56fd68e7dd
CCCL (CUDA C++ Core Libraries) provides: - CUB: device/block/warp-level GPU primitives (reduce, scan, sort, topk) - Thrust: high-level parallel algorithms (transform_reduce, sort, scan) - libcudacxx: CUDA C++ standard library (atomics, barriers, memory) - cudax: experimental features (memory resources, allocators) - Tuning policies: per-SM hardware-specific algorithm parameters Competition optimization vectors mapped to CCCL: - Output TPS (83% weight): warp_reduce, block_reduce, device_topk - Input TPS (14% weight): device_scan, block_load, prefetch - Cache TPS (3% weight): prefix caching strategy patterns - Memory (0.9 util): pooled/cached/buddy allocators Source: https://github.com/NVIDIA/cccl (shallow clone, HEAD only) License: Apache-2.0
31 lines
990 B
CMake
31 lines
990 B
CMake
file(
|
|
GLOB test_srcs
|
|
RELATIVE "${CMAKE_CURRENT_LIST_DIR}"
|
|
CONFIGURE_DEPENDS
|
|
*.cu
|
|
*.cpp
|
|
)
|
|
|
|
foreach (thrust_target IN LISTS THRUST_TARGETS)
|
|
thrust_get_target_property(config_device ${thrust_target} DEVICE)
|
|
if (NOT config_device STREQUAL "CUDA")
|
|
continue()
|
|
endif()
|
|
|
|
foreach (test_src IN LISTS test_srcs)
|
|
get_filename_component(test_name "${test_src}" NAME_WLE)
|
|
string(PREPEND test_name "cuda.")
|
|
|
|
# Create two targets, one with RDC enabled, the other without. This tests
|
|
# both device-side behaviors -- the CDP kernel launch with RDC, and the
|
|
# serial fallback path without RDC.
|
|
thrust_add_test(seq_test_target ${test_name}.cdp_0 "${test_src}" ${thrust_target})
|
|
thrust_configure_cuda_target(${seq_test_target} RDC OFF)
|
|
|
|
if (THRUST_ENABLE_RDC_TESTS)
|
|
thrust_add_test(cdp_test_target ${test_name}.cdp_1 "${test_src}" ${thrust_target})
|
|
thrust_configure_cuda_target(${cdp_test_target} RDC ON)
|
|
endif()
|
|
endforeach()
|
|
endforeach()
|