diff options
| author | Your Name <mahesha.comp@gmail.com> | 2020-06-24 00:38:56 +0530 |
|---|---|---|
| committer | Your Name <mahesha.comp@gmail.com> | 2020-06-24 00:39:41 +0530 |
| commit | cc9d69385659be32178506a38b4f2e112ed01ad4 (patch) | |
| tree | 5f222bf1c133b37f879c306f43b8cb6691d24081 /llvm/lib/Bitcode/Writer/BitcodeWriter.cpp | |
| parent | fb34345e363ae00e6a2f7979ef524f4a4ffae582 (diff) | |
| download | llvm-cc9d69385659be32178506a38b4f2e112ed01ad4.zip llvm-cc9d69385659be32178506a38b4f2e112ed01ad4.tar.gz llvm-cc9d69385659be32178506a38b4f2e112ed01ad4.tar.bz2 | |
[AMDGPU/MemOpsCluster] Implement new heuristic for computing max mem ops cluster size
Summary:
Make use of both the - (1) clustered bytes and (2) cluster length, to decide on
the max number of mem ops that can be clustered. On an average, when loads
are dword or smaller, consider `5` as max threshold, otherwise `4`. This
heuristic is purely based on different experimentation conducted, and there is
no analytical logic here.
Reviewers: foad, rampitec, arsenm, vpykhtin
Reviewed By: rampitec
Subscribers: llvm-commits, kerbowa, hiraditya, t-tye, Anastasia, tpr, dstuttard, yaxunl, nhaehnle, wdng, jvesely, kzhuravl, thakis
Tags: #llvm
Differential Revision: https://reviews.llvm.org/D82393
Diffstat (limited to 'llvm/lib/Bitcode/Writer/BitcodeWriter.cpp')
0 files changed, 0 insertions, 0 deletions