New The PC Build Thread

For AM5 users who have updated to AGESA 1.2.0.3a and were having issues with the chipset driver provided by their motherboard vendor. AMD has come around and published v7 on the 25th. It is a newer version (7.02.13.148) than that released by the motherboard vendors. Some components in the package have been updated.

I installed it yesterday, and all my issues went away. I knew there was something up with Gigabyte's driver package. I don't even have the temp spikes anymore. Gigabyte hasn't updated it on its website.

I guess AMD waited and prepared this to be AI 300 series ready.

Here are the release notes:


Article Number: RN-RYZEN-CHIPSET-7-02-13-148

Release Highlights​

  • New program support added.
  • Bug fixes.

Known Issues​

  • After installing the AMD Chipset Installer version “7.xx.xx.xx”, users cannot install version “6.xx.xx.xx” or earlier.
    Workaround:
    • Un-install the latest version of the AMD Chipset installer (version 7.xx.xx.xx or later).
    • Delete the “Qt_Dependencies” folder located at "C:\Program Files (x86)\AMD\Chipset_Software\".
    • Install any older version of the AMD Chipset installer (version 6.xx.xx.xx or earlier).
  • On a non-English OS, some driver names may appear in English. 
  • Occasionally, Ryzen PPKG may not install or upgrade.

That known issue described, the workaround... Find that out on my own when I downgraded my BIOS.

Someone on Reddit with the same issues as I had said that you needed to extract the Gigabyte package and then install it like you did in the old days. Ahem, that didn't go as expected. Moer.

All sorted now. I don't game on this system, but canned benchmarks all have slightly lower scores. I assume the improvements are in the interconnect latencies. That I won't be able to tell.

Anyhow, Happy system, happy work.
 
If you have any 7000 or 9000 AMD CPU, more particularly so the X3D and have AGESA 1.2.0.2a/b or 1.2.0.3a installed. Check your voltages, so far ASRock, ASUS and MSI boards have shown incredibly high voltage spikes.

It is a limited case, but as far as the internet sleuths can tell this seems to be related to some board vendors having an unlocked SoC voltage that is not within the safe perimeters. This is very similar to the Intel case...

If you have a hardware monitor, check for this,


This might not be the cause of the issue, but have been observed on some ASRock, ASUS and MSI boards.

Some suggest to manually set your SoC voltage.

This can be caused by AGESA of how it is handling the CPU. I won't lay it on the motherboard vendors, yet.

EDIT: Some are also say that it is a reporting issue, and nothing is wrong with the behaviour. I won't know.
on series 7 9700x with msi mb ive set my soc manual to 1.2v via ryzen master with expo enabled(6000mhz) pbo disabled. temps at 65-68 running all cores at boost speed(5500mhz) 24/7
very impressive chipset, draw is 100watt according to ryzen master. know this chip capable of OC and have tried with pbo manual overrides including curve control but didnt get more than a 100mhz from it running full bench load consistently. voltage spike with associated heat not worth it imho maybe im doing it wrong but for now im a happy chappie!
 
Since when has RebelTech been in Randburg? You mean I could go drive and collect all this time?
 
Last edited:
noticed some instability today think due to undervolt so manually increased cpu to 1.25v disabled expo/PBO and enabled xmp mode(same/similar timings) left cpu at 5500mhz temps have jumped to 75(day temps) still cool for full workload incl all threads. will keep updated.

***update*** set expo mode on/pbo off manual change to 5500mhz with volt at 1,25 no more restarts or hanging/ladies and gents can confirm it was indeed the undervolt previously applied that caused issues!!
 
Last edited:
For AM5 users who have updated to AGESA 1.2.0.3a and were having issues with the chipset driver provided by their motherboard vendor. AMD has come around and published v7 on the 25th. It is a newer version (7.02.13.148) than that released by the motherboard vendors. Some components in the package have been updated.

I installed it yesterday, and all my issues went away. I knew there was something up with Gigabyte's driver package. I don't even have the temp spikes anymore. Gigabyte hasn't updated it on its website.

I guess AMD waited and prepared this to be AI 300 series ready.

Here are the release notes:




That known issue described, the workaround... Find that out on my own when I downgraded my BIOS.

Someone on Reddit with the same issues as I had said that you needed to extract the Gigabyte package and then install it like you did in the old days. Ahem, that didn't go as expected. Moer.

All sorted now. I don't game on this system, but canned benchmarks all have slightly lower scores. I assume the improvements are in the interconnect latencies. That I won't be able to tell.

Anyhow, Happy system, happy work.

AMD always did have endless problems. This is why I stick to Intel.

Not a single issue with my 12th/14th gens or Z790 boards.
 
I wonder if all of these AI companies are the reason for the PSU drought.
 
AMD always did have endless problems. This is why I stick to Intel.

Not a single issue with my 12th/14th gens or Z790 boards.
Troll comment much?

No manufacturer is perfect, but have you been living under a rock with regards to what's been going on with 13th/14th gen intel desktop CPUs?
 
She's up and running!
3a52df0651bb7fde38570d955827c0e2.jpg
 
Nvidia aren't doing themselves any favour's with this generation.

Lacklustre generational improvement and missing rops.
It won't necessarily make a dent in the company itself since they're all in on the AI train but they might lose some of their shine as a gaming card company.

I won't be surprised if the issues are arising from some of the production being moved to the new fabs in the USA. They still do the packaging in Taiwan. It is most certainly a sampling issue, possibly as a result of indifferent standards.

It is bad, but it is also a teething issue. Jensen will be happy that the issue was identified early on, as it could have been much worse.

If it is more widespread, they should announce a recall.

EDIT: There is a new fab in Mexico too.
 
Where there's smoke, there's fire!


Retailer claims Nvidia is recalling ‘large number’ of RTX 5090 graphics cards due to fire hazard​

There have already been a few RTX 5090 graphics cards sold only to burn out due to the amount of power going through the 12V-2×6 power connector. When this happened with the first RTX 4090 graphics cards two years ago, the issue was primarily pinned on user error due to the power cable not being pushed all the way in. However, this time around, it may not have anything to do with the user at all, as we have received reports of a recall in direct relation to a potential fire hazard.
 
Nvidia aren't doing themselves any favour's with this generation.

Lacklustre generational improvement and missing rops.
It won't necessarily make a dent in the company itself since they're all in on the AI train but they might lose some of their shine as a gaming card company.

The transition has been lightning quick...data center revenue has grown like a 1000% in the last couple years while the gaming side has stagnated. It's not hard to see why they'd rather channel their output to the professional space where they get more per mm wafer space than in the consumer market.

Screenshot 2025-03-06 204724.jpg

Say what you like about the BS Jensen spills but he knows how to make a dollar.
 
Top
Sign up to the MyBroadband newsletter
X