happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

sampling : add assertions for contiguous tensors in async copy functions

This commit is contained in:
Daniel Bevenius 2025-11-24 21:00:03 +01:00
parent 883a87043a
commit a02adf4211
No known key found for this signature in database
1 changed files with 6 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
src/llama-context.cpp
View File

@ -1231,6 +1231,8 @@ static void copy_tensor_async_ints(
const uint32_t row = it->second; const uint32_t row = it->second;
GGML_ASSERT(row < sampled_size); GGML_ASSERT(row < sampled_size);
GGML_ASSERT(ggml_is_contiguous(tensor) && "sampled tokens tensor must be contiguous for async copy");
ggml_backend_t backend = ggml_backend_sched_get_tensor_backend(sched, tensor); ggml_backend_t backend = ggml_backend_sched_get_tensor_backend(sched, tensor);
ggml_backend_tensor_get_async(backend, tensor, sampled + row, 0, sizeof(sampled[row])); ggml_backend_tensor_get_async(backend, tensor, sampled + row, 0, sizeof(sampled[row]));
} }
@ -1253,6 +1255,8 @@ static void copy_tensor_async_floats(
const uint32_t row = it->second; const uint32_t row = it->second;
GGML_ASSERT(row < counts.size()); GGML_ASSERT(row < counts.size());
GGML_ASSERT(ggml_is_contiguous(tensor) && "logits/probs tensor must be contiguous for async copy");
ggml_backend_t backend = ggml_backend_sched_get_tensor_backend(sched, tensor); ggml_backend_t backend = ggml_backend_sched_get_tensor_backend(sched, tensor);
float * row_ptr = dst + (size_t) row * stride; float * row_ptr = dst + (size_t) row * stride;
ggml_backend_tensor_get_async(backend, tensor, row_ptr, 0, ggml_nbytes(tensor)); ggml_backend_tensor_get_async(backend, tensor, row_ptr, 0, ggml_nbytes(tensor));
@ -1279,6 +1283,8 @@ static void copy_tensor_async_candidates(
const uint32_t row = it->second; const uint32_t row = it->second;
GGML_ASSERT(row < counts.size()); GGML_ASSERT(row < counts.size());
GGML_ASSERT(ggml_is_contiguous(tensor) && "candidates tensor must be contiguous for async copy");
ggml_backend_t backend = ggml_backend_sched_get_tensor_backend(sched, tensor); ggml_backend_t backend = ggml_backend_sched_get_tensor_backend(sched, tensor);
llama_token * row_ptr = dst + (size_t) row * stride; llama_token * row_ptr = dst + (size_t) row * stride;
ggml_backend_tensor_get_async(backend, tensor, row_ptr, 0, ggml_nbytes(tensor)); ggml_backend_tensor_get_async(backend, tensor, row_ptr, 0, ggml_nbytes(tensor));