4 sub-group tests fail on CUDA

[alexbatashev]

The following sub-group LIT tests fail on CUDA:

• broadcast_fp16.cpp
• broadcast_fp64.cpp
• broadcast.cpp
• load_store.cpp

Log extracts are below:

Testing:  0.. 10.. 20.. 30.. 40.. 50
FAIL: SYCL :: on-device/sub_group/broadcast.cpp (297 of 533)
******************** TEST 'SYCL :: on-device/sub_group/broadcast.cpp' FAILED ********************
Script:
--
: 'RUN: at line 1';   /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/bin/clang --driver-mode=g++ -fsycl -fsycl-targets=nvptx64-nvidia-cuda-sycldevice /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.src/sycl/test/on-device/sub_group/broadcast.cpp -o /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/tools/sycl/test/on-device/sub_group/Output/broadcast.cpp.tmp.out
: 'RUN: at line 2';   env SYCL_DEVICE_TYPE=HOST /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/tools/sycl/test/on-device/sub_group/Output/broadcast.cpp.tmp.out
: 'RUN: at line 3';   true /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/tools/sycl/test/on-device/sub_group/Output/broadcast.cpp.tmp.out
: 'RUN: at line 4';    env SYCL_DEVICE_TYPE=GPU SYCL_BE=PI_CUDA  /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/tools/sycl/test/on-device/sub_group/Output/broadcast.cpp.tmp.out
: 'RUN: at line 5';   true /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/tools/sycl/test/on-device/sub_group/Output/broadcast.cpp.tmp.out
--
Exit Code: 255
Command Output (stdout):
--
$ ":" "RUN: at line 1"
$ "/localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/bin/clang" "--driver-mode=g++" "-fsycl" "-fsycl-targets=nvptx64-nvidia-cuda-sycldevice" "/localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.src/sycl/test/on-device/sub_group/broadcast.cpp" "-o" "/localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/tools/sycl/test/on-device/sub_group/Output/broadcast.cpp.tmp.out"
# command stderr:
ptxas fatal   : Unresolved extern function '_Z22__spirv_GroupBroadcastjij'
clang-12: error: ptxas command failed with exit code 255 (use -v to see invocation)
clang version 12.0.0 (https://github.com/intel/llvm 9446801fe1689ebd21028aecedab7be538f082ec)
Target: x86_64-unknown-linux-gnu
Thread model: posix
InstalledDir: /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/bin
clang-12: note: diagnostic msg: Error generating preprocessed source(s).
error: command failed with exit status: 255
--
********************
Testing:  0.. 10.. 20.. 30.. 40.. 50.
FAIL: SYCL :: on-device/sub_group/broadcast_fp16.cpp (301 of 533)
******************** TEST 'SYCL :: on-device/sub_group/broadcast_fp16.cpp' FAILED ********************
Script:
--
: 'RUN: at line 1';   /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/bin/clang --driver-mode=g++ -fsycl -fsycl-targets=nvptx64-nvidia-cuda-sycldevice /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.src/sycl/test/on-device/sub_group/broadcast_fp16.cpp -o /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/tools/sycl/test/on-device/sub_group/Output/broadcast_fp16.cpp.tmp.out
: 'RUN: at line 2';    env SYCL_DEVICE_TYPE=GPU SYCL_BE=PI_CUDA  /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/tools/sycl/test/on-device/sub_group/Output/broadcast_fp16.cpp.tmp.out
--
Exit Code: 255
Command Output (stdout):
--
$ ":" "RUN: at line 1"
$ "/localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/bin/clang" "--driver-mode=g++" "-fsycl" "-fsycl-targets=nvptx64-nvidia-cuda-sycldevice" "/localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.src/sycl/test/on-device/sub_group/broadcast_fp16.cpp" "-o" "/localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/tools/sycl/test/on-device/sub_group/Output/broadcast_fp16.cpp.tmp.out"
# command stderr:
ptxas fatal   : Unresolved extern function '_Z22__spirv_GroupBroadcastjDF16_j'
clang-12: error: ptxas command failed with exit code 255 (use -v to see invocation)
clang version 12.0.0 (https://github.com/intel/llvm 9446801fe1689ebd21028aecedab7be538f082ec)
Target: x86_64-unknown-linux-gnu
Thread model: posix
InstalledDir: /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/bin
clang-12: note: diagnostic msg: Error generating preprocessed source(s).
error: command failed with exit status: 255
--
********************
Testing:  0.. 10.. 20.. 30.. 40.. 50.
FAIL: SYCL :: on-device/sub_group/broadcast_fp64.cpp (304 of 533)
