xxx9999999 workframes ?
I just had my two gpu3 rig's essentially error out .....they went to xxx.99999 workframes then died , sent my 580's into full fan mode and everything.
Anyone see this ?
No one eh ?
It hasnt happened again , it was just odd both rigs did it at the same time.
Probably bad WUs pulled from the same server.
Or a simple hick-up from your PSU.
Ok , I have caught the machine in the act...
Here's what you first see ;
The last time I saw this , the rig just died....Now , it actually keeps going , and will return positive results to stanford.
However , if you stop the client , you will get this :
Then If you restart , you get this :
At this point you need to delete the std files and start from scratch.
Seems to me it's a flakey unit but .....
Edit : As far as I know "Client Died" is not a positive outcome....
I don't even remember reading about that. Ever.
The fahwiki says this could be due to bad memory/or timings issues:
Common Error Messages - FaHWiki
Tried memtest? I know its a GPU unit but give your rig's memory a check with 1-2 passes.
I havnt tried memtest but I will.....this is a 'formerly' stable rig ...no OC etc....
Yep....that's what makes me think workunit...
Mebbie stanford is crunching 'Hulk' genes....who knows....it's odd they return correctly.
:shok: Oh my god, they killed Client!
|All times are GMT -7. The time now is 10:58 PM.|