================
@@ -3623,6 +3623,9 @@ static llvm::omp::OpenMPOffloadMappingFlags
mapParentWithMembers(
LLVM::ModuleTranslation &moduleTranslation, llvm::IRBuilderBase &builder,
llvm::OpenMPIRBuilder &ompBuilder, DataLayout &dl, MapInfosTy
&combinedInfo,
MapInfoData &mapData, uint64_t mapDataIndex, bool isTargetParams) {
+ assert(!ompBuilder.Config.isTargetDevice() &&
+ "function only supported for host device codegen");
----------------
skatrak wrote:
Yes, I do agree that "host device" seems like a rather contradictory term,
since we generally just talk about "host" as the CPU and "device" as whatever
we're offloading to. The reason of using these terms is to align with OpenMP
terminology (5.2 spec, Section 1.2.1):
> **host device** The _device_ on which the _OpenMP program_ begins execution.
> **target device** A _device_ with respect to which the current _device_
> performs an operation, as specified by a _device construct_ or an OpenMP
> device memory routine.
This is also why the `-fopenmp-is-target-device` flag is called that and not
`-fopenmp-is-device` (which [used to be its
name](https://reviews.llvm.org/D154591)).
https://github.com/llvm/llvm-project/pull/137199
_______________________________________________
llvm-branch-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