******************** TEST 'SYCL :: on-device/sub_group/broadcast_fp64.cpp' FAILED ********************
Script:
--
: 'RUN: at line 1';   /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/bin/clang --driver-mode=g++ -fsycl -fsycl-targets=nvptx64-nvidia-cuda-sycldevice /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.src/sycl/test/on-device/sub_group/broadcast_fp64.cpp -o /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/tools/sycl/test/on-device/sub_group/Output/broadcast_fp64.cpp.tmp.out
: 'RUN: at line 2';   env SYCL_DEVICE_TYPE=HOST /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/tools/sycl/test/on-device/sub_group/Output/broadcast_fp64.cpp.tmp.out
: 'RUN: at line 3';   true /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/tools/sycl/test/on-device/sub_group/Output/broadcast_fp64.cpp.tmp.out
: 'RUN: at line 4';    env SYCL_DEVICE_TYPE=GPU SYCL_BE=PI_CUDA  /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/tools/sycl/test/on-device/sub_group/Output/broadcast_fp64.cpp.tmp.out
: 'RUN: at line 5';   true /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/tools/sycl/test/on-device/sub_group/Output/broadcast_fp64.cpp.tmp.out
--
Exit Code: 255
Command Output (stdout):
--
$ ":" "RUN: at line 1"
$ "/localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/bin/clang" "--driver-mode=g++" "-fsycl" "-fsycl-targets=nvptx64-nvidia-cuda-sycldevice" "/localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.src/sycl/test/on-device/sub_group/broadcast_fp64.cpp" "-o" "/localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/tools/sycl/test/on-device/sub_group/Output/broadcast_fp64.cpp.tmp.out"
# command stderr:
ptxas fatal   : Unresolved extern function '_Z22__spirv_GroupBroadcastjdj'
clang-12: error: ptxas command failed with exit code 255 (use -v to see invocation)
clang version 12.0.0 (https://github.com/intel/llvm 9446801fe1689ebd21028aecedab7be538f082ec)
Target: x86_64-unknown-linux-gnu
Thread model: posix
InstalledDir: /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/bin
clang-12: note: diagnostic msg: Error generating preprocessed source(s).
error: command failed with exit status: 255
--
********************
Testing:  0.. 10.. 20.. 30.. 40.. 50.
FAIL: SYCL :: on-device/sub_group/load_store.cpp (310 of 533)
******************** TEST 'SYCL :: on-device/sub_group/load_store.cpp' FAILED ********************
Script:
--
: 'RUN: at line 6';   /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/bin/clang --driver-mode=g++ -fsycl -fsycl-targets=nvptx64-nvidia-cuda-sycldevice /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.src/sycl/test/on-device/sub_group/load_store.cpp -o /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/tools/sycl/test/on-device/sub_group/Output/load_store.cpp.tmp.out
: 'RUN: at line 7';   env SYCL_DEVICE_TYPE=HOST /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/tools/sycl/test/on-device/sub_group/Output/load_store.cpp.tmp.out
: 'RUN: at line 8';   true /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/tools/sycl/test/on-device/sub_group/Output/load_store.cpp.tmp.out
: 'RUN: at line 9';    env SYCL_DEVICE_TYPE=GPU SYCL_BE=PI_CUDA  /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/tools/sycl/test/on-device/sub_group/Output/load_store.cpp.tmp.out
: 'RUN: at line 10';   true /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/tools/sycl/test/on-device/sub_group/Output/load_store.cpp.tmp.out
--
Exit Code: 255
Command Output (stdout):
--
$ ":" "RUN: at line 6"
$ "/localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/bin/clang" "--driver-mode=g++" "-fsycl" "-fsycl-targets=nvptx64-nvidia-cuda-sycldevice" "/localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.src/sycl/test/on-device/sub_group/load_store.cpp" "-o" "/localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/tools/sycl/test/on-device/sub_group/Output/load_store.cpp.tmp.out"
# command stderr:
ptxas fatal   : Unresolved extern function '_Z30__spirv_SubgroupBlockReadINTELIhET_PU3AS1Kh'
clang-12: error: ptxas command failed with exit code 255 (use -v to see invocation)
clang version 12.0.0 (https://github.com/intel/llvm 9446801fe1689ebd21028aecedab7be538f082ec)
Target: x86_64-unknown-linux-gnu
Thread model: posix
InstalledDir: /localdisk2/sycl_ci/buildbot/worker/Lit_With_Cuda/llvm.obj/bin
clang-12: note: diagnostic msg: Error generating preprocessed source(s).
error: command failed with exit status: 255
--