What the fuck are you going to do about it nig?

>>108660844
you mean calibration?
its the same as all the other parts of POST

>>108660844
holy shit we have a memory training schizo now.

>>108660858
My old PC can boot from power off to Windows login screen in a couple of seconds.
Now apparently it takes at least 20 seconds to just get out of POST.

File: 1763325549017133.png (187.9 KB)

187.9 KB PNG

>>108660897
there are massive differences between MBs look it up before you pick one. i cant find benchmarks that only go up to the bootloader they include winblows in this shit. still it blows my mind that they cant make an MB that basically goes instantly to the bootloader

File: macbook-6_hgfm.jpg (58.5 KB)

58.5 KB JPG

>>108660981
And then PC nerds wonder why people prefer this.

>>108660844
subspace quantum leaks and depolarization

>>108660857
Something a white man couldn't do, clearly.

>>108661722
Yeah but this is the pc nerd place not the normie zone. Who sent you?

>>108660863
I can memorize a deck of playing cards. Can you?

>>108662381
I once memorized pi. All of it.

>>108661756
I don't believe in quantum mechanics, the cat is either dead or alive whether someone is looking at it or not.

>>108660844
Memory training has been a thing since before you were born. It's just more prominent these days because training speeds haven't kept up with memory capacities. Also, modern consumer hardware shouldn't have to re-train on every boot unless it's defective or misconfigured. Fix your shit.

>>108662392
Very impressive. Now we can finally have someone who can solve the elusive Math problems that involve a circle.

>>108660844

anon with Asus boad its possible to disable AEMP II so it will memory train only after install or long off time

File: 1754134338537766.jpg (140.5 KB)

140.5 KB JPG

>>108661722
you prefer that one because you can see it with they eyes and yo nigga hand be fire grab that shid off fag whitey yo. meanwhile whats a nigga gonna do to take a PC? he gotta unplug everythang and try not to get bit by they dog carrin out that heavy shieeet.

>>108662758
But if nobody is looking at it, which one is it? Dead, or alive?

>>108661722
https://www.lttlabs.com/articles/2026/03/25/macbook-thermal-testing

>>108660981
My budget Dell PC I bought 5 months ago boots up in like 10 seconds max, often faster

>>108665371
Whichever is true

File: m1-chip-upgrade-ram.jpg (662.3 KB)

662.3 KB JPG

Anything to foolishly resist the inevitability of on-chip or closely soldered memory

>>108666236
on package memory is ideal for consumer slop that gets replaced every 2 years

>>108665598
your dell pc has less memory bandwidth than a mobile phone

File: 1763784016745659.jpg (420 KB)

420 KB JPG

>>108665598
they really have to advertise stuff like this if they started adding 'boots in 10 sec' on the box i bet a lot of people would go for those

>>108666236
i think in the future cpus,ram and ssds will all use very similar looking interfaces. might even use the same one with different 'keys' like different m.2 cards today.

>>108660844
Because storing the calibration parameters somewhere between boots is apparently too much to ask.

File: file.png (860 KB)

860 KB PNG

>>108661952

File: 1674003948121723.png (846.9 KB)

846.9 KB PNG

>>108660844
i haven't turned off my computers for 2 months each why would i give a shit if it takes 20 seconds longer to boot, boot time is the most irrelevant metric on earth, ever since SSD it does not matter

>>108665388
there is a regulation i think about low temperature burns, so if you do the mod, you can get a low temp burn if you use it on your lap

>>108660844
Just get some Japanese dude to anthropomorphize it into a cute anime girl and reframe it to “correction.”

>>108665388
Bro you can just enable power saver mode and have the cpu consume over 70% less power and have 50% less performance as a downside, which still matches up roughly with an 5700x.

>>108666804
>why would i give a shit
Because PC fans are loud as fuck and you can't reliably use sleep because of terrible power management implementation in both Windows and Linux.

>>108667421
What do you mean, reliably? I've never had issues with sleep on any OS.

>>108666272
so basically, SoM's like the PI compute module? and the "motherboard" is just a carrier board with all the IO and power delivery stuff

>>108669318
PC turning itself on (because of updates, or whatever, sometimes no reason) and then never going back to sleep (unless you intervene) because fuck you, that's why.

