[clang] [llvm] [Clang][SYCL] Compile device code once per IGCA target (PR #239010)

Yury Plyakhin via cfe-commits cfe-commits at lists.llvm.org
Thu Nov 12 18:10:31 PST 2026


================
@@ -97,6 +98,12 @@ class Compilation {
   };
   std::map<TCArgsKey, llvm::opt::DerivedArgList *> TCArgs;
 
+  /// For SYCL, the Intel GPUs given in --offload-arch for each IGCA target,
+  /// e.g. xe-mtl-h and xe-mtl-u for igca_30r. Each list is sorted.
----------------
YuriPlyakhin wrote:

```suggestion
  /// For SYCL, maps a device toolchain and an IGCA target, e.g. igca_30r, to
  /// the Intel GPUs from --offload-arch that share that target, e.g. xe-mtl-h
  /// and xe-mtl-u. Each list is sorted and has no duplicates.
```

https://github.com/llvm/llvm-project/pull/239010


More information about the cfe-commits mailing list