aboutsummaryrefslogtreecommitdiff
path: root/clang/lib/CodeGen/CodeGenFunction.cpp
diff options
context:
space:
mode:
authorS. VenkataKeerthy <31350914+svkeerthy@users.noreply.github.com>2025-10-09 15:59:52 -0700
committerGitHub <noreply@github.com>2025-10-09 15:59:52 -0700
commit9e0d3bcbf2a6597353c3ab5844941b92c4a2e70e (patch)
tree970740a972a33856b2fd4c66c9b815c7ab3a79a4 /clang/lib/CodeGen/CodeGenFunction.cpp
parente9205ca8cf4235647c3ec2a1946850751ce4e1c8 (diff)
downloadllvm-9e0d3bcbf2a6597353c3ab5844941b92c4a2e70e.zip
llvm-9e0d3bcbf2a6597353c3ab5844941b92c4a2e70e.tar.gz
llvm-9e0d3bcbf2a6597353c3ab5844941b92c4a2e70e.tar.bz2
[IR2Vec] Restrict caching only to Flow-Aware computation (#162559)
Removed all the caching maps (BB, Inst) in `Embedder` as we don't want to cache embeddings in general. Our earlier experiments on Symbolic embeddings show recomputation of embeddings is cheaper than cache lookups. OTOH, Flow-Aware embeddings would benefit from instruction level caching, as computing the embedding for an instruction would depend on the embeddings of other instructions in a function. So, retained instruction embedding caching logic only for Flow-Aware computation. This also necessitates an `invalidate` method that would clean up the cache when the embeddings would become invalid due to transformations.
Diffstat (limited to 'clang/lib/CodeGen/CodeGenFunction.cpp')
0 files changed, 0 insertions, 0 deletions