gistfile1.txt

mlir-opt matmult.mlir -convert-linalg-to-loops -lower-affine -convert-scf-to-cf -convert-linalg-to-llvm -convert-memref-to-llvm -convert-func-to-llvm -reconcile-unrealized-casts > out.mlir
mlir-cpu-runner out.mlir -O3 -e main -entry-point-result=void --shared-libs=libmlir_runner_utils.dylib

matmult.mlir

    // C += A * B.
func.func @matmul(%A: memref<2048x2048xf64>, %B: memref<2048x2048xf64>, %C: memref<2048x2048xf64>) {
    affine.for %arg3 = 0 to 2048 {
        affine.for %arg4 = 0 to 2048 {
            affine.for %arg5 = 0 to 2048 {
                %a = affine.load %A[%arg3, %arg5] : memref<2048x2048xf64>
                %b = affine.load %B[%arg5, %arg4] : memref<2048x2048xf64>
                %ci = affine.load %C[%arg3, %arg4] : memref<2048x2048xf64>
                %p = arith.mulf %a, %b : f64
                %co = arith.addf %ci, %p : f64
                affine.store %co, %C[%arg3, %arg4] : memref<2048x2048xf64>
            }
        }
    }
return
}

func.func @main() {
    %A = memref.alloc() : memref<2048x2048xf64>
    %B = memref.alloc() : memref<2048x2048xf64>
    %C = memref.alloc() : memref<2048x2048xf64>
    %cf1 = llvm.mlir.constant(1.00000e+00 : f64) : f64

    linalg.fill ins(%cf1 : f64) outs(%A : memref<2048x2048xf64>)
    linalg.fill ins(%cf1 : f64) outs(%B : memref<2048x2048xf64>)
    linalg.fill ins(%cf1 : f64) outs(%C : memref<2048x2048xf64>)
    call @matmul(%A, %B, %C) : (memref<2048x2048xf64>, memref<2048x2048xf64>, memref<2048x2048xf64>) -> ()
    %C.cast = memref.cast %C : memref<2048x2048xf64> to memref<*xf64>
    call @printMemrefF64(%C.cast): (memref<*xf64>) -> ()
    return
}

func.func private @printMemrefF64(memref<*xf64>)