If the size is 0, probably any allocation of that resource will fail. Is the size reported as '0' due to driver or application bug on Windows?
The beta app was released on 28 Aug 2021 but it seems the issue is not being debugged for AMD windows hosts since then: allowing beta apps only makes sense if the developers actively debug the issue and it seems very repeatable since it 100% fails on windows with AMD hardware? Seems there should be enough data points to either fix it or stop scheduling the version for AMD windows hosts.
Links can be copied and pasted: are you saying that doesn't work? I haven't posted anything other than plain text, not sure what you want me to do about anything auto-generated (or not).
Disable beta has already been suggested.
My point: what is the point of beta testing if the bugs are never resolved? I would have thought the point was to eventually get it working on as much compatible hardware as possible. If no one ever runs these, the status will never change.
Can you explain what the beta testing is for since I seem to have the wrong conception?
I will happily run them as long as they error out quickly and don't cause any stability issues, but only if there is a larger point.
your url that is not a hyperlink (“link”) does work.
your url that IS a link, does not work. Hence “your link doesn’t work”.
I don’t think you should expect any further development in this app. It should probably have been removed but i can’t speak to why it wasn’t, probably because it works for some cards and not others. Your RX560 is a card that doesn’t work. The best option for you is to disable beta since the FGRPB1G project will end in about 4 months and I doubt the devs are doing any active development on it anymore.
My point: what is the point of beta testing if the bugs are never resolved? I would have thought the point was to eventually get it working on as much compatible hardware as possible. If no one ever runs these, the status will never change.
Can you explain what the beta testing is for since I seem to have the wrong conception?
Gary, it's okay if you want to run BETA testing, that's what they are there for, to 'test' the project.
It is wrong to act like your upset that they're not working as you'd expect. It would be better if you could post your findings in the Technical News Forum under the name of the project your testing. You cannot create a thread there, only respond to one created by the Administrators. So look for a thread closely related to what you are experiencing. If there is not a related thread to your posts, then you should be a bit more careful when using this forum; maybe start with "Experiencing problems with beta test ...blah...blah...blah".
Another method would be to send Bernd a private message so that he would be sure to read it. It's a BETA test, it is expected to not be correct until all of the bugs are worked out of it.
My point: what is the point of beta testing if the bugs are never resolved? I would have thought the point was to eventually get it working on as much compatible hardware as possible. If no one ever runs these, the status will never change.
Can you explain what the beta testing is for since I seem to have the wrong conception?
Gary, it's okay if you want to run BETA testing, that's what they are there for, to 'test' the project.
It is wrong to act like your upset that they're not working as you'd expect. It would be better if you could post your findings in the Technical News Forum under the name of the project your testing. You cannot create a thread there, only respond to one created by the Administrators. So look for a thread closely related to what you are experiencing. If there is not a related thread to your posts, then you should be a bit more careful when using this forum; maybe start with "Experiencing problems with beta test ...blah...blah...blah".
Another method would be to send Bernd a private message so that he would be sure to read it. It's a BETA test, it is expected to not be correct until all of the bugs are worked out of it.
Maybe Bernd would even start a thread if none exists so all the Beta Testers can come forward with their problems, if it's an easy fix, like cancelling the existing tasks and reissuing them with better memory values then it could only take a day or so to get things back on track.
I'm not acting like I'm upset, I am just pointing out obvious facts: beta testing...is for testing, which implies fixing. Why else is it being tested?
"It is wrong to act like your upset that they're not working as you'd expect."
I am upset that the beta testing is not being done because that implies using feedback (from our computers running and expending time and energy on it) to fix issues. I do not expect that a beta version is 100% working (you can read that I said exactly this previously). But, it cannot be reproducibly failing since day 1 of rollout 2 years ago and not even be looked at, there is no "testing" going on. And so time and energy are being wasted.
I think this thread has plenty of feedback/details already, I just thought I would ask what exactly is going on with the "beta test".
The tasks are not kept indefinitely (I don't control that, I used the first task in the list and these are the oldest and get dropped first).
A more recent working task on same linux host: https://einsteinathome.org/task/1485536124
Snip from that log:
Using OpenCL platform provided by: Advanced Micro Devices, Inc.
Using OpenCL device "gfx900:xnack-" by: Advanced Micro Devices, Inc.
