[BACKEND] Update LLVM version to llvm/llvm-project@7752e0a (triton-lang#6735)

Removed `addArgumentMaterialization` since that method was removed from
MLIR in
llvm/llvm-project@23e3cbb

Author: gflegar

third_party/amd/lib/TritonAMDGPUToLLVM/TritonGPUToLLVM.cpp

@@ -26,6 +26,7 @@
 #include "triton/Dialect/Triton/IR/Dialect.h"
 #include "triton/Dialect/TritonGPU/IR/Dialect.h"
 #include "triton/Dialect/TritonNvidiaGPU/IR/Dialect.h"
+#include "llvm/TargetParser/TargetParser.h"
 
 namespace mlir::triton {
 #define GEN_PASS_DEF_CONVERTTRITONAMDGPUTOLLVM
@@ -106,6 +107,15 @@ struct ConvertTritonAMDGPUToLLVM
       mod.emitError("unsupported target: '") << this->arch.getValue() << "'";
       return signalPassFailure();
     }
+    llvm::StringRef chipset =
+        llvm::AMDGPU::getArchNameAMDGCN(targetInfo.getGPUKind());
+    llvm::FailureOr<mlir::amdgpu::Chipset> maybeChipset =
+        mlir::amdgpu::Chipset::parse(chipset);
+    if (failed(maybeChipset)) {
+      mlir::emitError(mlir::UnknownLoc::get(&getContext()),
+                      "Invalid chipset name: " + chipset);
+      return signalPassFailure();
+    }
 
     mlir::LowerToLLVMOptions option(context);
     option.overrideIndexBitwidth(32);