I am using SpectraST for creating a spectral library out of a 71 GB msp.file, which lead to a very long remapping step. Like 4 days for 10 % progress. The PC I am using has an available 350 GB RAM, but it only runs with a fraction of the possible power ~ 10 GB. It also only uses 2 of 9 available cores.
I have seen that there are options for Comet and X!Tandem to increase the processing speed, like “spectrum,threads, so My question is, are there any ways to do the same for SpectraST? Or is there a specific limit for the performance level for SpectraST?
Thank you in advance
Hi Philip, while I am not completely certain, I don’t think SpectraST has any options for multi-threaded execution in the library import/conversion steps. I recall Henry saying in the past that these operations tend to be I/O bound mostly so multi-threading wouldn’t help very much? Maybe that was true when disks were slower than today. Since you have so much RAM, you might consider setting up a RAM disk:
to see if that speeds things up considerably? Maybe not for the reading so much as the writing? Read from an SSD and write to the RAM disk?
That may not help a lot, but it’s the only thing I can think of at the moment. Maybe Henry has other ideas?
You received this message because you are subscribed to the Google Groups "spctools-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to spctools-discu...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/spctools-discuss/848867d7-705b-412c-a6a3-eae181ae31e5n%40googlegroups.com.