[Mlir-commits] [mlir] [MLIR][XeGPU] Clean up xegpu op tests (PR #144592)

[Jianhui Li]

================
@@ -154,89 +85,89 @@ gpu.func @test_load_nd_simt(%src: memref<8x16xf16>) {
   gpu.return
 }
 
-// CHECK: func @test_load_nd_vc_2(%[[arg0:.*]]: memref<8x16xf16>) {
-gpu.func @test_load_nd_vc_2(%src: memref<8x16xf16>) {
+// CHECK: func @test_subgroup_load_nd_2(%[[arg0:.*]]: memref<8x16xf16>) {
+gpu.func @test_subgroup_load_nd_2(%src: memref<8x16xf16>) {
   // CHECK: %[[R0:.*]] = xegpu.create_nd_tdesc %arg0[0, 0] : memref<8x16xf16> -> !xegpu.tensor_desc<16xf16>
   %1 = xegpu.create_nd_tdesc %src[0, 0] : memref<8x16xf16> -> !xegpu.tensor_desc<16xf16>
   // CHECK: %[[R1:.*]] = xegpu.load_nd %[[R0]] <{l1_hint = #xegpu.cache_hint<cached>, l2_hint = #xegpu.cache_hint<uncached>}> : !xegpu.tensor_desc<16xf16> -> vector<16xf16>
   %2 = xegpu.load_nd %1 <{l1_hint = #xegpu.cache_hint<cached>, l2_hint = #xegpu.cache_hint<uncached>}> : !xegpu.tensor_desc<16xf16> -> vector<16xf16>
   gpu.return
 }
 
-// CHECK: func @test_load_nd_simt_2(%[[arg0:.*]]: memref<8x16xf16>) {
-gpu.func @test_load_nd_simt_2(%src: memref<8x16xf16>) {
+// CHECK: func @test_lane_load_nd_2(%[[arg0:.*]]: memref<8x16xf16>) {
+gpu.func @test_lane_load_nd_2(%src: memref<8x16xf16>) {
   // CHECK: %[[R0:.*]] = xegpu.create_nd_tdesc %arg0[0, 0] : memref<8x16xf16> -> !xegpu.tensor_desc<16xf16>
   %1 = xegpu.create_nd_tdesc %src[0, 0] : memref<8x16xf16> -> !xegpu.tensor_desc<16xf16>
   // CHECK: %[[R1:.*]] = xegpu.load_nd %[[R0]] <{l1_hint = #xegpu.cache_hint<cached>, l2_hint = #xegpu.cache_hint<uncached>}> : !xegpu.tensor_desc<16xf16> -> vector<1xf16>
   %2 = xegpu.load_nd %1 <{l1_hint = #xegpu.cache_hint<cached>, l2_hint = #xegpu.cache_hint<uncached>}> : !xegpu.tensor_desc<16xf16> -> vector<1xf16>
   gpu.return
 }
 
-// CHECK: func @test_load_nd_vc_3(%[[arg0:.*]]: memref<24x32xf32>) {
-gpu.func @test_load_nd_vc_3(%src: memref<24x32xf32>) {
+// CHECK: func @test_subgroup_load_nd_3(%[[arg0:.*]]: memref<24x32xf32>) {
+gpu.func @test_subgroup_load_nd_3(%src: memref<24x32xf32>) {
   // CHECK: %[[R0:.*]] = xegpu.create_nd_tdesc %arg0[0, 0] : memref<24x32xf32> -> !xegpu.tensor_desc<8x16xf32>
   %1 = xegpu.create_nd_tdesc %src[0, 0] : memref<24x32xf32> -> !xegpu.tensor_desc<8x16xf32>
   // CHECK: %[[R1:.*]] = xegpu.load_nd %[[R0]] <{l1_hint = #xegpu.cache_hint<cached>, l2_hint = #xegpu.cache_hint<uncached>}> : !xegpu.tensor_desc<8x16xf32> -> vector<8x16xf32>
   %2 = xegpu.load_nd %1 <{l1_hint = #xegpu.cache_hint<cached>, l2_hint = #xegpu.cache_hint<uncached>}> : !xegpu.tensor_desc<8x16xf32> -> vector<8x16xf32>
   gpu.return
 }
 
-// CHECK: func @test_load_nd_simt_3(%[[arg0:.*]]: memref<24x32xf32>) {
-gpu.func @test_load_nd_simt_3(%src: memref<24x32xf32>) {
+// CHECK: func @test_lane_load_nd_3(%[[arg0:.*]]: memref<24x32xf32>) {
+gpu.func @test_lane_load_nd_3(%src: memref<24x32xf32>) {
   // CHECK: %[[R0:.*]] = xegpu.create_nd_tdesc %arg0[0, 0] : memref<24x32xf32> -> !xegpu.tensor_desc<8x16xf32>
   %1 = xegpu.create_nd_tdesc %src[0, 0] : memref<24x32xf32> -> !xegpu.tensor_desc<8x16xf32>
   // CHECK: %[[R1:.*]] = xegpu.load_nd %[[R0]] <{l1_hint = #xegpu.cache_hint<cached>, l2_hint = #xegpu.cache_hint<uncached>}> : !xegpu.tensor_desc<8x16xf32> -> vector<8xf32>
   %2 = xegpu.load_nd %1 <{l1_hint = #xegpu.cache_hint<cached>, l2_hint = #xegpu.cache_hint<uncached>}> : !xegpu.tensor_desc<8x16xf32> -> vector<8xf32>
   gpu.return
 }
 
-// CHECK: func @test_load_nd_vc_4(%[[arg0:.*]]: memref<24x32xf16>) {
-gpu.func @test_load_nd_vc_4(%src: memref<24x32xf16>) {
+// CHECK: func @test_subgroup_load_nd_4(%[[arg0:.*]]: memref<24x32xf16>) {
+gpu.func @test_subgroup_load_nd_4(%src: memref<24x32xf16>) {
----------------
Jianhui-Li wrote:

now we have prefix: subgroup and simt.  For operations which doesn't differ under these two level, we don't have either prefix. 

https://github.com/llvm/llvm-project/pull/144592