I have a 2080 ti and it fails on PrimeGrid GFN18, GFN19, ..... It is not like the Einstein abort, but the system screen goes blank and the system reboots. GFN17 and below work fine.
I am running pps-sieve's on it but just went thru the gfn 21 challenge without a problem for either. So far only einstien has failed for me with this latest driver, however I have not tried all primegrid variants yet.
I have a 2080 ti and it fails on PrimeGrid GFN18, GFN19, ..... It is not like the Einstein abort, but the system screen goes blank and the system reboots. GFN17 and below work fine.
Were you able to run the larger GFN WU?
2080 non ti runs everything at primegrid. manual sieves, pps sieves, all the GFN ranges, and AP search. have had no errors.
417.71 Einstein does not work with this new driver as I just had four computational errors and aborted all tasks in this this test under a 2070 card. Bummer, at least for me. I am in a primegrid cpu challenge right now and cannot test this driver further for several more days.
417.71 Einstein does not work with this new driver as I just had four computational errors and aborted all tasks in this this test under a 2070 card. Bummer, at least for me. I am in a primegrid cpu challenge right now and cannot test this driver further for several more days.
I tried this driver with my 2080TI as well and it immediately crashed the display driver same as before. I believe there is something wrong with the Einstein program as all SETI tasks, and all games, etc. work fine.
As soon as I enabled the Einstein project it generated 2 driver timeout errors in my Windows Event Log and the screen froze for several seconds. As soon as the screen started working again I suspended Einstein and I guess it will stay that way until there is a new program for the Einstein@home tasks.
I got a copy of the v1.16 Einstein GPU binary from Oliver and tried it out. I thought the problem might be a latent bug in the Einstein code and introduced with major optimization changes. Nope. It still fails with the opencl -36 error.
I do now have another stand alone test case for Nvidia OR an easy way to test individual binaries on my Windows machine.
The fastest way to get this fixed? 8-) Donate an Nvidia Series-20 board to the Einstein developers.
gandolph1 wrote:
williamd007 wrote:
417.71 Einstein does not work with this new driver as I just had four computational errors and aborted all tasks in this this test under a 2070 card. Bummer, at least for me. I am in a primegrid cpu challenge right now and cannot test this driver further for several more days.
I tried this driver with my 2080TI as well and it immediately crashed the display driver same as before. I believe there is something wrong with the Einstein program as all SETI tasks, and all games, etc. work fine.
As soon as I enabled the Einstein project it generated 2 driver timeout errors in my Windows Event Log and the screen froze for several seconds. As soon as the screen started working again I suspended Einstein and I guess it will stay that way until there is a new program for the Einstein@home tasks.
The fastest way to get this fixed? 8-) Donate an Nvidia Series-20 board to the Einstein developers
I offered that months ago. Are you serious or just speculating? To be specific, I have a 2080 card sitting on the floor I am ready to send to an appropriate interested Einstein recipient. If there is interest I hope to learn of it here or by PM.
Although I am not at all sure that this is an Einstein bug as distinct from an Nvidia hardware or software problem, I think the best chance of getting Nvidia attention would be from a developer report. I'm pretty sure developer reports rate far higher in Nvidia prioritization than do end user feedback. In any case, I submitted feedback including the portable test case I posted at:
Half serious. I have one board and I am using it ... for everything except Einstein. I chatted with Nvidia and asked about the bug status. I added my voice to the urgency. Thanks for sharing the number. They are going to "get back to me". Just like they told you.
archae86 wrote:
rjs5 wrote:
The fastest way to get this fixed? 8-) Donate an Nvidia Series-20 board to the Einstein developers
I offered that months ago. Are you serious or just speculating? To be specific, I have a 2080 card sitting on the floor I am ready to send to an appropriate interested Einstein recipient. If there is interest I hope to learn of it here or by PM.
Although I am not at all sure that this is an Einstein bug as distinct from an Nvidia hardware or software problem, I think the best chance of getting Nvidia attention would be from a developer report. I'm pretty sure developer reports rate far higher in Nvidia prioritization than do end user feedback. In any case, I submitted feedback including the portable test case I posted at:
I have a 2080 ti and it fails
)
I have a 2080 ti and it fails on PrimeGrid GFN18, GFN19, ..... It is not like the Einstein abort, but the system screen goes blank and the system reboots. GFN17 and below work fine.
Were you able to run the larger GFN WU?
I am running pps-sieve's on
)
I am running pps-sieve's on it but just went thru the gfn 21 challenge without a problem for either. So far only einstien has failed for me with this latest driver, however I have not tried all primegrid variants yet.
rjs5 wrote:I have a 2080 ti
)
2080 non ti runs everything at primegrid. manual sieves, pps sieves, all the GFN ranges, and AP search. have had no errors.
417.71 Einstein does not work
)
417.71 Einstein does not work with this new driver as I just had four computational errors and aborted all tasks in this this test under a 2070 card. Bummer, at least for me. I am in a primegrid cpu challenge right now and cannot test this driver further for several more days.
williamd007 wrote:417.71
)
I tried this driver with my 2080TI as well and it immediately crashed the display driver same as before. I believe there is something wrong with the Einstein program as all SETI tasks, and all games, etc. work fine.
As soon as I enabled the Einstein project it generated 2 driver timeout errors in my Windows Event Log and the screen froze for several seconds. As soon as the screen started working again I suspended Einstein and I guess it will stay that way until there is a new program for the Einstein@home tasks.
I got a copy of the v1.16
)
I got a copy of the v1.16 Einstein GPU binary from Oliver and tried it out. I thought the problem might be a latent bug in the Einstein code and introduced with major optimization changes. Nope. It still fails with the opencl -36 error.
I do now have another stand alone test case for Nvidia OR an easy way to test individual binaries on my Windows machine.
The fastest way to get this fixed? 8-) Donate an Nvidia Series-20 board to the Einstein developers.
rjs5 wrote:The fastest way to
)
I offered that months ago. Are you serious or just speculating? To be specific, I have a 2080 card sitting on the floor I am ready to send to an appropriate interested Einstein recipient. If there is interest I hope to learn of it here or by PM.
Although I am not at all sure that this is an Einstein bug as distinct from an Nvidia hardware or software problem, I think the best chance of getting Nvidia attention would be from a developer report. I'm pretty sure developer reports rate far higher in Nvidia prioritization than do end user feedback. In any case, I submitted feedback including the portable test case I posted at:
http://pastoll.info/BOINC/
months ago, and it was acknowledged and assigned a bug number 2434391 about October 30, 2018.
Half serious. I have one
)
Half serious. I have one board and I am using it ... for everything except Einstein. I chatted with Nvidia and asked about the bug status. I added my voice to the urgency. Thanks for sharing the number. They are going to "get back to me". Just like they told you.