Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
 
  • Post
  • Reply
Truga
May 4, 2014
Lipstick Apathy

Subjunctive posted:

WTF is "AMD Link"?

literally just steam link but red, afaik

Adbot
ADBOT LOVES YOU

HalloKitty
Sep 30, 2005

Adjust the bass and let the Alpine blast

Subjunctive posted:

WTF is "AMD Link"?

I hadn't heard of it either. It's a streaming program that lets you.. stream your games from your PC to something else.

runaway dog
Dec 11, 2005

I rarely go into the field, motherfucker.

Subjunctive posted:

My ROG Strix X670-E Gaming Wifi has a new BIOS (1202) as of April 21, which I'm hoping has the fix I need because I just installed it!

yeah I got that one too, I figured that the if hero got a new one one a few hours ago that an even newer one is otw

hobbesmaster
Jan 28, 2008

Paul MaudDib posted:

Yes, the idea of 1.3v being normal is kind of crazy, AM4 is more like 0.8-1.1v right?

I get the impression this Io die is really really iffy, AMD has always been a bit crappy compared to intel (apart from like, X99) but this time they can’t even run all the slots populated etc and intel didn’t have those problems even on their first gen DDR5 stuff (or X99!). If there ends up being a new IO die next gen it’ll be interesting to see how it does.

Default is about 1.0V I think, the fabric voltages or whatever were the ones at 0.85 and 0.95. It’s like 1.1V was considered something fine to blindly set for XMP, 1.15V was getting a bit more aggressive with 1:1 FCLK, 1.2V+ was for trying to set a record with your B die at 1.6V and not caring if you burn everything out.

AutismVaccine
Feb 26, 2017


SPECIAL NEEDS
SQUAD

Dr. Video Games 0031 posted:

Correct me if I'm wrong, but heat transfer doesn't have any hard "bottlenecks" like this, and you still get a lot of benefit from water cooling. I've seen plenty of results from people using AIOs and custom loops on AM5 that show clear improvements to thermal performance, without delidding. The cooling solution doesn't become irrelevant just because of some heat dissipation limitations in the silicon.

Pretty much this. The bigger the heat difference, the better heat flows

Paul MaudDib
May 3, 2006

TEAM NVIDIA:
FORUM POLICE
It’s all about thermal deltas. If the hot side is X degrees hotter than the cold side at steady state, for a cooler of a given quality (delta-t over watts? Or whatever) you’ll move Y watts. Any given cooler can move more watts at a higher delta-t, and vice versa.

Like the reason blower coolers run hot is they just need a really high delta-t because they need a high temperature differential to move any heat because they suck.

Now if you really want to push maximum cooling what you need is not a radiative cooler but a heat pump :q:

https://m.youtube.com/watch?v=185VQ_d2bYU

(h/t Swiss Army Druid forever ago, I believe)

gradenko_2000
Oct 5, 2010

HELL SERPENT
Lipstick Apathy

Subjunctive posted:

WTF is "AMD Link"?

AMD Link is AMD's equivalent of Nvidia game streaming. It's technically a GPU feature, but their APUs/SOCs are capable of it.

HalloKitty
Sep 30, 2005

Adjust the bass and let the Alpine blast

Paul MaudDib posted:

It’s all about thermal deltas. If the hot side is X degrees hotter than the cold side at steady state, for a cooler of a given quality (delta-t over watts? Or whatever) you’ll move Y watts. Any given cooler can move more watts at a higher delta-t, and vice versa.

Like the reason blower coolers run hot is they just need a really high delta-t because they need a high temperature differential to move any heat because they suck.

Now if you really want to push maximum cooling what you need is not a radiative cooler but a heat pump :q:

https://m.youtube.com/watch?v=185VQ_d2bYU

(h/t Swiss Army Druid forever ago, I believe)

Strapping a freezer to your cpu used to be a "mainstream" thing, in that you could run out and buy a Prometia or Vapochill case ready to go.

