@@ -214,9 +214,9 @@ gpu.module @test {
214214
215215
216216 // CHECK-LABEL: load_with_offsets
217- // CHECK-SAME: [[arg0:%.+]]: memref<64xf32>
218- // CHECK-COUNT-2: xegpu.load {{.*}}[{{.*}}], {{.*}} <{chunk_size = 1 : i64, l1_hint = #xegpu.cache_hint<cached>}> : memref<64xf32> , vector<16xindex>, vector<16xi1> -> vector<16xf32>
219- gpu.func @load_with_offsets (%src: memref < 64 x f32 > ) -> vector <32 xf32 > {
217+ // CHECK-SAME: [[arg0:%.+]]: ui64
218+ // CHECK-COUNT-2: xegpu.load {{.*}}[{{.*}}], {{.*}} <{chunk_size = 1 : i64, l1_hint = #xegpu.cache_hint<cached>}> : ui64 , vector<16xindex>, vector<16xi1> -> vector<16xf32>
219+ gpu.func @load_with_offsets (%src: ui64 ) -> vector <32 xf32 > {
220220 %cst = arith.constant dense <[
221221 0 , 8 , 16 , 24 , 32 , 40 , 48 , 56 ,
222222 64 , 72 , 80 , 88 , 96 , 104 , 112 , 120 ,
@@ -226,7 +226,7 @@ gpu.module @test {
226226
227227 %c17 = arith.constant 17 : index
228228 %mask = vector.create_mask %c17: vector <32 xi1 >
229- %ld = xegpu.load %src [%cst ], %mask {chunk_size = 1 , layout_result_0 = #xegpu.layout <inst_data = [16 ]>, l1_hint = #xegpu.cache_hint <cached >} : memref < 64 x f32 > , vector <32 xindex >, vector <32 xi1 > -> vector <32 xf32 >
229+ %ld = xegpu.load %src [%cst ], %mask {chunk_size = 1 , layout_result_0 = #xegpu.layout <inst_data = [16 ]>, l1_hint = #xegpu.cache_hint <cached >} : ui64 , vector <32 xindex >, vector <32 xi1 > -> vector <32 xf32 >
230230
231231 gpu.return %ld : vector <32 xf32 >
232232 }
@@ -279,9 +279,9 @@ gpu.module @test {
279279 //-----
280280
281281 // CHECK-LABEL: store_with_offsets
282- // CHECK-SAME: [[arg0:%.+]]: memref<64xf32>
283- // CHECK-COUNT-2: xegpu.store {{.*}}[{{.*}}], {{.*}} <{chunk_size = 1 : i64, l1_hint = #xegpu.cache_hint<cached>}> : vector<16xf32>, memref<64xf32> , vector<16xindex>, vector<16xi1>
284- gpu.func @store_with_offsets (%src: memref < 64 x f32 > ) {
282+ // CHECK-SAME: [[arg0:%.+]]: ui64
283+ // CHECK-COUNT-2: xegpu.store {{.*}}[{{.*}}], {{.*}} <{chunk_size = 1 : i64, l1_hint = #xegpu.cache_hint<cached>}> : vector<16xf32>, ui64 , vector<16xindex>, vector<16xi1>
284+ gpu.func @store_with_offsets (%src: ui64 ) {
285285 %cst = arith.constant dense <[
286286 0 , 8 , 16 , 24 , 32 , 40 , 48 , 56 ,
287287 64 , 72 , 80 , 88 , 96 , 104 , 112 , 120 ,
@@ -293,7 +293,7 @@ gpu.module @test {
293293 %mask = vector.create_mask %c17: vector <32 xi1 >
294294
295295 %st_vec = arith.constant dense <1023.0 >: vector <32 xf32 >
296- xegpu.store %st_vec , %src [%cst ], %mask {chunk_size = 1 , layout = #xegpu.layout <inst_data = [16 ]>, l1_hint = #xegpu.cache_hint <cached >} : vector <32 xf32 >, memref < 64 x f32 > , vector <32 xindex >, vector <32 xi1 >
296+ xegpu.store %st_vec , %src [%cst ], %mask {chunk_size = 1 , layout = #xegpu.layout <inst_data = [16 ]>, l1_hint = #xegpu.cache_hint <cached >} : vector <32 xf32 >, ui64 , vector <32 xindex >, vector <32 xi1 >
297297
298298 gpu.return
299299 }
@@ -364,14 +364,14 @@ gpu.module @test {
364364
365365//-----
366366 // CHECK-LABEL: load_with_offsets_chunk
367- // CHECK-SAME: [[arg0:%.+]]: memref<64xf32>
367+ // CHECK-SAME: [[arg0:%.+]]: ui64
368368 // CHECK: [[cst:%.+]] = arith.constant dense<0.000000e+00> : vector<32x4xf32>
369369 // CHECK: [[cst0:%.+]] = arith.constant dense<[130, 138, 146, 154, 162, 170, 178, 186, 194, 202, 210, 218, 226, 234, 242, 250]> : vector<16xindex>
370370 // CHECK: [[cst1:%.+]] = arith.constant dense<[2, 10, 18, 26, 34, 42, 50, 58, 66, 74, 82, 90, 98, 106, 114, 122]> : vector<16xindex>
371371 // CHECK: [[cst2:%.+]] = arith.constant dense<[128, 136, 144, 152, 160, 168, 176, 184, 192, 200, 208, 216, 224, 232, 240, 248]> : vector<16xindex>
372372 // CHECK: [[cst3:%.+]] = arith.constant dense<[0, 8, 16, 24, 32, 40, 48, 56, 64, 72, 80, 88, 96, 104, 112, 120]> : vector<16xindex>
373- // CHECK-COUNT-4: xegpu.load {{.*}}[{{.*}}], {{.*}} <{chunk_size = 2 : i64, l1_hint = #xegpu.cache_hint<cached>}> : memref<64xf32> , vector<16xindex>, vector<16xi1> -> vector<16x2xf32>
374- gpu.func @load_with_offsets_chunk (%src: memref < 64 x f32 > ) -> vector <32 x4 xf32 > {
373+ // CHECK-COUNT-4: xegpu.load {{.*}}[{{.*}}], {{.*}} <{chunk_size = 2 : i64, l1_hint = #xegpu.cache_hint<cached>}> : ui64 , vector<16xindex>, vector<16xi1> -> vector<16x2xf32>
374+ gpu.func @load_with_offsets_chunk (%src: ui64 ) -> vector <32 x4 xf32 > {
375375 %cst = arith.constant dense <[
376376 0 , 8 , 16 , 24 , 32 , 40 , 48 , 56 ,
377377 64 , 72 , 80 , 88 , 96 , 104 , 112 , 120 ,
@@ -381,7 +381,7 @@ gpu.module @test {
381381
382382 %c17 = arith.constant 17 : index
383383 %mask = vector.create_mask %c17: vector <32 xi1 >
384- %ld = xegpu.load %src [%cst ], %mask {chunk_size = 4 , layout_result_0 = #xegpu.layout <inst_data = [16 , 2 ]>, l1_hint = #xegpu.cache_hint <cached >} : memref < 64 x f32 > , vector <32 xindex >, vector <32 xi1 > -> vector <32 x4 xf32 >
384+ %ld = xegpu.load %src [%cst ], %mask {chunk_size = 4 , layout_result_0 = #xegpu.layout <inst_data = [16 , 2 ]>, l1_hint = #xegpu.cache_hint <cached >} : ui64 , vector <32 xindex >, vector <32 xi1 > -> vector <32 x4 xf32 >
385385 gpu.return %ld : vector <32 x4 xf32 >
386386 }
387387
@@ -410,14 +410,14 @@ gpu.module @test {
410410
411411//-----
412412 // CHECK-LABEL: store_with_offsets_chunk
413- // CHECK-SAME: [[arg0:%.+]]: memref<64xf32>
413+ // CHECK-SAME: [[arg0:%.+]]: ui64
414414 // CHECK: [[cst:%.+]] = arith.constant dense<1.023000e+03> : vector<16x2xf32
415415 // CHECK: [[cst0:%.+]] = arith.constant dense<[130, 138, 146, 154, 162, 170, 178, 186, 194, 202, 210, 218, 226, 234, 242, 250]> : vector<16xindex>
416416 // CHECK: [[cst1:%.+]] = arith.constant dense<[2, 10, 18, 26, 34, 42, 50, 58, 66, 74, 82, 90, 98, 106, 114, 122]> : vector<16xindex>
417417 // CHECK: [[cst2:%.+]] = arith.constant dense<[128, 136, 144, 152, 160, 168, 176, 184, 192, 200, 208, 216, 224, 232, 240, 248]> : vector<16xindex>
418418 // CHECK: [[cst3:%.+]] = arith.constant dense<[0, 8, 16, 24, 32, 40, 48, 56, 64, 72, 80, 88, 96, 104, 112, 120]> : vector<16xindex>
419- // CHECK-COUNT-4: xegpu.store {{.*}}[{{.*}}], {{.*}} <{chunk_size = 2 : i64, l1_hint = #xegpu.cache_hint<cached>}> : vector<16x2xf32>, memref<64xf32> , vector<16xindex>, vector<16xi1>
420- gpu.func @store_with_offsets_chunk (%src: memref < 64 x f32 > ) {
419+ // CHECK-COUNT-4: xegpu.store {{.*}}[{{.*}}], {{.*}} <{chunk_size = 2 : i64, l1_hint = #xegpu.cache_hint<cached>}> : vector<16x2xf32>, ui64 , vector<16xindex>, vector<16xi1>
420+ gpu.func @store_with_offsets_chunk (%src: ui64 ) {
421421 %cst = arith.constant dense <[
422422 0 , 8 , 16 , 24 , 32 , 40 , 48 , 56 ,
423423 64 , 72 , 80 , 88 , 96 , 104 , 112 , 120 ,
@@ -429,7 +429,7 @@ gpu.module @test {
429429 %mask = vector.create_mask %c17: vector <32 xi1 >
430430
431431 %st_vec = arith.constant dense <1023. >: vector <32 x4 xf32 >
432- xegpu.store %st_vec , %src [%cst ], %mask {chunk_size = 4 , layout = #xegpu.layout <inst_data = [16 , 2 ]>, l1_hint = #xegpu.cache_hint <cached >} : vector <32 x4 xf32 >, memref < 64 x f32 > , vector <32 xindex >, vector <32 xi1 >
432+ xegpu.store %st_vec , %src [%cst ], %mask {chunk_size = 4 , layout = #xegpu.layout <inst_data = [16 , 2 ]>, l1_hint = #xegpu.cache_hint <cached >} : vector <32 x4 xf32 >, ui64 , vector <32 xindex >, vector <32 xi1 >
433433 gpu.return
434434 }
435435
0 commit comments