Hey everybody. I'm new to this, so I apologize if I've missed something basic here
I just set up my first "rig," which is really a test using a tower I've had for a few years. I picked up a couple of these locally and got them set up. Wasn't too hard to get them going but they're only doing about 21 MH/s each and I can't figure out why. I was expecting them to be above 25 MH/s, and closer to 30.
I'm running ubuntu 16.04.2 with kernel 4.8.0-53. amdgpu-pro 17.10 is installed and working as far as I can tell.
Poking around in the driver files in /sys/class/card0/device, I find:
When I try to modify the pp_mclk_od and pp_sclk_od, I get these messages in journalctl and no change in performance:
After this, if I look at the pp_dpm_mclk and sclk files again, this is what I see:
Although, given that there's no change in performance, I think the driver has lost its mind. The only way to recover the driver from here seems to be a reboot.
Has anybody seen this specific problem before?
What do I need to do to get these cards up to expected speeds?
I just set up my first "rig," which is really a test using a tower I've had for a few years. I picked up a couple of these locally and got them set up. Wasn't too hard to get them going but they're only doing about 21 MH/s each and I can't figure out why. I was expecting them to be above 25 MH/s, and closer to 30.
I'm running ubuntu 16.04.2 with kernel 4.8.0-53. amdgpu-pro 17.10 is installed and working as far as I can tell.
Poking around in the driver files in /sys/class/card0/device, I find:
root@newton:/sys/class/drm/card0/device# cat pp_dpm_mclk
0: 300Mhz
1: 1000Mhz
2: 2000Mhz *
root@newton:/sys/class/drm/card0/device# cat pp_dpm_sclk
0: 300Mhz
1: 600Mhz
2: 900Mhz
3: 1145Mhz
4: 1215Mhz
5: 1300Mhz
6: 1355Mhz
7: 1405Mhz *
When I try to modify the pp_mclk_od and pp_sclk_od, I get these messages in journalctl and no change in performance:
May 29 09:27:12 newton kernel: amdgpu: [powerplay] VDDCI is larger than max VDDCI in VDDCI Voltage Table!
May 29 09:27:12 newton kernel: amdgpu: [powerplay]
failed to send message 18a ret is 0
May 29 09:27:13 newton kernel: amdgpu: [powerplay]
failed to send pre message 145 ret is 0
May 29 09:27:13 newton kernel: amdgpu: [powerplay]
failed to send pre message 146 ret is 0
May 29 09:27:14 newton kernel: amdgpu: [powerplay]
failed to send message 146 ret is 0
After this, if I look at the pp_dpm_mclk and sclk files again, this is what I see:
root@newton:/sys/class/drm/card0/device# cat pp_dpm_mclk
0: 300Mhz *
1: 1000Mhz
2: 2000Mhz
root@newton:/sys/class/drm/card0/device# cat pp_dpm_sclk
0: 300Mhz *
1: 600Mhz
2: 900Mhz
3: 1145Mhz
4: 1215Mhz
5: 1300Mhz
6: 1355Mhz
7: 1419Mhz
root@newton:/sys/class/drm/card0/device#
Although, given that there's no change in performance, I think the driver has lost its mind. The only way to recover the driver from here seems to be a reboot.
Has anybody seen this specific problem before?
What do I need to do to get these cards up to expected speeds?