File size: 1,709 Bytes
9ab9a5e |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 |
#blocked = #triton_gpu.blocked<{sizePerThread = [8], threadsPerWarp = [32], warpsPerCTA = [4], order = [0], CTAsPerCGA = [1], CTASplitNum = [1], CTAOrder = [0]}>
module attributes {"triton_gpu.compute-capability" = 89 : i32, "triton_gpu.num-ctas" = 1 : i32, "triton_gpu.num-warps" = 4 : i32, "triton_gpu.threads-per-warp" = 32 : i32} {
tt.func public @triton__0d1d2de(%arg0: !tt.ptr<f32, 1> {tt.divisibility = 16 : i32}, %arg1: !tt.ptr<bf16, 1> {tt.divisibility = 16 : i32}, %arg2: i32 {tt.divisibility = 16 : i32, tt.max_divisibility = 16 : i32}) attributes {noinline = false} {
%cst = arith.constant dense<12865792> : tensor<1024xi32, #blocked>
%c1024_i32 = arith.constant 1024 : i32
%0 = tt.get_program_id x : i32
%1 = arith.muli %0, %c1024_i32 : i32
%2 = tt.make_range {end = 1024 : i32, start = 0 : i32} : tensor<1024xi32, #blocked>
%3 = tt.splat %1 : (i32) -> tensor<1024xi32, #blocked>
%4 = arith.addi %3, %2 : tensor<1024xi32, #blocked>
%5 = arith.cmpi slt, %4, %cst : tensor<1024xi32, #blocked>
%6 = tt.splat %arg0 : (!tt.ptr<f32, 1>) -> tensor<1024x!tt.ptr<f32, 1>, #blocked>
%7 = tt.addptr %6, %4 : tensor<1024x!tt.ptr<f32, 1>, #blocked>, tensor<1024xi32, #blocked>
%8 = tt.load %7, %5 {cache = 1 : i32, evict = 1 : i32, isVolatile = false} : tensor<1024xf32, #blocked>
%9 = tt.splat %arg1 : (!tt.ptr<bf16, 1>) -> tensor<1024x!tt.ptr<bf16, 1>, #blocked>
%10 = tt.addptr %9, %4 : tensor<1024x!tt.ptr<bf16, 1>, #blocked>, tensor<1024xi32, #blocked>
%11 = arith.truncf %8 : tensor<1024xf32, #blocked> to tensor<1024xbf16, #blocked>
tt.store %10, %11, %5 {cache = 1 : i32, evict = 1 : i32} : tensor<1024xbf16, #blocked>
tt.return
}
}
|