Topic: weird 100% usage problem in my rig (Read 1338 times)
Dunno what part, it just wouldn't boot properly, stopping at GPU initialisation. Cards would sometimes power on and often wouldn't. Taking them all out and putting a basic card in it still wouldn't reliably start up.
hmm, the board still boots up fine, had it offline for some time too.
However, I cant really check on the board, since the only GPU on it that still delivers a display signal is too far off to reach the peg slot that the board is pushing its display information to. However, did you remember what part on the board failed?
also, does anyone know how to flatten the mainboard again(the board is screwed on the rig, and, well, it has taken a kind of wavy shape over the last weeks...
However, I cant really check on the board, since the only GPU on it that still delivers a display signal is too far off to reach the peg slot that the board is pushing its display information to. However, did you remember what part on the board failed?
also, does anyone know how to flatten the mainboard again(the board is screwed on the rig, and, well, it has taken a kind of wavy shape over the last weeks...
I had this sort of weird "almost working" behaviour when my motherboard died. The workshop only believed me when they couldn't get it to boot.
Hi there,
I'm currently having some kind of weird issue in my rig. The rig itself consists of some cheap asus board, together with a 5850, 6850, 6950 and a 6770.
they all used to work just fine for the last months.
So the other night, I noticed on my mining tool, that my 6770 went offline, while all other cards continued to work. I thought, fine, I'll reboot the rig and it'll be all fine again.
after rebooting, all cards were still detected, X11 was starting up and not freezing. however when I started the second worker, I noticed that the system became laggy, so I tried to htop and found that all poclbm instances were using an entire core. So I did another reboot, noticed the problem still being there, used another pocl(newer version) - problem still there. I went to hashkill-gpu, which refused to start with no reason. fine again. I then tried CGMiner which told me on start that it couldnt fire the 5850, then started all other GPUs and also had the 100% issue. I thought, ok, maybe I raped the 5850 too hard over the past. but it wasnt that. after rebooting with the 5850 detached, I found the other gpus still using 100% each. So I reattached it, and tried mining anyways. despite all gpus coming online and having OK stale rates, my mining power was only about 50% of what it should be. after killing two pocl instances, the other 2 instances went to full capacity, however using both of my cores to the limit.
then I went installing a new linux stick. after everything was set up, I still had that same error.
finally, I tried each GPU single, and different combinations of them... all tries lead to each pocl instance having 100% CPU usage each.
So I'm out of elaborate guesses now, anyone else have any hints what's wrong with the rig?
I'm currently having some kind of weird issue in my rig. The rig itself consists of some cheap asus board, together with a 5850, 6850, 6950 and a 6770.
they all used to work just fine for the last months.
So the other night, I noticed on my mining tool, that my 6770 went offline, while all other cards continued to work. I thought, fine, I'll reboot the rig and it'll be all fine again.
after rebooting, all cards were still detected, X11 was starting up and not freezing. however when I started the second worker, I noticed that the system became laggy, so I tried to htop and found that all poclbm instances were using an entire core. So I did another reboot, noticed the problem still being there, used another pocl(newer version) - problem still there. I went to hashkill-gpu, which refused to start with no reason. fine again. I then tried CGMiner which told me on start that it couldnt fire the 5850, then started all other GPUs and also had the 100% issue. I thought, ok, maybe I raped the 5850 too hard over the past. but it wasnt that. after rebooting with the 5850 detached, I found the other gpus still using 100% each. So I reattached it, and tried mining anyways. despite all gpus coming online and having OK stale rates, my mining power was only about 50% of what it should be. after killing two pocl instances, the other 2 instances went to full capacity, however using both of my cores to the limit.
then I went installing a new linux stick. after everything was set up, I still had that same error.
finally, I tried each GPU single, and different combinations of them... all tries lead to each pocl instance having 100% CPU usage each.
So I'm out of elaborate guesses now, anyone else have any hints what's wrong with the rig?
Jump to: