gaspersic/triton
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

[Triton-MLIR][Backend] Fix the order in linear/delinear and a few bugs in reduce conversion (#851)

Browse Source 
1, fix the order in linearize/delinearize, which fix the error of order
in emitIndices;
2, fix the selecting of fast implementation in reduce codegen;
3, fix the redundant barrier in reduce codegen;
4, fix the index mapping of the second round of warp_shuffle in shuffle
version of reduce codegen.

Co-authored-by: Keren Zhou <kerenzhou@openai.com>
This commit is contained in:
goostavz
2022-11-09 02:10:09 +08:00
committed by GitHub
parent 303790da88
commit 080b4addf8
4 changed files with 282 additions and 246 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
include/triton/Analysis/Utility.h
View File

@@ -21,11 +21,11 @@ template <typename Int> Int product(llvm::ArrayRef<Int> arr) {
template <typename Int> Int ceil(Int m, Int n) { return (m + n - 1) / n; }
// output[i] = input[order[i]]
template <typename T>
SmallVector<T> reorder(ArrayRef<T> input, ArrayRef<unsigned> order) {
template <typename T, typename RES_T = T>
SmallVector<RES_T> reorder(ArrayRef<T> input, ArrayRef<unsigned> order) {
size_t rank = order.size();
assert(input.size() == rank);
SmallVector<T> result(rank);
SmallVector<RES_T> result(rank);
for (auto it : llvm::enumerate(order)) {
result[it.index()] = input[it.value()];
}