>>108660844
I just go get some water and it's done before I get back. Who cares if it takes a little time to boot. Under a minute and it's good for the entire day.

Its getting to the point were everything is bleeding edge, and they are producing so much, that the training cant be realistically dont in facility, and set to a base line. And with the temperature, and humidity, and hardware, and power supply degradation, and input voltage from the wall, and bios version, and os differences, and preset settings, you can tune stuff to be a hell of a lot faster, instead of relying on the manufacturer baseline tune.

>>108669548
>who cares that my city is flooded with more and more brown people every day
>I close my eyes when I walk home, close the blinds and it's great at the end of the day!
This is you.

File: 1757013895606859.jpg (37.8 KB)

37.8 KB JPG

>>108669358
>so basically, SoM's like the PI compute module?
im talking about the physical interface, compute modules of raspberry use s D-PHY connector (pic related)

i think the new framework laptop is one of the first if not the first computer with LPCAMM2 which has a land grid array connector. i think its starting with laptops but will soon move on to all PC ram and all PC SSDs being connected via land grid connectors just like the CPU

>is just a carrier board with all the IO and power delivery stuff
i mean that's what a motherboard is

>>108669584
What is the amount of RAM people have now?
64 GB? 128GB if rich (and kinda pointless since its slow RAM for the CPU and not proper fast GDDR on your graphics card you could do actually anything useful with)?

So, 4 times, at best 8x the amount of RAM on my 15 year old PC?
Wow, such bleeding edge, such progress...

>>108669665
BTW: graphics cards don't do any "training" bullshit (at least don't take any significant time to do it) despite their RAM being way way faster.

>>108669665
Yeah doofus im talking about actual bleeding edge. Not ddr4 or ddr5. Im talking about hbm3/4 and ddr7+ ram. Also, did you even fucking read everything I wrote? Do you actually think all these factors DONT play a role in the performance of these chips?

>>108660844
>you have to "train" your memory (whatever it means) now every boot?
you have to train every PCIE connection to adjust filter paramaters, every motherboard is different so every GPU goes throug a training cycle

>>108660897
>Now apparently it takes at least 20 seconds to just get out of POST.
select an xmp profile

File: Mac_Studio_(2022)_front.jpg (3 MB)

3 MB JPG

>>108669843
>to make your computer work you must set blipingoo doop to plapity pep in floopity bloop settings after you enter gliptidy moo menu
And then you nerds wonder why normal people choose products that just work from companies that respect them.

>>108660897
>My old PC can boot from power off to Windows login screen in a couple of seconds.
No. It cannot. It wasn't powered off, just hibernating.

>>108670102
>company respects you so much it's paying a small army of shill to spam your friendly neighborhood forum with unwanted advertisements
Of course, Chang... Of course...
Please contact the nearest CCP agent, and get your organ harvested already.

My desktop booted windows 10 in maybe 15 seconds. Windows 11 I haven't bothered to check, but Linux boots in probably 20. My laptop boots extreamly quick into linux. Probably 10 seconds.

File: IMG_3023.png (208.8 KB)

208.8 KB PNG

>>108660844
Thank you OP for reminding me yet again that there is still no good reason for me to upgrade from my Ryzen 9 5950X paired with 128GB of CL16 DDR4 3600.

>>108660844
>>108660897
hell on earth

>>108666272
I hope so, the dimm format just isnt working anymore, servers are stuck at dual socket 24 channels because of the size.

>>108670270
You live in India so you're accustomed to being surrounded by shit.

>>108670224
>windows: 15 seconds
>linux: 20 seconds
>out of linux user mouth
Linux B T F O

>>108670405
>boot is 5s slower on cherry-picked hw
>literally unusable

>>108670850
It starts at boot - literally the beginning - and only gets worse from there.

