2025-04-02 10:40 PM - edited 2025-04-02 10:41 PM
Greetings!
After updating to the latest v5 of NEAI yesterday, benchmarks have failed to cross the initial step of loading in the first library. I suspect this has to do with the number of cores/ threads and memory being allocated to the NEAI process, as the 3221225773 and 3221225794 pwsh codes are dumped in the logs. For reference, my system has a Ryzen 7 4800H and 16 GB of RAM onboard, which hasn't had an issue with running benchmarks in previous versions.
Updating to the new 5.0.1 version released earlier today hasn't fixed the issues either. I've attached my logs below.
Solved! Go to Solution.
2025-04-03 2:04 AM
Hello @Haze,
Concerning the slider, it was moved to the settings.
You can access the setting page by clicking the first icon on the bottom left corner.
The slider is at the bottom of the setting page:
For the error itself, we are digging further as it does not seem to be the disk space.
I will let you know when I know more.
Sorry for the inconvenience,
Have a good day,
Julian
2025-04-03 1:21 AM
Hello @Haze,
Thank you for the logs.
How much disk space do you have on your pc?
The benchmark allocate space per available CPU core.
In my case, I have around 20 cores used in NanoEdge (check in the option, at the botton) and it allocate around 40GB for the benchmark.
You may not have enough disk space available.
if it seems to be the case, you can reduce the number of cores used in the NanoEdge settings or free up some disk space.
Have a good day,
Julian
2025-04-03 1:55 AM
Hi @Julian E.
I've got over 500 GB free on the same disk NEAI is on, which shouldn't be an issue. However, I don't seem to have the core selection slider during benchmark setup like v4.6. Has this moved to a different location?
Thanks,
Haze
2025-04-03 2:04 AM
Hello @Haze,
Concerning the slider, it was moved to the settings.
You can access the setting page by clicking the first icon on the bottom left corner.
The slider is at the bottom of the setting page:
For the error itself, we are digging further as it does not seem to be the disk space.
I will let you know when I know more.
Sorry for the inconvenience,
Have a good day,
Julian
2025-04-03 2:21 AM
Thanks for the reply, Julian
Dropping the number of threads with the slider seems to have fixed the issue for my system. Looks like NEAI was indeed trying to use all available threads and was crashing the benchmark. I didn't notice the slider had moved and assumed it was using some automatically decided number. It would probably help if there was the slider or a notification for thread selection on page 1 of setting up a project.
Congrats to the team on the new patch, benchmark statistics like the search space are pretty handy when deciding our own preprocessing steps.
Cheers
Haze
2025-04-03 2:26 AM
Thank you very much for your comments and improvement ideas.
And as I said, I will update you when we find the cause of your issue.
Julian