Closed cmnbroad closed 9 months ago
I think the likelihood of issues is low, but its probably worth at least profiling to see if inlining helps. The main issue is that the variable being tested isn't mutated by this call; so another option would be to move this to where its mutated, which might be much less overhead. So I'll experiment a bit more.
We're integrating this as part of Yash's codecs PRs.
The containing method gets called frequently, and the validateArg call shows up as a hotspot (3-4%) when profiling writing CRAM .