File: images(17).jpg (8.3 KB)

8.3 KB JPG

>>108670926

>>108670264
There isn't. Used to run a system with 4x16 DDR4 3600CL14. Upgraded the whole thing, wanted to run 4x32GB DDR5 from the manufacturers QVL and it took over an hour to train, 20+ minutes to retrain every other boot with all DDR5 cope options turned on. Only solution is accepting that DDR5 is a scam since using more than two DIMMs might technically work but it's still not in a state I'd call anywhere near acceptable considering the hefty fucking markup DDR5 always had over DDR4 while offering no benefits. If you can just skip DDR5 altogether. Seriously it's that fucking shit. I regret upgrading and not taking my own advice and waiting for DDR6 to fix DDR5 fuckups.

File: 0xixvxvc32jf1.png (190 KB)

190 KB PNG

>>108670972

File: 1753335725814667.png (115.8 KB)

115.8 KB PNG

>>108671066
old stats. 43 is out, 44 is almost done.

Unless you're doing something retarded, memory training should only ever happen after changing the physical configuration of the memory in any way (including just reseating) or changing xmp/docp/manual tuning profiles, and it should only happen once until the next change. What retard shit are you doing to make this an issue?

>>108671158
>y-you're h-holding it w-wrong!
No faster way to admit your product is shit than pulling out this tired cope.

>>108660844
>DDR5
im still on ddr3
its literally fine? i don't *need* ddr4
luckily im not using windows and forced to update to slower and slower operating system so i haven't needed to update ram in years.
i do have a work laptop on w11 and every time i leave it alone for 5 minutes the fan spins up to 100% like its about to take off. i just use it for emails what the fuck is it indexing? so i can see that some cucks will think oh yes please the latest thing, 'latest and greatest' as my indian colleague says.

but its not for me. dont need it thanks.

File: 1764823913416339.gif (2.3 MB)

2.3 MB GIF

>it's just one of those low quality troll threads jannies let shit up the board
ebin

>>108671158
everyone complaining about this is on AM5
(you can figure out the rest)

File: IMG_8652.jpg (43 KB)

43 KB JPG

>>108671941
you mean glued together CPUs have latency issues? Shocker!

>>108671259
DDR3 is still amazing if you have a morherboard capable of supporting triple channel which i do on my Mac Pro

>>108671265
It's a real issue with DDR5.

>>108671158
am5 is shit
you can configure bios options to reduce trainings but at the very least it will still happen twice a month

>>108671941
>seen it personally happen with Xeons
>seen it personally happen with EPYC
>seen it personally happen with AM5
>haven't seen it happen with current gen Intel because I haven't bought any consumer intel chips since the 8700K
Hmmmmmmmmm must be an AM5 problem.

>>108675886
No it's not.
t. using DDR5, zero training required beyond the first time enabling the xmp profile for the sticks after building the system.

>>108660981
Gone are the days of fast booting.
I have a B550 MSI board and a 5900X and this shit takes a good half a minute boot into linux. I remember booting in half of that with an old machine on 8.1 (and with fast startup disabled, so there's no excuse there).
Nowadays, I turn on my mac mini with macOS installed on an external drive, and that takes about 15 seconds tops.

File: 1768823299929580.jpg (229.5 KB)

229.5 KB JPG

>>108671066

>>108678278
>pleb settings

>>108678278
mine doesn't have a "max upscale" option

>>108678278
Why the actual fuck are you looking at benchmarks for a brand new game using NINE YEAR OLD GPUS and presenting that as an argument in favor of amd/linux?

>>108660844
>So, you have to "train" your memory (whatever it means) now every boot?
AMDrones put up with some real retarded shit all in the name of 6 more 4fps at 4K

>>108660844
DRAM always had to be trained?

>>108661984
doesn't really feel like it these days

>>108666804
I think on Ryzen 9000s sleep fries the CPU, so you would need to keep it idle not sleeping or using lower c states

>>108666236
why don't they just make little lga sockets for bare dram/nand chips?

>>108660844
to adapt other components to ram speed and timings?

>>108660981
>replace my ancient cheapo b350m with high end x670e
>boot goes from ~12s to ~35s
>sometimes ram fails to clock high enough on boot and the entire pc is a jittery pile of shit so you need to restart and hope for good rng (no im not joking, admittedly I run ram at very high oc)
sigh

>>108660844
Memory training is a hack to support replaceable DIMMs at high memory speeds.
Soldered RAM would fix this but you niggers don't want to accept it because... MUH modularity o algo.

