Posts by Henk Haneveld

1) Message boards : Number crunching : calculation error (Message 988)
Posted 17 Jan 2025 by Profile Henk Haneveld
Post:
Hi everyone,

I'm currently applying a new version of the loneseed application (1.07) with the following changes/improvements:

1. Smaller batch sizes when we detect a GPU on the Maxwell architecture
(Resolves the 500 or so instances where older 900 and 700 series GPUs cannot complete the work since the kernel runs for too long and gets killed due to timeout)
2. More robust error handling for cuda functions prior to work beginning
(Should help us diagnose the remaining segmentation faults we're getting from some users)

Please let us know if you encounter any issues using this new version.


Thank you!

No, still not working on GTX750Ti with error
GPUassert: the launch timed out and was terminated (code 702) main.cu 4885

But the problem with a black screen is gone.
2) Message boards : Number crunching : calculation error (Message 970)
Posted 15 Jan 2025 by Profile Henk Haneveld
Post:
Hi everyone,

I'm going to push a 1.06 version which reverts a change we made to the maxrregcount argument Nvidia's compiler uses.
It sped up certain cards (Titan X Maxwell, RTX 3090, V100, GTX 1060 tested) but RTX 4090, RTX 2080Ti, and seemingly GTX 750Ti have been negatively impacted.

For the error on GTX 750Ti, it's just a timeout error as there's a limit to how long you can run a single kernel wave. Since we changed the max register count, it's possible this impacted the weaker GTX 750Ti enough to fail this timeout.

Hopefully this addresses both issues.

Thanks!

Thanks for the attempt to fix the problem on GTX750Ti but it still gives a error but a different one.

GPUassert: unknown error (code 999) main.cu 4870

Also my screen went completely black for a few secondes. It starts to look like that the demands of the app are to high for my GPU card.
3) Message boards : Number crunching : calculation error (Message 968)
Posted 15 Jan 2025 by Profile Henk Haneveld
Post:
Hello,

We've pushed another update (1.05) that may address the speed change.

Please let me know if this helps.

Thanks!

Before you start working at speed optimization may I suggest that you first fix the errors in the app.

I get a Computation error with message:
GPUassert: the launch timed out and was terminated (code 702) main.cu 4862

This is on Windows 10, Boinc 8.02 with GPU card
NVIDIA GeForce GTX 750 Ti (driver version 566.36, CUDA version 12.7, compute capability 5.0, 2048MB, 2048MB available, 1388 GFLOPS peak)

Thanks
4) Message boards : Number crunching : Pausing tasks causes tast restart (Message 106)
Posted 29 Jun 2020 by Profile Henk Haneveld
Post:
It is not about checkpointing. When a task is paused, I so far assumed it is just halted and kept in memory. The experience I described above, however, suggests that Minecraft discards the entire data after a certain period of time? Which I found baffling. ;-)

Michael.

When a GPU task is paused it is removed from the GPU memory. Without checkpointing it has to restart from the beginning.
The "Leave in Memory" setting works only for CPU tasks.
5) Message boards : News : We are live (Message 104)
Posted 29 Jun 2020 by Profile Henk Haneveld
Post:
Please remove the useless once a minute scheduler call from your settings.

If there is a real need for the client to call home make the interval at most once per hour.


Removed, this was dangling from testing, apologies

It is back again.
6) Message boards : News : We are live (Message 50)
Posted 27 Jun 2020 by Profile Henk Haneveld
Post:
Please remove the useless once a minute scheduler call from your settings.

If there is a real need for the client to call home make the interval at most once per hour.