fastmean: hoist INTEGER_RO()/REAL_RO() out of the element loops

[m-muecke]

base::mean() is currently quite a bit faster, REAL_ELT()/INTEGER_ELT() would also work and is used by base R, but was quite a bit slower than hoisting. For reference: https://github.com/r-devel/r-svn/blob/main/src/main/summary.c#L463

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.02%. Comparing base (d6b2382) to head (33392a1).
⚠️ Report is 1 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #7914   +/-   ##
=======================================
  Coverage   99.02%   99.02%           
=======================================
  Files          88       88           
  Lines       17378    17381    +3     
=======================================
+ Hits        17208    17211    +3     
  Misses        170      170           

☔ View full report in Codecov by Harness.
📢 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.

[MichaelChirico]

Makes sense, are you able to share any benchmark demonstrating the impact?

[m-muecke]

Apple M2 Pro (arm64), macOS 27.0.1, R 4.6.1, Homebrew clang 23.1.3:

library(data.table)
setDTthreads(1L)
set.seed(1)
N = 1e7
DT = data.table(g = sample(10L, N, TRUE), d = rnorm(N), i = sample(1e6L, N, TRUE))
options(datatable.optimize = 1L)
bench::mark(DT[, mean(d), by = g], DT[, mean(i), by = g], min_iterations = 10, check = FALSE)
setkey(DT, g)
bench::mark(DT[, mean(d), by = g], DT[, mean(i), by = g], min_iterations = 10, check = FALSE)

1e7 rows, 10 groups, optimize=1L	master	PR
DT[, mean(d), by=g], double, keyed by g	54.1ms	30.0ms	1.81x
DT[, mean(d), by=g], double, unkeyed	103.1ms	81.2ms	1.27x
DT[, mean(i), by=g], integer, keyed by g	43.4ms	20.3ms	2.14x
DT[, mean(i), by=g], integer, unkeyed	87.9ms	64.6ms	1.36x

[MichaelChirico]

Great! Sorry, one more thing to add is base::mean() since that's emphasized in the PR description