Max allocation limit: 7287183768
Global mem size: 8573157376
Unless the plan is to not use any more OpenCL apps, then they can solve this problem for future apps.
Seeing same for windows host
)
Seeing same for windows host for 1.28 version: https://einsteinathome.org/task/1484667057
Also reports wrong "Global mem size: 0'
This is non-zero on linux host for 1.28 version: https://einsteinathome.org/task/1479678918
If the size is 0, probably any allocation of that resource will fail. Is the size reported as '0' due to driver or application bug on Windows?
The beta app was released on 28 Aug 2021 but it seems the issue is not being debugged for AMD windows hosts since then: allowing beta apps only makes sense if the developers actively debug the issue and it seems very repeatable since it 100% fails on windows with AMD hardware? Seems there should be enough data points to either fix it or stop scheduling the version for AMD windows hosts.
Your link doesn’t
)
Your link doesn’t work.
but this beta app doesn’t really work with older AMD GPUs pretty much only works with Navi and newer.
disable beta tasks and you won’t get them.
_________________________________________________________________________
Links can be copied and
)
Links can be copied and pasted: are you saying that doesn't work? I haven't posted anything other than plain text, not sure what you want me to do about anything auto-generated (or not).
Disable beta has already been suggested.
My point: what is the point of beta testing if the bugs are never resolved? I would have thought the point was to eventually get it working on as much compatible hardware as possible. If no one ever runs these, the status will never change.
Can you explain what the beta testing is for since I seem to have the wrong conception?
I will happily run them as long as they error out quickly and don't cause any stability issues, but only if there is a larger point.
I don’t think you should
)
your url that is not a hyperlink (“link”) does work.
your url that IS a link, does not work. Hence “your link doesn’t work”.
I don’t think you should expect any further development in this app. It should probably have been removed but i can’t speak to why it wasn’t, probably because it works for some cards and not others. Your RX560 is a card that doesn’t work. The best option for you is to disable beta since the FGRPB1G project will end in about 4 months and I doubt the devs are doing any active development on it anymore.
_________________________________________________________________________
gary wrote: Disable beta has
)
Gary, it's okay if you want to run BETA testing, that's what they are there for, to 'test' the project.
It is wrong to act like your upset that they're not working as you'd expect. It would be better if you could post your findings in the Technical News Forum under the name of the project your testing. You cannot create a thread there, only respond to one created by the Administrators. So look for a thread closely related to what you are experiencing. If there is not a related thread to your posts, then you should be a bit more careful when using this forum; maybe start with "Experiencing problems with beta test ...blah...blah...blah".
Another method would be to send Bernd a private message so that he would be sure to read it. It's a BETA test, it is expected to not be correct until all of the bugs are worked out of it.
Proud member of the Old Farts Association
GWGeorge007 wrote: gary
)
Maybe Bernd would even start a thread if none exists so all the Beta Testers can come forward with their problems, if it's an easy fix, like cancelling the existing tasks and reissuing them with better memory values then it could only take a day or so to get things back on track.
Good idea Mikey! Why don't
)
Good idea Mikey! Why don't you make the suggestion to Bernd since it is your idea.
Proud member of the Old Farts Association
I'm not acting like I'm
)
I'm not acting like I'm upset, I am just pointing out obvious facts: beta testing...is for testing, which implies fixing. Why else is it being tested?
"It is wrong to act like your upset that they're not working as you'd expect."
I am upset that the beta testing is not being done because that implies using feedback (from our computers running and expending time and energy on it) to fix issues. I do not expect that a beta version is 100% working (you can read that I said exactly this previously). But, it cannot be reproducibly failing since day 1 of rollout 2 years ago and not even be looked at, there is no "testing" going on. And so time and energy are being wasted.
I think this thread has plenty of feedback/details already, I just thought I would ask what exactly is going on with the "beta test".
The tasks are not kept
)
The tasks are not kept indefinitely (I don't control that, I used the first task in the list and these are the oldest and get dropped first).
A more recent working task on same linux host: https://einsteinathome.org/task/1485536124
Snip from that log:
Unless the plan is to not use any more OpenCL apps, then they can solve this problem for future apps.
GWGeorge007 wrote: Good idea
)
Done!!