I have friends that did have them back in the day, when we thought it was a good idea to buy a mid range chip to save money, then buy a massively expensive cooling system to gain performance

One of my friends had a lovely experience when removing a cpu, the pins had become extremely brittle

HalloKitty fucked around with this message at 09:48 on Apr 26, 2023

Bloodplay it again
Aug 25, 2003

Oh, Dee, you card. :-*

Dr. Video Games 0031 posted:

Gigabyte has joined Asus in straight-up deleting the first several bios versions from their support pages. I suspect the latest bios versions will be fine to use with the X3D chips. I'm on the latest B650 Aorus Pro bios, and I am not noticing any anomalies with any of the voltage readouts with expo enabled.

Are you looking at the listed voltage in the BIOS or using hwinfo64? I'm looking at hwinfo64 and seeing CPU VDDCR_SOC (SVI3 TFN) listed at 1.245V with EXPO enabled at 6000. When I look at the BIOS, all of my Vcore-related options are set to Auto and I'm unable to find a listed voltage like with the UI from ASUS. Just wanna make sure I am looking at the right value.

Is there any way to see the automatically-set voltage in the BIOS? Only way I can see them is by setting manual values. I am still using f3c from Nov. 14 2022, fwiw.

Paul MaudDib
May 3, 2006

TEAM NVIDIA:
FORUM POLICE

HalloKitty posted:

Strapping a freezer to your cpu used to be a "mainstream" thing, in that you could run out and buy a Prometia or Vapochill case ready to go.

I have friends that did have them back in the day, when we thought it was a good idea to buy a mid range chip to save money, then buy a massively expensive cooling to gain performance

One of my friends had a lovely experience when removing a cpu, the pins had become extremely brittle

the novel thing is that this is the difference between an A/C unit and a heatpump. this is a really direct way to move heat from one side to the other - it's a mechanical peltier. And the other side gets really hot, but it's air, so, if you can design around expelling a giant jet of 400C air then awesome!

over a million rpm

Cao Ni Ma
May 25, 2010



Gigabyte removed all the previous bios from the web page so I'm thinking its more than just a manufacturer doing something weird

Enos Cabell
Nov 3, 2004


I built my 7800x3d system on Monday evening with a Gigabyte x670 board that came on firmware F6. Booted with xmp/expo(not sure of the difference, bios calls it both) enabled and seemed to work fine. Yesterday I updated to the most recent for that board, F8a which came out last month, and it would no longer boot with xmp enabled. This morning F8c was out, so updated again and it's back to working fine with xmp enabled.

:shrug:

Dr. Video Games 0031
Jul 17, 2004

Enos Cabell posted:

I built my 7800x3d system on Monday evening with a Gigabyte x670 board that came on firmware F6. Booted with xmp/expo(not sure of the difference, bios calls it both) enabled and seemed to work fine. Yesterday I updated to the most recent for that board, F8a which came out last month, and it would no longer boot with xmp enabled. This morning F8c was out, so updated again and it's back to working fine with xmp enabled.

:shrug:

Oh, looks like gigabyte pushed a new update for all of their AM5 boards. It's probably a good idea to jump right on that, given the circumstances.

Cao Ni Ma
May 25, 2010



The gigabyte bios with letters on them are beta bios so be aware of that

Enos Cabell
Nov 3, 2004


Cao Ni Ma posted:

The gigabyte bios with letters on them are beta bios so be aware of that

Good to know! I had to leave for work shortly after updating mine to F8c, but it made it through a CP2077 bench and a few 3dmark runs so fingers crossed.

Dr. Video Games 0031
Jul 17, 2004

Cao Ni Ma posted:

The gigabyte bios with letters on them are beta bios so be aware of that

yeah, but i'll take potentially buggy behavior over the potential of my cpu and motherboard spontaneously combusting.

Cao Ni Ma
May 25, 2010



Dr. Video Games 0031 posted:

yeah, but i'll take potentially buggy behavior over the potential of my cpu and motherboard spontaneously combusting.

