FYI, we're going to look into this problem as soon as we possibly can.
Thank you very much!
I'm sure there are quite a few people who have invested in these cards who will be rather relieved to know that a solution might be on the horizon :-).
I'm ready to uncomment or remove my gpu exclusion for my 2080 as soon as the project supports Turing. Looking forward to seeing what the card is capable of crunching.
The thing is we always need get hold of the hardware affected before we can start proper debugging of course. Luckily we now have a RTX so we hope to get started this or next week :-)
You have all the failing Turing hardware at your access. The Einstein PROJECT -> PREFERENCES -> BETA SETTINGS allows you to release your test binary under that option and verify that it is fixed. Just release your test version and let the community test it for you. 8-)
Since that is a option, those that don't want to participate can just turn it off, It will also enable you to test if older boards get broken with the changes.
Oliver Behnke wrote:
The thing is we always need get hold of the hardware affected before we can start proper debugging of course. Luckily we now have a RTX so we hope to get started this or next week :-)
That will certainly be our testing approach to validate a potential fix, but it's not well-suited for the actual debugging as that needs much more direct interaction and faster turnaround times.
A new data file LATeah0104Y.dat came into play more than 12 hours ago. It has the same size as, and would appear to be a continuation of a previous series that ended with LATeah0104X.dat (first mentioned in the opening post of this thread). The tasks based on the new file will most likely crunch faster than the previous 2103L tasks.
Based on the fact that 0104X tasks did fail on Turing GPUs, I imagine the new ones would also fail as well, unfortunately.
A new data file LATeah0104Y.dat came into play more than 12 hours ago. It has the same size as, and would appear to be a continuation of a previous series that ended with LATeah0104X.dat (first mentioned in the opening post of this thread). The tasks based on the new file will most likely crunch faster than the previous 2103L tasks.
Based on the fact that 0104X tasks did fail on Turing GPUs, I imagine the new ones would also fail as well, unfortunately.
Guys, the datafiles won't change the Turing issue. We're going to release a new app version if and when that problem got fixed, probably preceded by a beta release. We'll also let you know about this via a post or PM.
Until I'm going to unsubscribe from this thread again to increase the SNR for me
Guys, the datafiles won't change the Turing issue. We're going to release a new app version if and when that problem got fixed, probably preceded by a beta release. We'll also let you know about this via a post or PM.
Until I'm going to unsubscribe from this thread again to increase the SNR for me
Cheers
But before you go, please be aware that some data files have processed successfully on Turing cards with the current app. That's a debug data point that you shouldn't ignore.
Guys, the datafiles won't change the Turing issue.
Then how come Turing GPUs are quite able to crunch tasks based on certain series of data files that have particular characteristics but not on others that have different characteristics? This has been a repeated observation that has gone on for months. The list of people running into this problem continues to grow.
To characterize this as just 'noise' is rather demeaning to those who have spent a lot of cash buying the latest hardware with the intention of supporting this project.
Oliver Behnke wrote:FYI,
)
Thank you very much!
I'm sure there are quite a few people who have invested in these cards who will be rather relieved to know that a solution might be on the horizon :-).
Cheers,
Gary.
I'm ready to uncomment or
)
I'm ready to uncomment or remove my gpu exclusion for my 2080 as soon as the project supports Turing. Looking forward to seeing what the card is capable of crunching.
The thing is we always need
)
The thing is we always need get hold of the hardware affected before we can start proper debugging of course. Luckily we now have a RTX so we hope to get started this or next week :-)
Oliver
Einstein@Home Project
You have all the failing
)
You have all the failing Turing hardware at your access. The Einstein PROJECT -> PREFERENCES -> BETA SETTINGS allows you to release your test binary under that option and verify that it is fixed. Just release your test version and let the community test it for you. 8-)
Since that is a option, those that don't want to participate can just turn it off, It will also enable you to test if older boards get broken with the changes.
That will certainly be our
)
That will certainly be our testing approach to validate a potential fix, but it's not well-suited for the actual debugging as that needs much more direct interaction and faster turnaround times.
Cheers
Einstein@Home Project
A new data file
)
A new data file LATeah0104Y.dat came into play more than 12 hours ago. It has the same size as, and would appear to be a continuation of a previous series that ended with LATeah0104X.dat (first mentioned in the opening post of this thread). The tasks based on the new file will most likely crunch faster than the previous 2103L tasks.
Based on the fact that 0104X tasks did fail on Turing GPUs, I imagine the new ones would also fail as well, unfortunately.
Cheers,
Gary.
Still fails on my 2080
)
Still fails on my 2080 ti
Guys, the datafiles won't
)
Guys, the datafiles won't change the Turing issue. We're going to release a new app version if and when that problem got fixed, probably preceded by a beta release. We'll also let you know about this via a post or PM.
Until I'm going to unsubscribe from this thread again to increase the SNR for me
Cheers
Einstein@Home Project
Oliver Behnke wrote:Guys, the
)
But before you go, please be aware that some data files have processed successfully on Turing cards with the current app. That's a debug data point that you shouldn't ignore.
Oliver Behnke wrote:Guys, the
)
Then how come Turing GPUs are quite able to crunch tasks based on certain series of data files that have particular characteristics but not on others that have different characteristics? This has been a repeated observation that has gone on for months. The list of people running into this problem continues to grow.
To characterize this as just 'noise' is rather demeaning to those who have spent a lot of cash buying the latest hardware with the intention of supporting this project.
Cheers,
Gary.