Reenable some mxfp tests for XPU (#5379)

Addresses #4062

* skips same as CUDA
* skip large block tests for PVC 
* add num_warps=8 for empty kernels case

Author: HBN-MichalSzy

python/test/unit/language/test_matmul.py

@@ -1217,7 +1217,15 @@ def test_mxfp8_mxfp4_matmul(M, N, K, BLOCK_M, BLOCK_N, BLOCK_K, NUM_STAGES, B_TR
         if (A_DATA_TYPE == 'float4' and not WITH_A_SCALE) or (B_DATA_TYPE == 'float4' and not WITH_B_SCALE):
             pytest.skip("Float4 without scale is tested in test_block_scale_fp4")
     elif is_xpu():
-        pytest.xfail("XPU does not natively support scaled mxfp8 & mxfp4 matmul")
+        if not (WITH_A_SCALE and WITH_B_SCALE):
+            pytest.skip("None scale has not been tested on XPU backend")
+        if not (A_DATA_TYPE == "float8e5" and B_DATA_TYPE == "float4"):
+            pytest.skip(f"(A: {A_DATA_TYPE}, B: {B_DATA_TYPE}) has not been tested on XPU backend")
+        if (BLOCK_M, BLOCK_N,
+                BLOCK_K) == (128, 256,
+                             256) and CONST_SCALE and triton.runtime.driver.active.utils.get_device_properties(
+                                 triton.runtime.driver.active.get_current_device())["max_shared_mem"] < 196608:
+            pytest.skip("XPU: Not enough shared memory")
     if not PACK_B_ALONG_K and B_DATA_TYPE != "float4":
         pytest.xfail("Pack along K can only be False for float4")
 
@@ -1288,6 +1296,8 @@ def create_operand(dtype: str, size0: int, size1: int, k_dim: int, transpose: bo
     kernel_kwargs = {}
     if is_hip():
         kernel_kwargs["matrix_instr_nonkdim"] = nonKDim
+    if is_xpu() and (128, 256, 256) == (BLOCK_M, BLOCK_N, BLOCK_K) and not CONST_SCALE and not PACK_B_ALONG_K:
+        kernel_kwargs["num_warps"] = 8
     out = mxfp8_mxfp4_matmul[grid](a, b, output, a_scale, b_scale, M, N, K, stride_scale, a.stride(0), a.stride(1),
                                    b.stride(0), b.stride(1), output.stride(0), output.stride(1), not CONST_SCALE,
                                    dtype_converter[A_DATA_TYPE], dtype_converter[B_DATA_TYPE], BLOCK_M, BLOCK_N,

scripts/test-triton.sh

@@ -402,7 +402,7 @@ run_mxfp_tests() {
   cd $TRITON_PROJ/python/test/unit
 
   TRITON_DISABLE_LINE_INFO=1 TRITON_TEST_SUITE=mxfp \
-    run_pytest_command -vvv -n ${PYTEST_MAX_PROCESSES:-8} --device xpu intel/test_mxfp_matmul.py
+    run_pytest_command -vvv -n ${PYTEST_MAX_PROCESSES:-8} --device xpu intel/test_mxfp_matmul.py language/test_matmul.py::test_mxfp8_mxfp4_matmul
 }
 
 run_scaled_dot_tests() {