New CRAN release before Bioc 3.8

[lima1]

The avoidance of gc() in https://github.com/HenrikBengtsson/PSCBS/blob/develop/NEWS#L12-L13 reduces the R CMD check runtime of one of my packages by a couple of minutes. Is it possible to make a new release before the next Bioconductor release?

Thanks a lot, Markus

[HenrikBengtsson]

Yes, I should be able to roll out PSCBS on CRAN before the next Bioc release. When is it?

[lima1]

Not in a while, usually mid October to early November.

[HenrikBengtsson]

It this is your PureCN (v1.11.11), correct? When I run revdepcheck, I don't see a speed up for PureCN's package tests between old and new PSCBS:

# PSCBS 0.63.0
> proc.time()
   user  system elapsed 
240.848   1.019 243.612 

# PSCBS 0.63.0-9000
> proc.time()
   user  system elapsed 
269.500   1.071 272.176 

[lima1]

Thanks for looking it this.

On my Macbook the devel version seems to be much faster:

Examples with CPU or elapsed time > 5s
                    user system elapsed
segmentationPSCBS 32.588  0.672  34.036
runAbsoluteCN     24.571  0.303  21.574
filterIntervals   18.108  0.299  17.124
segmentationCBS   14.809  0.196  13.355
findFocal         14.669  0.188  13.246
annotateTargets    7.006  0.650   8.750

Examples with CPU or elapsed time > 5s
                     user system elapsed
segmentationPSCBS 104.262  0.572 103.430
runAbsoluteCN      24.235  0.283  21.105
filterIntervals    19.096  0.455  18.697
findFocal          16.780  0.388  15.974
segmentationCBS    15.029  0.173  13.419
annotateTargets     6.281  0.380   7.257

[HenrikBengtsson]

Wow, but on the other hand, it makes sense if there are lots of segments - we basically got one gc() call per segment, which really adds up.

[HenrikBengtsson]

PSCBS 0.64.0 is now on CRAN.