>>108661984
No it's /pol/ with slight tech aestetics these days, pretty sure that 80% of regular users have never typed a single line of code

>>108669611
A brief delay I don't care about is nothing like a populace replacement. Your argument is either ignorant, or intentionally stupid. No further engagement required.

>>108660981
Currently using an MSI Z790 Pro-A Wifi, maybe 5 seconds boot time with I5 13600k + 2x16GB DDR5 6400mhz CL32.
Old PC with Asrock Z270 + I7 7700k and 4x8GB DDR4 at 2133mhz CL16, almost same boot time.
Relative's PC with i5 12400 and a cheap Gigabyte M-ATX MOBO (B at that) B760M Gaming X AX + 2x8GB DDR4 3200MHZ CL16 , almost same boot time.
Yeah no, shitty post times only happen with AMD garbage CPU's, they had the same problems with the 7800X3D, same with 9800X3D and 9950X3D...
Those CPU's also hate RAM with high clock speed unlike Intel and eat up to 35 watts when idle (lol).
Can't wait for Intel to get back on its feet and make better gaming CPU's (it will happen), AMD always fucks everything up somehow, like 9800X3D still frying themselves for stupid reasons.

File: file.png (269.1 KB)

269.1 KB PNG

>>108660857

>>108678880
whats more retarded is presenting outdated stats like in >>108671066, linux and proton move very fast

>>108687100
Prove it got better.

>>108672521
Three channels of ddr3 1866 barely match a single ddr4 3200 stick.

>>108660844
I have a 9800x3d and am using a 6000 cl30 kit w/ expo. the initial training took a while, but it saves the config and reloads it every boot. it's been ~10 months and the ram has retrained 3 times on boot with me booting my computer at least once a day (shorter than the inital, but still like a 1-2 minutes). I'm not sure if it's a coincidence, but it has only happened after windows updates so maybe something changes and windows requests a retrain? anways, it basically doesn't matter, you will spend more time with your pc being unusable due to updates than from ram training

I am actually curious if someone had left their pc on for the same 10 month period as me if the ram would have become unstable or performance worsened since it doesn't recalibrate every day / on those rare times it needs to retrain

>>108688000
9800x3D 6000Mhz CL30-36-33-126 here. (manually tuned, no EXPO/XMP)

I built my PC in mid December and have turned my PC off twice in that time, once in February and once again just a few days ago.

Zero issues and my RAM didn't do any retraining, but the two times I restarted were for windows updates, so boot took longer than "normal" on both occasions due to that more than anything else.

But yeah, I basically only restart every 2-3 months and so far zero retraining after the first initial training back in December.

>>108660981
>>108678099
>>108684453

LOL. Us Intel chads don't have this problem. My system boots in less than 5 seconds. Also run 4x16 DDR5 6200 Mhz with perfect system stability.

>Memory training

HAHAHAHAHA. AMD tards have to train their memory with every single boot. Imagine actually paying for this trash.

>>108671941
>Memory context restore : Enabled
Problem solved.

>>108689212
right but AMD is the normalfag brand now so changing a BIOS option is too much for most users

>>108689242
Let's not be too charitable to AMD. Early AM5 AGESA was terrible and turning this on had a 50/50 of causing failure to boot. Something like this should be on by default and the reason why it wasn't was purely AMD's fault. It's still not in a good enough state to have on by default.

>>108670264
>there is still no good reason for me to upgrade from my Ryzen 9 5950X paired with 128GB
There isn't, unironically. You have WAY more RAM than most end-users have even a decade past when it was "good enough" (8GB). Even 16GB is the bare minimum if you're doing gaming.

If you're 32+GB, you're in complete overkill territory and unless you're doing something intensive (running a complete server off that hardware) it's complete overkill for most UseCase?.meme's.

>>108669365
That's a Windows problem. I don't recall Linux waking from S3 to do an update because Linux respects the user and PROMPTS an update request instead of doing it because it's users are iToddler-tier.

Windows is the only time I've had the computer wake from sleep and never go back to sleep because of an update.