I think I remember debaur saying that the older bios were removed from websites and that the newer ones had potential fixes already applied (probably looking into more), the only extra line that the new beta bios for my motherboard has is "Optimized for Ryzen 7000 X3D series CPU"

I still swapped to it though just in case

Kibner
Oct 21, 2008

Acguy Supremacy
A screen shot of the two most recent BIOS for my Asus motherboard:



I'm holding off on the beta BIOS, especially since it seems like the production one I am on already implements some better monitoring.

Klyith
Aug 3, 2007

GBS Pledge Week
For people who are continuing to run EXPO/XMP with a older possibly-bad BIOS and just checking your SOC voltage to make sure it's safe, you should check while running something that stresses the ram. Prime95 has an option that hits main memory.

I may be wrong about this, but I saw on twitter that the SOC voltage is now more variable on AM5? On AM4 ryzens VSOC was pretty static -- I get an extra .05V or so with heavy memory load. If that's higher now it wouldn't take much to go from safe to dangerous depending on test conditions.


Bloodplay it again posted:

Is there any way to see the automatically-set voltage in the BIOS? Only way I can see them is by setting manual values. I am still using f3c from Nov. 14 2022, fwiw.

If gigabyte's 650/670 BIOS layout is the same as their X570 was, look at Settings -> PC Health (in advanced mode). That will give you a readout of sensors including SOC. Not exactly what you want, but a good 2nd confirmation of hwinfo.

Cao Ni Ma posted:

The gigabyte bios with letters on them are beta bios so be aware of that

This is a case where I'd easily prefer running a new beta BIOS rather than a old BIOS known to destroy CPUs.

And if I didn't want to run a beta BIOS, I'd probably be resetting to stock and running with no EXPO/XMP at all until this poo poo gets definitively cleared up.

Dr. Video Games 0031
Jul 17, 2004

I have checked that, and I haven't noticed any notable fluctuations in SOC voltages.

Gigabyte put out a press release about the new beta bios, and they do say that it tightens up the voltage ranges to prevent any issues: https://www.techpowerup.com/307805/gigabyte-releases-updated-motherboard-bios-preventing-ryzen-7000x3d-cpu-issues

I haven't really noticed any changes after updating, but all of my voltages were in the acceptable range already anyway.

edit:

quote:

Meanwhile, through the GIGABYTE exclusive Performance Bung in the BIOS option, users can easily optimize the tuning process of CPU voltage setting,

Kibner
Oct 21, 2008

Acguy Supremacy
Do you think there will be a way to check for damage that does not involve unmounting the cooling solution to visually check the CPU?

I did get the Microcenter insurance for my motherboard and maybe my CPU, so maybe I should just take advantage of that once this all gets sorted out.

Klyith
Aug 3, 2007

GBS Pledge Week

Kibner posted:

Do you think there will be a way to check for damage that does not involve unmounting the cooling solution to visually check the CPU?

If you have visual damage, the CPU is already dead. The bulging happens during the final failure when it burns itself out.

So no, there's zero way to tell if you have degraded / unsafe silicon until it pops.


Kibner posted:

I did get the Microcenter insurance for my motherboard and maybe my CPU, so maybe I should just take advantage of that once this all gets sorted out.

On the side on not worrying about this too much: how many CPUs do we think have actually failed this way? This isn't widespread, it's just extremely visible because it happens to enthusiasts who post about it.

Like, prerequisite number one is to have high-end ram in the DDR5-6000 zone. The unsafe voltage gets selected at high memory speed. But I'm sure way more people have fast memory than have burned CPUs. We're not seeing hundreds of people with the problem. So it seems like only a few CPUs that are exposed to dangerous voltage actually pop. (At least, so far -- they'll get worse over time.)

hobbesmaster
Jan 28, 2008

Klyith posted:

I may be wrong about this, but I saw on twitter that the SOC voltage is now more variable on AM5? On AM4 ryzens VSOC was pretty static -- I get an extra .05V or so with heavy memory load. If that's higher now it wouldn't take much to go from safe to dangerous depending on test conditions.

