Copying sequences around as the heap resized is significantly expensive.
This speeds up glrParse by ~35% (2.4 => 3.25 MB/s)
Paths
| Differential D128307
[pseudo] Store reduction sequences by pointer in heaps, instead of by value. ClosedPublic Authored by sammccall on Jun 21 2022, 3:20 PM.
Details Summary Copying sequences around as the heap resized is significantly expensive. This speeds up glrParse by ~35% (2.4 => 3.25 MB/s)
Diff Detail
Event Timelinehokein added inline comments.
This revision is now accepted and ready to land.Jun 22 2022, 7:19 AM sammccall added inline comments.
This revision was landed with ongoing or failed builds.Jun 23 2022, 10:41 AM Closed by commit rG7aff663b2a04: [pseudo] Store reduction sequences by pointer in heaps, instead of by value. (authored by sammccall). · Explain Why This revision was automatically updated to reflect the committed changes. sammccall marked an inline comment as done.
Diff 439462 clang-tools-extra/pseudo/lib/GLR.cpp
|
I start to have a nervous feeling that we're going to add more and more intermediate data structure, which increases the complexity of the code, but I think the improvement is large enough to justify.