Strange results in global CTF refinement

Rado · March 7, 2025, 1:27am

@hsnyder Yes, I will be happy to share the data.

hsnyder · March 7, 2025, 8:25pm

Thanks, @Rado. I will send a direct message with instructions.
–Harris

hsnyder · April 10, 2025, 9:35pm

We were able to identify the cause of the incorrect behaviour of the global CTF refinement job on certain systems and the issue is fixed in CryoSPARC v4.7, released yesterday. Thanks for your help in reporting and debugging this.

–Harris

rbs_sci · April 10, 2025, 11:52pm

Out of curiosity, what was the cause?

Rado · April 11, 2025, 8:07am

That’s great to hear @hsnyder! Thank you for your effort in figuring this out.

Rado

hsnyder · April 14, 2025, 2:14pm

@rbs_sci it was a concurrency bug. It was there for a while, but the performance improvements from the IO changes in 4.6 caused it to surface. That’s why turning off io_uring helped.