Join PrimeGrid
Returning Participants
Community
Leader Boards
Results
Other
drummers-lowrise
|
Message boards :
Generalized Fermat Prime Search :
Genefer 17 Mega v4.02 (OCLcudaGFN17MEGA) errors
| Author |
Message |
|---|
|
|
|
Mint 20.3 with 5.15 kernel on a dual card box:
Any pointers. All WUs of this type are not failing but probably about 1/3 of them on a particular RTX 3070 card.
Stderr output
<core_client_version>7.20.5</core_client_version>
< compositeVolunteer tester
Send message
Joined: 16 Feb 10
Posts: 1172
ID: 55391
Credit: 1,208,228,947
RAC: 1,105,945
                        
|
|
Mr. Google says the error might be related to a memory address misalignment using a shared buffer.
This would be a software issue if the problem happens at all, so it would likely happen every time.
Hence, it's more likely to be a hardware issue or driver issue.
For a hardware issue, a misaligned address is a symptom of a memory bus address line
having insufficient time to latch the address before the data strobe, either because the
clock is too fast, or excessive loading of the bus is causing the voltage to slew too slowly.
Which leads to a couple of questions about the hardware.
1) Is the card with the failing workunits overclocked?
2) Does the card with failing workunits work properly when the other card is removed from the box? | |
|
|
|
|
1) Yes, looking at the OC now. I now strongly suspect this is the issue.
2) I haven't removed the 3060ti yet. But I've had bus/bandwidth issues when I had a pair of RX 6600s in it and it had different symptoms.
Will report back. Thanx, Skip
____________
- da shu @ HeliOS,
"A child's exposure to technology should never be predicated on an ability to afford it." | |
|
|
|
1) Yes, looking at the OC now. I now strongly suspect this is the issue.
2) I haven't removed the 3060ti yet. But I've had bus/bandwidth issues when I had a pair of RX 6600s in it and it had different symptoms.
Will report back. Thanx, Skip
Well I need to run thru some more Gen17 Mega WUs but so far so good. I have a twin to this card in another box so I checked it. No errors and sure enough it has a bit lower clocks. I've now lowered both sclk & mclk on the problem child.
Prior to this 31 of 164 got the error previously posted. Since dropping the clocks no errors so far.
Thanx, Skip
____________
- da shu @ HeliOS,
"A child's exposure to technology should never be predicated on an ability to afford it." | |
|
Post to thread
Message boards :
Generalized Fermat Prime Search :
Genefer 17 Mega v4.02 (OCLcudaGFN17MEGA) errors |
