nikola/llama_cpp
1
0
Fork 0
Code Issues Pull Requests Actions 4 Packages Projects Releases Wiki Activity

CUDA: Various fixes to cpy.cu (#25000)

Browse Source 
* Add failing test-case to test-backend-ops

Extracted from https://github.com/ggml-org/llama.cpp/issues/24072

* Minimize repro with help of AI

N = 8 * (65535 - 1) + 1 = 524273

* Port and adjust workaround from 0ba798341e

Fall-back should share code, also relax y-z constraint to be inclusive

* Add test-case + fallback also for y dim

* Fix x-guards which is 2^{31}-1, so inlusive of INT_MAX

* Fix overflow problems for transposed copy kernel
This commit is contained in:
Oliver Simons
2026-06-25 17:29:23 +02:00
committed by GitHub
parent c7cddefcbd
commit 1ec44d178d
2 changed files with 37 additions and 29 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
tests/test-backend-ops.cpp
View File

@@ -8176,6 +8176,8 @@ static std::vector<std::unique_ptr<test_case>> make_test_cases_eval() {
test_cases.emplace_back(new test_cpy(GGML_TYPE_I32, GGML_TYPE_I32, {256, 4, 1, 1}, {-1,-1,-1,-1}, {0, 0, 0, 0}, {0, 0, 0, 0}, true));
test_cases.emplace_back(new test_cpy(GGML_TYPE_I32, GGML_TYPE_I32, {256, 1, 4, 1}, {-1,-1,-1,-1}, {1, 2, 0, 3}, {0, 0, 0, 0}));
test_cases.emplace_back(new test_cpy(GGML_TYPE_F32, GGML_TYPE_F32, {256, 1, 4, 1}, {-1,-1,-1,-1}, {1, 2, 0, 3}, {0, 0, 0, 0}));
test_cases.emplace_back(new test_cpy(GGML_TYPE_F32, GGML_TYPE_F32, {2, 2097121, 1, 1}, {-1,-1,-1,-1}, {1, 0, 2, 3}));
test_cases.emplace_back(new test_cpy(GGML_TYPE_F32, GGML_TYPE_F32, {2, 2, 524281, 1}, {-1,-1,-1,-1}, {1, 0, 2, 3}));
// CPY - different src/dst shapes (reshaping via CPY)
// Use permutations of {3, 5, 7, 32}. Total elements: 3*5*7*32 = 3360.