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
45 lines
1.1 KiB
Plaintext
45 lines
1.1 KiB
Plaintext
=============== array.normal begin ===============
|
|
cuda::std::array<int, 3> = {
|
|
[0] = -7,
|
|
[1] = 0,
|
|
[2] = 42
|
|
}
|
|
=============== array.normal end ===============
|
|
=============== array.empty begin ===============
|
|
cuda::std::array<int, 0>
|
|
=============== array.empty end ===============
|
|
=============== array.nested begin ===============
|
|
cuda::std::array<cuda::std::array<int, 2>, 2> = {
|
|
[0] = cuda::std::array<int, 2> = {
|
|
[0] = 13,
|
|
[1] = -5
|
|
},
|
|
[1] = cuda::std::array<int, 2> = {
|
|
[0] = 0,
|
|
[1] = 88
|
|
}
|
|
}
|
|
=============== array.nested end ===============
|
|
=============== array.alias begin ===============
|
|
cuda::std::array<int, 4> = {
|
|
[0] = -31,
|
|
[1] = 17,
|
|
[2] = 8,
|
|
[3] = -64
|
|
}
|
|
=============== array.alias end ===============
|
|
=============== array.update.before begin ===============
|
|
cuda::std::array<int, 3> = {
|
|
[0] = 6,
|
|
[1] = -91,
|
|
[2] = 52
|
|
}
|
|
=============== array.update.before end ===============
|
|
=============== array.update.after begin ===============
|
|
cuda::std::array<int, 3> = {
|
|
[0] = 3,
|
|
[1] = 85,
|
|
[2] = -12
|
|
}
|
|
=============== array.update.after end ===============
|