Bump IREE to f27feffa3d44c834a184ed3ea4b704316fd1769e (#1001)

Author: yzhang93

.github/workflows/ci-linux.yml

@@ -57,7 +57,7 @@ jobs:
       - name: Python deps
         run: |
           pip install -r third_party/iree/runtime/bindings/python/iree/runtime/build_requirements.txt
-          pip install pyyaml pybind11==2.13.6 nanobind==2.2.0
+          pip install pyyaml pybind11==2.13.6 nanobind==2.4.0
 
       - name: Enable cache
         uses: actions/cache/restore@v3

.github/workflows/ci-macos.yml

@@ -78,7 +78,7 @@ jobs:
       - name: Python deps
         run: |
           pip install -r third_party/iree/runtime/bindings/python/iree/runtime/build_requirements.txt
-          pip install pytest pybind11==2.13.6 nanobind==2.2.0
+          pip install pytest pybind11==2.13.6 nanobind==2.4.0
 
       - name: Enable cache
         uses: actions/cache/restore@v3

.github/workflows/ci-windows.yml

@@ -81,7 +81,7 @@ jobs:
       - name: Python deps
         run: |
           pip install -r third_party\iree\runtime\bindings\python\iree\runtime\build_requirements.txt
-          pip install pyyaml pybind11==2.13.6 nanobind==2.2.0
+          pip install pyyaml pybind11==2.13.6 nanobind==2.4.0
 
       - name: Enable cache
         uses: actions/cache/restore@v3

compiler/plugins/target/AMD-AIE/aievec/ConvertVectorToAIEVec.cpp

@@ -72,7 +72,7 @@ struct RedundantLoadStoreOptimizationPass
     patterns.add<SetInboundsToReadOp, SetInboundsToWriteOp>(
         patterns.getContext());
 
-    (void)applyPatternsAndFoldGreedily(op, std::move(patterns));
+    (void)applyPatternsGreedily(op, std::move(patterns));
     IRRewriter rewriter(&getContext());
     vector::transferOpflowOpt(rewriter, op);
   }

compiler/plugins/target/AMD-AIE/aievec/VectorToVectorConversions.cpp

@@ -1042,14 +1042,14 @@ struct CanonicalizeVectorForAIEVecPass
       RewritePatternSet patterns(context);
       patterns.add<CanonicalizeTrivialReadAccessSubviewOpPattern,
                    CanonicalizeTrivialWriteAccessSubviewOpPattern>(context);
-      (void)applyPatternsAndFoldGreedily(op, std::move(patterns));
+      (void)applyPatternsGreedily(op, std::move(patterns));
     }
     {
       // These must run before 'populateVectorBroadcastLoweringPatterns'
       // so that broadcasts can be matched before conversion to insert.
       RewritePatternSet patterns(context);
       populateBubbleSignExtensionsLate(patterns);
-      (void)applyPatternsAndFoldGreedily(op, std::move(patterns));
+      (void)applyPatternsGreedily(op, std::move(patterns));
     }
 
     {
@@ -1067,7 +1067,7 @@ struct CanonicalizeVectorForAIEVecPass
       mlir::vector::populateShapeCastFoldingPatterns(patterns);
       mlir::vector::populateDropUnitDimWithShapeCastPatterns(patterns);
       mlir::vector::populateVectorBroadcastLoweringPatterns(patterns);
-      (void)applyPatternsAndFoldGreedily(op, std::move(patterns));
+      (void)applyPatternsGreedily(op, std::move(patterns));
     }
 
     {
@@ -1076,7 +1076,7 @@ struct CanonicalizeVectorForAIEVecPass
       // cycles.
       RewritePatternSet patterns(context);
       populateBubbleSignExtensionsLate(patterns);
-      (void)applyPatternsAndFoldGreedily(op, std::move(patterns));
+      (void)applyPatternsGreedily(op, std::move(patterns));
     }
   }
 };

compiler/plugins/target/AMD-AIE/iree-amd-aie/Target/XCLBinGen.cpp

@@ -1021,6 +1021,7 @@ void addLowerToLLVMPasses(OpPassManager &pm) {
   ConvertFuncToLLVMPassOptions opts;
   opts.useBarePtrCallConv = true;
   pm.addPass(createConvertFuncToLLVMPass(opts));
+  pm.addPass(createArithToLLVMConversionPass());
   pm.addPass(createCanonicalizerPass());
   pm.addPass(createCSEPass());
   pm.addPass(createConvertControlFlowToLLVMPass());

compiler/plugins/target/AMD-AIE/iree-amd-aie/Transforms/AMDAIEAssignTiles.cpp

@@ -321,7 +321,7 @@ LogicalResult assignNonLocalTiles(RewriterBase &rewriter, Operation *op,
   // Find and fill the tile candidates.
   RewritePatternSet fillTilePatterns(context);
   fillTilePatterns.insert<FillTiles>(context, deviceModel);
-  if (failed(applyPatternsAndFoldGreedily(op, std::move(fillTilePatterns)))) {
+  if (failed(applyPatternsGreedily(op, std::move(fillTilePatterns)))) {
     return op->emitOpError()
            << "collection of tile candidates for logical objectFifos failed";
   }

compiler/plugins/target/AMD-AIE/iree-amd-aie/Transforms/AMDAIECanonicalizeDoublyStridedOp.cpp

@@ -211,7 +211,7 @@ void AMDAIECanonicalizeDoublyStridedOpPass::runOnOperation() {
   }
   populateCanonicalizeDoublyStridedOpPatterns(patterns, foldSingleDims,
                                               deviceModel);
-  if (failed(applyPatternsAndFoldGreedily(parentOp, std::move(patterns)))) {
+  if (failed(applyPatternsGreedily(parentOp, std::move(patterns)))) {
     parentOp->emitOpError(
         "failed to canonicalize doubly strided DMA operations");
     return signalPassFailure();

compiler/plugins/target/AMD-AIE/iree-amd-aie/Transforms/AMDAIECombineStridedOps.cpp

@@ -183,7 +183,7 @@ void AMDAIECombineStridedOpsPass::runOnOperation() {
   MLIRContext *context = &getContext();
   RewritePatternSet patterns(context);
   populateStridedOpCombinationPattern(patterns);
-  if (failed(applyPatternsAndFoldGreedily(parentOp, std::move(patterns)))) {
+  if (failed(applyPatternsGreedily(parentOp, std::move(patterns)))) {
     parentOp->emitOpError("failed to combine strided operations");
     return signalPassFailure();
   }

compiler/plugins/target/AMD-AIE/iree-amd-aie/Transforms/AMDAIEDistributeCoresAndObjectFifos.cpp

@@ -478,7 +478,7 @@ void AMDAIEDistributeCoresAndObjectFifosPass::runOnOperation() {
   // possible.
   RewritePatternSet unrollLocalLoopsPatterns(context);
   unrollLocalLoopsPatterns.insert<AMDAIEUnrollLocalLoops>(context);
-  if (failed(applyPatternsAndFoldGreedily(
+  if (failed(applyPatternsGreedily(
           moduleOp, std::move(unrollLocalLoopsPatterns)))) {
     moduleOp.emitOpError()
         << "loop unrolling of loops selected for parallel execution failed";