frollapply throttle usage of multiple threads

[jangorecki]

closes #7305

using same mechanism as we use for openmp throttle.
verbose nicely reports all details.

also updated timings in manual, which were 3 years old.

library(data.table)

x = rnorm(1024)
n = 10L
f = mean

options(datatable.verbose=TRUE)
setDTthreads(1)
system.time(frollapply(x, n, f, simplify=unlist))
#frollapply running on single CPU thread
#   user  system elapsed 
#  0.004   0.000   0.005 

setDTthreads(2)
system.time(frollapply(x, n, f, simplify=unlist))
#frollapply run on 2 CPU threads throttled to 1 threads, input length 1024
#   user  system elapsed 
#  0.005   0.000   0.004 

x = rnorm(1025)
system.time(frollapply(x, n, f, simplify=unlist))
#frollapply running on 2 CPU threads
#   user  system elapsed 
#  0.005   0.010   0.013 

x = rnorm(2048)
setDTthreads(3)
system.time(frollapply(x, n, f, simplify=unlist))
#frollapply run on 3 CPU threads throttled to 2 threads, input length 2048
#   user  system elapsed 
#  0.000   0.005   0.014 

options(datatable.verbose=FALSE)
x = rnorm(1024)
setDTthreads(1)
system.time(for (i in 1:1e3) frollapply(x, n, f, simplify=unlist))
#   user  system elapsed 
#  4.540   0.005   4.546 
setDTthreads(2)
system.time(for (i in 1:1e3) frollapply(x, n, f, simplify=unlist))
#   user  system elapsed 
#  4.572   0.004   4.578

---

This work was supported by the Medical Research Council, UK [grant number 1573].

[github-actions]

No obvious timing issues in HEAD=frollapply-throttle

Generated via commit b786d4a

Download link for the artifact containing the test results: ↓ atime-results.zip

Task	Duration
R setup and installing dependencies	3 minutes and 0 seconds
Installing different package versions	43 seconds
Running and plotting the test cases	2 minutes and 18 seconds

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.09%. Comparing base (251e88d) to head (b786d4a).
⚠️ Report is 2 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #7306   +/-   ##
=======================================
  Coverage   99.09%   99.09%           
=======================================
  Files          83       83           
  Lines       16073    16085   +12     
=======================================
+ Hits        15928    15940   +12     
  Misses        145      145           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.