMD, only on the motherboard manufactures. Oh, but there is no favorites! Just buy AMD!! It just works!!

>>106752860
In some ways, MB vendors do share the blame.
The shit with AMD is a compound of unrealistic standards. Vendors are doing anything to sell their slop and then AMD goes on reddit talking about shit their hardware actually isn't truly capable of.
Tech toobers also don't want to explain themselves when they say don't buy anything because their entire existence is based on graphs and telling people what to buy.

Everyone wants to delude themselves that the PC market somehow not in a bad spot.

>>106749502
asrock am4 mobos don't do that. It indicates an x3d problem.

>>106749502
use case for ever buying new hardware?

>>106753026
Use case for into be buyingly new: Make old cheap for me.

I wonder if asrock or AMD has a interpose to measure current and voltage for each pad of the socket.

Steve is spiritually trans. Any adult male who makes drama into their main source of income is just a woman in a man's body. You will seethe about this because you know I'm right.

>>106750908
it's like come on man you made a video about VR headsets from 1994 and you think it's going to get millions of views lmao

>>106753072
teach me

>>106753487
shut the fuck up zoomer, not every faggot is trans some are just gay

>>106751193
It was always a thing. The problem simply got worse as we pushing towards hard limits of physics and demand reliable data transmission. DDR4 and DDR5 are much more touchy then their predecessors. Running them beyond JEDEC-spec without them crapping themselves is rolling the dice.
The barrier of entry is too low. You use have to manual input the timing and clockspeed before XMP/EXPO were a thing. Now it is just a click and done affair. To be fair, on both ecosystems you do get warning that enabling XMP/EXPO mode may cause data corruption and instability before you turn it on. Most people don't bother reading this warning or the manual that outlines this.

>>106751737
No, it isn't. The problem is DRAM itself being limited to 32-bit bus and having to operate DIMMs at insane clockspeeds starting with DDR4. That's why CUDIMM is being entertained in the first place. It is just trying reduce tracing and distance involved. We are approaching the hard limits of what can done with electrically conductive mediums.

>>106752726
The problem with Raptor Lake was that most of the K series yields were misbinned from the started. Intel need keep-up on silly gayming benchmarks for optics and thought their 7nm node would tank like their 14nm node (see Coffee Lake R/Comet Lake).
If Raptor Lake K series SKus were binned correctly (See the Xeon versions). They would be only somewhat faster then Alder Lake and fail to catch up to Zen4 at gayming shit. Intel would be forced to sell K at lower price points. Intel marketing and shareholder would not accept this. This is what led to Raptor Lake K series fisaco. Intel cannot admit this openly as it unleash of can of legal worms. They are just downplaying as "firmware" issues with a patch a.k.a dialing down the boost.

>>106751961
>ballistix
They didn't make $$$$ nor stand out from the other factory overclocked memory incumbents. It was an experiment that ended. Pure business decision.

>>106753601
>. The problem is DRAM itself being limited to 32-bit bus
Each DIMM is 64 bits actually
>That's why CUDIMM is being entertained in the first place. It is just trying reduce tracing and distance involved
CUDIMM didn't do anything about trace length, it added what is essentially buffer for the clock signals. Accepting the clock signal and regenerating it.
I\s also not the limit, not close. Registered/buffered RAM allows for much more RAM and higher clocks than what unbuffered DIMMs allow. It's just not something that has been afforded to regular desktop yet.