[Chapel Merge] Calculate the numBytes correctly

Branch: refs/heads/main
Revision: 57e356810e43e6a60f712505508705155f347e11
Author: ShreyasKhandekar
Link: Calculate the numBytes correctly by ShreyasKhandekar · Pull Request #24686 · chapel-lang/chapel · GitHub
Log Message:
Calculate the numBytes correctly (#24686)

In the sort performance explorers for CPU and GPU sorting, number of
bytes was hard-coded to be 8. This was resulting in incorrect
calculation for the MiB/s (sort speed) for types which were not 64 bits.
Replaced 8 with numBytes(eltType) to fix this issue.

[Reviewed by @e-kayrakli ]

Compare: Comparing a37c49637bbf2e6ba7fc5b296abbfa088e45e498...27cb60a64078c3b40c085f4d4b0e4c9c24bd08f3 · chapel-lang/chapel · GitHub

Diff:
M test/gpu/native/studies/sort/gpu-sort-performance-explorer.chpl
M test/library/packages/Sort/performance/sort-performance-explorer.chpl
https://github.com/chapel-lang/chapel/pull/24686.diff