================
@@ -132,6 +130,12 @@ class NVPTXTargetCIRGenInfo : public TargetCIRGenInfo {
public:
NVPTXTargetCIRGenInfo(CIRGenTypes &cgt)
: TargetCIRGenInfo(std::make_unique<NVPTXABIInfo>(cgt)) {}
+
+ mlir::Type getCUDADeviceBuiltinSurfaceDeviceType() const override {
+ // CUDA surface is represented as a 64-bit handle on device
----------------
koparasy wrote:
I find the OG side comment more informative:
```
// On the device side, surface reference is represented as an object handle
// in 64-bit integer.
```
https://github.com/llvm/llvm-project/pull/196079
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
