Displaying 3 results from an estimated 3 matches for "loopops".
2020 Apr 23
7
Cannot build master
Hi,
Using master at b0a1c0b72c9c61f8b0a223e08f43498abb64f5e8, I cannot build
LLVM. I configured with:
CC=clang CXX=clang++ cmake -DCMAKE_INSTALL_PREFIX=$HOME/opt/llvm11-git \
-DCMAKE_BUILD_TYPE=Release \
-DLLVM_BUILD_LLVM_DYLIB=ON \
-DLLVM_LINK_LLVM_DYLIB=ON \
-DBUILD_SHARED_LIBS=OFF \
-DLLVM_ENABLE_EH=ON \
-DLLVM_ENABLE_RTTI=ON \
2020 Jan 15
3
[RFC] Writing loop transformations on the right representation is more productive
..., this is a major focus of the affine dialect
> <https://mlir.llvm.org/docs/Dialects/Affine/> (e.g. see affine.for on
> that page). MLIR is not limited to affine loops, that is just a choice
> made by the affine dialect - the loop dialect
> <https://mlir.llvm.org/docs/Dialects/LoopOps/> has more general
> constructs that are less developed.
>
This is definitely subjective question. I think that MLIR is closer to
LLVM-IR for how it is processed. Both have a sequence of passes running
over a single source of truth. Both allow walking the entire structure from
every inst...
2020 Jan 03
10
Writing loop transformations on the right representation is more productive
In the 2018 LLVM DevMtg [1], I presented some shortcomings of how LLVM
optimizes loops. In summary, the biggest issues are (a) the complexity
of writing a new loop optimization pass (including needing to deal
with a variety of low-level issues, a significant amount of required
boilerplate, the difficulty of analysis preservation, etc.), (b)
independent optimization heuristics and a fixed pass