Merge triton-mlir branch - Complete rewrite of the backend from scratch (#1004)

This PR merges the `triton-mlir` branch, in which we have been quietly
rewriting the Triton backend from scratch to increase maintainability,
stability and ultimately performance. Changes to the runtime are
minimal, and this new version aims to remain backward-compatible with
the previous commit. The legacy backend is now officially deprecated,
but can still be accessed via the `legacy-backend` tag.

Co-authored-by: Keren Zhou <kerenzhou@openai.com>
Co-authored-by: Yan Chunwei <yanchunwei@outlook.com>
Co-authored-by: goostavz <109190422+goostavz@users.noreply.github.com>
Co-authored-by: Shintaro Iwasaki <siwasaki@fb.com>
Co-authored-by: Yan Da <dyanab@connect.ust.hk>
Co-authored-by: Jun Yang <yangjunpro@gmail.com>
Co-authored-by: Ian Bearman <ianb@microsoft.com>
Co-authored-by: Jason Ansel <jansel@jansel.net>
Co-authored-by: Qingyi Liu <qingyil@nvidia.com>
Co-authored-by: ben-zhang-609 <110140741+ben-zhang-609@users.noreply.github.com>
Co-authored-by: Chenggang Zhao <lyricz@yeah.net>
Co-authored-by: ben-zhang-609 <benzh609@gmail.com>
Co-authored-by: dongdongl <dongdongl@nvidia.com>

lib/Conversion/TritonGPUToLLVM/ConvertLayoutOpToLLVM.h

@@ -0,0 +1,24 @@
+#ifndef TRITON_CONVERSION_TRITONGPU_TO_LLVM_CONVERT_LAYOUT_OP_H
+#define TRITON_CONVERSION_TRITONGPU_TO_LLVM_CONVERT_LAYOUT_OP_H
+
+#include "TritonGPUToLLVMBase.h"
+
+using namespace mlir;
+using namespace mlir::triton;
+
+using ::mlir::triton::gpu::DotOperandEncodingAttr;
+
+bool isMmaToDotShortcut(MmaEncodingAttr &mmaLayout,
+                        DotOperandEncodingAttr &dotOperandLayout);
+
+void storeBlockedToShared(Value src, Value llSrc, ArrayRef<Value> srcStrides,
+                          ArrayRef<Value> srcIndices, Value dst, Value smemBase,
+                          Type elemPtrTy, Location loc,
+                          ConversionPatternRewriter &rewriter);
+
+void populateConvertLayoutOpToLLVMPatterns(
+    mlir::LLVMTypeConverter &typeConverter, RewritePatternSet &patterns,
+    int numWarps, AxisInfoAnalysis &axisInfoAnalysis,
+    const Allocation *allocation, Value smem, PatternBenefit benefit);
+
+#endif