Unless something changed that should be a “static” voltage, if there is an increase with load instead of a decrease it’s due to load line calibration. .05V is actually a lot for the zen2/3 IOD but if you’re at like 1.05 default it might be why your system is stable. You probably would have a flat voltage by bumping LLC down a couple notches.

kliras
Mar 27, 2021
reminder to check if you can get a complimentary jedi survivor code if you bought a 7X3D cpu lately, in case you want something new to throw at your new explode-y cpu

game is out in two days

Zedsdeadbaby
Jun 14, 2008

You have been called out, in the ways of old.
PC reviews say it's yet another bad PC port with poor CPU usage problems.

Ardryn
Oct 27, 2007

Rolling around at the speed of sound.


Dr. Video Games 0031 posted:

I have checked that, and I haven't noticed any notable fluctuations in SOC voltages.

Gigabyte put out a press release about the new beta bios, and they do say that it tightens up the voltage ranges to prevent any issues: https://www.techpowerup.com/307805/gigabyte-releases-updated-motherboard-bios-preventing-ryzen-7000x3d-cpu-issues

I haven't really noticed any changes after updating, but all of my voltages were in the acceptable range already anyway.

edit:

I feel like I'm going crazy because I have a BIOS version labeled F8a while the only version I see on the gigabyte site is F7, I downloaded F8a back when I assembled the 78 machine. Should I still flash to F7 or just sit on F8a?

Which SoC in hwinfo am I supposed to be worried about here? VDDCR_SoC is 1.245 and VCORE SoC is 1.296 but goes up to 1.308 at times.

hobbesmaster
Jan 28, 2008

Zedsdeadbaby posted:

PC reviews say it's yet another bad PC port with poor CPU usage problems.

So it’s a good pairing with the x3Ds?

Klyith
Aug 3, 2007

GBS Pledge Week

hobbesmaster posted:

Unless something changed that should be a “static” voltage, if there is an increase with load instead of a decrease it’s due to load line calibration. .05V is actually a lot for the zen2/3 IOD but if you’re at like 1.05 default it might be why your system is stable. You probably would have a flat voltage by bumping LLC down a couple notches.

Aha, that makes sense! And yeah my SOC is running at default voltage. The ram I have now is not particularly great, and I wasn't able to see worth-while improvements from putting more juice into vdimm and vsoc. So I said gently caress it and run the memory pretty vanilla.

I definitely don't want to play around with the LLC since I have a stable undervolt and that did improve my boost clocks.



Ardryn posted:

Which SoC in hwinfo am I supposed to be worried about here? VDDCR_SoC is 1.245 and VCORE SoC is 1.296 but goes up to 1.308 at times.

These are probably the same thing, but measured by different sensors and at different locations. So the low one has resistance losses or something. Even the worse of the two is 1.3V -- high but not dangerous. The mobos that are killing CPUs were reportedly going higher than that.

You could try setting your SoC manually to 1.2 or 1.25 and see what happens
a) do you still boot & run stable?
b) how does that change the measured voltage in hwinfo? (For example, if you set it to 1.25V and get the same numbers as above, you can relax and go back to auto. One of the sensors just reads higher than what's actually going into the CPU, which is a normal thing.)

Ardryn
Oct 27, 2007

Rolling around at the speed of sound.


Klyith posted:

Aha, that makes sense! And yeah my SOC is running at default voltage. The ram I have now is not particularly great, and I wasn't able to see worth-while improvements from putting more juice into vdimm and vsoc. So I said gently caress it and run the memory pretty vanilla.

I definitely don't want to play around with the LLC since I have a stable undervolt and that did improve my boost clocks.

These are probably the same thing, but measured by different sensors and at different locations. So the low one has resistance losses or something. Even the worse of the two is 1.3V -- high but not dangerous. The mobos that are killing CPUs were reportedly going higher than that.

You could try setting your SoC manually to 1.2 or 1.25 and see what happens
a) do you still boot & run stable?
b) how does that change the measured voltage in hwinfo? (For example, if you set it to 1.25V and get the same numbers as above, you can relax and go back to auto. One of the sensors just reads higher than what's actually going into the CPU, which is a normal thing.)

