https://llvm.org/bugs/show_bug.cgi?id=26341
Bug ID: 26341
Summary: Cuda __device__ lambda does not work
Product: clang
Version: trunk
Hardware: PC
OS: Linux
Status: NEW
Severity: normal
Priority: P
Component: -New Bugs
Assignee: unassignedclangb...@nondot.org
Reporter: crtr...@sandia.gov
CC: llvm-bugs@lists.llvm.org
Classification: Unclassified
Created attachment 15731
--> https://llvm.org/bugs/attachment.cgi?id=15731&action=edit
Reproduction code
This is an experimental feature in CUDA 7.5 and expected to be non experimental
in CUDA 8. The attached simple test code fails to compile.
Code:
run( [=] __device__ (int i) {
d_a[i] = d_c;
});
Error:
main.cpp:23:12: error: lambda requires '()' before attribute specifier
run( [=] __device__ (int i) {
^
()
/usr/local/cuda/include/host_defines.h:189:9: note: expanded from macro
'__device__'
__location__(device)
^
/usr/local/cuda/include/host_defines.h:88:9: note: expanded from macro
'__location__'
__annotate__(a)
^
/usr/local/cuda/include/host_defines.h:86:9: note: expanded from macro
'__annotate__'
__attribute__((a))
^
main.cpp:23:23: error: expected body of lambda expression
run( [=] __device__ (int i) {
^
main.cpp:23:28: error: expected '(' for function-style cast or type
construction
run( [=] __device__ (int i) {
~~~ ^
3 errors generated.
I actually like to see to be able to do __host__ __device__ (which does not
work with NVCC right now) so that you can dispatch the lambda to the host or to
the device. Effectively I just want the operator of the auto generated struct
to be marked __host__ __device__. I am responsible to only capture stuff which
is accesible on the device. Furthermore one might want to restrict a __device__
lambda to only capture by value. i.e. [=] __host__ __device__ (...) {...}
--
You are receiving this mail because:
You are on the CC list for the bug.
_______________________________________________
llvm-bugs mailing list
llvm-bugs@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-bugs
