Make inductor work with new triton kernel launch API (#610)

intel · Nov 7, 2024 · a2dbfc3 · a2dbfc3
1 parent a13f579
commit a2dbfc3
Showing 1 changed file with 3 additions and 1 deletion.
diff --git a/intel_extension_for_pytorch/_inductor/xpu/triton_heuristics.py b/intel_extension_for_pytorch/_inductor/xpu/triton_heuristics.py
@@ -31,6 +31,7 @@
 
     try:
         from triton.compiler.compiler import ASTSource
+        from triton.backends.compiler import GPUTarget
     except ImportError:
         warnings.warn(
             "XPU: Import error on ASTSource, if this is not the case, \
@@ -128,7 +129,8 @@ def _precompile_config(self, cfg: Config, warm_cache_only_with_cc: Dict):
                 ),
             )
 
-            target = (compile_meta["device_type"], cc)
+            warp_size = 32
+            target = GPUTarget(compile_meta["device_type"], cc, warp_size)
             options = {
                 "num_warps": compile_meta["num_warps"],
                 "num_stages": compile_meta["num_stages"],