Weirdly, the VDDCR went up to the 1.25 but the VCORE stayed the same when at "idle" and went up the same amount when under load, like firing up rimworld with too many mods. I think I'll just go back to auto, the only reason i was concerned in the first place was the corsair vengeance 6000 ram I got had those buildzoid tight timings. Thanks for the assistance at any rate!

Ardryn fucked around with this message at 19:17 on Apr 26, 2023

wargames
Mar 16, 2008

official yospos cat censor

Enos Cabell posted:

I built my 7800x3d system on Monday evening with a Gigabyte x670 board that came on firmware F6. Booted with xmp/expo(not sure of the difference, bios calls it both) enabled and seemed to work fine. Yesterday I updated to the most recent for that board, F8a which came out last month, and it would no longer boot with xmp enabled. This morning F8c was out, so updated again and it's back to working fine with xmp enabled.

:shrug:

when i go to the page for the gigabyte b650 aorus elite ax i only see F4 came out on 2023/03/22

kliras
Mar 27, 2021
this is just every major pc release from now on i guess

https://twitter.com/EAStarWars/status/1651294688786403329

Enos Cabell
Nov 3, 2004


wargames posted:

when i go to the page for the gigabyte b650 aorus elite ax i only see F4 came out on 2023/03/22

Huh, weird. I just went back to the page for the x670 elite ax and this morning it showed a bunch of downloads, F8c being the most recent, and right now the only one listed is F7.

https://www.gigabyte.com/Motherboard/X670-AORUS-ELITE-AX-rev-10/support#support-dl-bios

Lowen SoDium
Jun 5, 2003

Highen Fiber
Clapping Larry

Enos Cabell posted:

Huh, weird. I just went back to the page for the x670 elite ax and this morning it showed a bunch of downloads, F8c being the most recent, and right now the only one listed is F7.

https://www.gigabyte.com/Motherboard/X670-AORUS-ELITE-AX-rev-10/support#support-dl-bios

Same, and I just updated to F8c. Now I am not feeling very good about it.

Cygni
Nov 12, 2005

raring to post

https://twitter.com/GamersNexus/status/1651308183745331202

Its been documented here a few times on forums.SomethingAwful.com, the best hardware website, but the board manufacturers have a history since at least the Z170 days of running higher than stock voltages and enabling obtuse auto-overclocking "features" when users did things like enable XMP in order to win barchart graphs. I wonder how much of this latest ASUS disaster is going to end up being related to that practice?

Cao Ni Ma
May 25, 2010



Enos Cabell posted:

Huh, weird. I just went back to the page for the x670 elite ax and this morning it showed a bunch of downloads, F8c being the most recent, and right now the only one listed is F7.

https://www.gigabyte.com/Motherboard/X670-AORUS-ELITE-AX-rev-10/support#support-dl-bios

Lmao god loving drat it. Maybe I should go back to 7

repiv
Aug 13, 2009

gigabyte are leading the way with their innovative BIOS gacha system

the rare drop is your CPU doesn't catch fire

Klyith
Aug 3, 2007

GBS Pledge Week

Cao Ni Ma posted:

Lmao god loving drat it. Maybe I should go back to 7

Just reset your BIOS to defaults and leave everything running at stock clocks for the next week or so. Then you don't have to care about which BIOS is safe or will explode your CPU.

Oh no, losing 5 frames per second in a video game, the horror.

BlankSystemDaemon
Mar 13, 2009



Having to run the firmware at stock settings for a few weeks while this gets figured is hardly the end of the world.

My former workstation motherboard spent almost its entire time with dead SATA2 ports and only two SATA3 chips working, and that was impossible to fix in software, so I just lived with it.

Zedsdeadbaby
Jun 14, 2008

You have been called out, in the ways of old.
If people buy enthusiast hardware, they're gonna want to do enthusiast things with them

Adbot
ADBOT LOVES YOU

BlankSystemDaemon
Mar 13, 2009



Zedsdeadbaby posted:

If people buy enthusiast hardware, they're gonna want to do enthusiast things with them
Absolutely, but this isn't a thing that's not going to be fixed.

It might take a little while, that's all.

  • 1
  • 2
  • 3
  • 4
  • 5
  • Post
  • Reply