What is the minimum number of processing cores in a Mac computer required to run a QIIME2 analysis in a faster way? I am thinking specifically in dada2 step. When I ran that step in a Linux server that uses 4 cores, it took me more that 10 hours for 64 samples. Possibly I will transition to a Mac computer soon and I want to select the best number of cores to speed up my analysis, even when I have more samples.
I would appreciate your feedback.
It depends on your computer. You can, in theory, use as many cores as you have on your mac. (I think mine has 4 but Ive never really checked?) If you go over, your process will slow down. This also depends on the sequencing depth of the samples; 64 samples on a single run will run close to the same time as 300 samples in my experience.
My suggestion (for what its worth) is that its often most efficient to do the compute-intensive steps off your laptop. These are things like quality filtering/denoising; tree building; and diversity calculations. Running these in parallel on a server potentially saves a lot of time on your end. Then, the rest can be run locally. If your ethics/IT allows it, there are secondary cluster/server options that are relatively cheap, like AWS, as you think about getting the right computational resources.
Just a small addition, with most modern macs (and other laptops as well actually) your CPU is capable of hyperthreading meaning that a quad-core processor can actually simulate 8 cores and you can assign up to 8 cores to a parallelizable task even if you technically only have four cores.
Thanks @jwdebelius and @Mehrbod_Estaki for the valuable information. I will be considering all these details!
This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.