fix(gpu): fix tree_add_chunks after 5547d92

zama-ai · Aug 2, 2024 · c3d88f6 · c3d88f6
1 parent 7fa9f33
commit c3d88f6
Showing 1 changed file with 3 additions and 1 deletion.
diff --git a/backends/tfhe-cuda-backend/cuda/src/integer/multiplication.cuh b/backends/tfhe-cuda-backend/cuda/src/integer/multiplication.cuh
@@ -234,7 +234,9 @@ __host__ void host_integer_sum_ciphertexts_vec_kb(
   int32_t h_smart_copy_in[r * num_blocks];
   int32_t h_smart_copy_out[r * num_blocks];
 
-  auto max_shared_memory = cuda_get_max_shared_memory(gpu_indexes[0]);
+  auto max_shared_memory = 0;
+  cudaDeviceGetAttribute(&max_shared_memory, cudaDevAttrMaxSharedMemoryPerBlock,
+                         0);
 
   // create lut object for message and carry
   // we allocate luts_message_carry in the host function (instead of scratch)