I am working on a new project to develop a set of utilities to manage AMD GPUs in the Linux environment. I have a working basic monitor. I plan to include a simple tool to manage performance also. Here is the repository on GitHub:
https://github.com/Ricks-Lab/amdgpu-utils
Let me know of any feedback.

---

GitHub: Ricks-Lab
Instagram: ricks_labs

Very cool Rick. I'm sure the diehard AMD/ATI fans will enjoy the tools.

On another matter, have you been bugged lately with AP tasks failing on your AMD gpu hardware because of "time limit exceeded" errors. There has been a discussion on this matter here.

https://setiathome.berkeley.edu/forum_thread.php?id=83829

See that Dr. Anderson has recognized the bug and has opened an issue for it at github.

https://github.com/BOINC/boinc/issues/2988

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Very cool Rick. I'm sure the diehard AMD/ATI fans will enjoy the tools.

On another matter, have you been bugged lately with AP tasks failing on your AMD gpu hardware because of "time limit exceeded" errors. There has been a discussion on this matter here.

https://setiathome.berkeley.edu/forum_thread.php?id=83829

See that Dr. Anderson has recognized the bug and has opened an issue for it at github.

https://github.com/BOINC/boinc/issues/2988

I stopped running AP tasks long ago since it requires a very old version of AMD drivers to get valid results.

I have been testing with no problems on my RX Vega64 system, but for my ProDuo system, the tool takes a very long time to read power. Let me know if you have any issues on other AMD GPUs. Thanks!

Can't help you there friend as I don't have any AMD gpus. But I'm sure to pass on your post to those I've tried to help with AMD gpu hardware.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Very cool Rick. I'm sure the diehard AMD/ATI fans will enjoy the tools.

On another matter, have you been bugged lately with AP tasks failing on your AMD gpu hardware because of "time limit exceeded" errors. There has been a discussion on this matter here.

https://setiathome.berkeley.edu/forum_thread.php?id=83829

See that Dr. Anderson has recognized the bug and has opened an issue for it at github.

https://github.com/BOINC/boinc/issues/2988

I stopped running AP tasks long ago since it requires a very old version of AMD drivers to get valid results.

I was almost sure its driver related.

---

With each crime and every kindness we birth our future.

I have an R7-240 that I haven't bothered to load drivers for - if they even exist.

---

I have an R7-240 that I haven't bothered to load drivers for - if they even exist.

I was running an AMD FX system with a r7 240 alongside a HD6950, it processed everything without errors if I recall correctly. And there should be drivers for the R7 on AMD's site (for Windoze, if not Linux).
Only stopped 'coz of the power costs rising significantly (am a pensioner, must watch the pennies ;-) )

P.

---

Very cool Rick. I'm sure the diehard AMD/ATI fans will enjoy the tools.

On another matter, have you been bugged lately with AP tasks failing on your AMD gpu hardware because of "time limit exceeded" errors. There has been a discussion on this matter here.

https://setiathome.berkeley.edu/forum_thread.php?id=83829

See that Dr. Anderson has recognized the bug and has opened an issue for it at github.

https://github.com/BOINC/boinc/issues/2988

I stopped running AP tasks long ago since it requires a very old version of AMD drivers to get valid results.

I was almost sure its driver related.

Well they had to change the code already for the Turing cards and that was not because of any driver change. It has to do with the architecture of the gpu and how it gets interpreted and exported by the CUDA and OpenCL API libraries. The BOINC code polls a card's capabilities from the manufacturer API library. The very same thing could have happened with the ATI/AMD cards.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

I am working on a new project to develop a set of utilities to manage AMD GPUs in the Linux environment. I have a working basic monitor. I plan to include a simple tool to manage performance also. Here is the repository on GitHub:
https://github.com/Ricks-Lab/amdgpu-utils
Let me know of any feedback.

I have the same problem as many others. I have a 2400G (apu type) but I have been running it under Windows because I didn't know that there was going to be much improvement under Linux.

Now I have to decide :)

Tom

---

A proud member of the OFA (Old Farts Association).

See that Dr. Anderson has recognized the bug and has opened an issue for it at github.

https://github.com/BOINC/boinc/issues/2988

Well they had to change the code already for the Turing cards and that was not because of any driver change. It has to do with the architecture of the gpu and how it gets interpreted and exported by the CUDA and OpenCL API libraries. The BOINC code polls a card's capabilities from the manufacturer API library. The very same thing could have happened with the ATI/AMD cards.

I have (just) enough knowledge of the NVidia architecture to write that Turing patch for them, with help and checks by users here. The patch was accepted without quibble.

If anyone here has sufficient knowledge of ATI cards to supply the fix, I can submit it on your behalf. I've put a link to the current code area in #2988, so you can see what's necessary.

I wonder if asking some of the Windows monitoring app developers like Ray Hinchcliffe of SIV or Martin Malik of HwInfo64 for hints about how they poll for a ATI/AMD devices capabilities in their programs.

Seems the developer of GPU-Z would be the perfect person to ask but I don't know who that person is. Might be an effort by the team at TechPowerUp since they host the forums and downloads for the program.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Good idea. Ray was one of those who helped me with NV.

I wonder if asking some of the Windows monitoring app developers like Ray Hinchcliffe of SIV or Martin Malik of HwInfo64 for hints about how they poll for a ATI/AMD devices capabilities in their programs.

Seems the developer of GPU-Z would be the perfect person to ask but I don't know who that person is. Might be an effort by the team at TechPowerUp since they host the forums and downloads for the program.

My starting point was 2 repositories on GitHub that I originally found through a Reddit posting:
https://github.com/sibradzic/amdgpu-clocks
https://github.com/BoukeHaarsma23/WattmanGTK
There were also a few articles on Phoronix that were helpful, but I didn't save the links. I think the issue I have with the ProDuo is specific to that card. I also had GPU-z issues with that setup. I have just updated a new version which is a major rewrite to make it easier to update data to a Gtk display. But I am new to Gtk, so the GUI is very simple. I am digging into Gtk, so I can use it in the the utility I am planning to build to modify p-state details.

---

GitHub: Ricks-Lab
Instagram: ricks_labs

I am working on a new project to develop a set of utilities to manage AMD GPUs in the Linux environment. I have a working basic monitor. I plan to include a simple tool to manage performance also. Here is the repository on GitHub:
https://github.com/Ricks-Lab/amdgpu-utils
Let me know of any feedback.

I have the same problem as many others. I have a 2400G (apu type) but I have been running it under Windows because I didn't know that there was going to be much improvement under Linux.

Now I have to decide :)

Tom

Let me know if you try it. It would be good to know if it works on an APU. Not sure how different the interaction mechanism would be, vs a card in a PCIe slot.

That WattmanGTK repository gives some hints of where you are heading with GTK interface I suspect.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

That WattmanGTK repository gives some hints of where you are heading with GTK interface I suspect.

I learned a lot from that repository, but will take a different direction. I think Wattman is great for a single card, but a bit tedious for a multi card setup. Plus combining edit and monitor in a single app adds too much complexity. My plans is for a simple view to give the status of all cards at once. The latest release has a initial version of this. I also plan a simple interface to view and edit all p-states of all GPUs in a single view.

Here is a screen shot of the simple Gtk view of status:

Maybe its a bit too simple. Still a lot to learn about Gtk...

What does the ps in the monitor stand for. Guessing ---- power state?

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

What does the ps in the monitor stand for. Guessing ---- power state?

Yes, thats correct.

---

GitHub: Ricks-Lab
Instagram: ricks_labs