HomeSort by relevance Sort by last modified time
    Searched refs:shared_memory_size (Results 1 - 3 of 3) sorted by null

  /external/tensorflow/tensorflow/core/kernels/
bias_op_gpu.cu.cc 202 int32 shared_memory_size = 0; local
204 shared_memory_size = bias_size * sizeof(typename AccumulatorType<T>::type);
207 if (shared_memory_size <= max_shared_memory_size) {
210 <<<config.block_count, config.thread_per_block, shared_memory_size,
topk_op_gpu.cu.cc 377 // shared_memory_size / (2 * (sizeof(int) + sizeof(T))) < k.
381 constexpr auto shared_memory_size = 48 << 10; // 48 KB
383 // shared_memory_size = (num_shards + 1) * heap_size <=>
384 num_shards = shared_memory_size / heap_size - 1;
400 auto shared_memory_size = (num_shards + 1) * k * sizeof(Entry<T>);
402 TopKKernel<<<batch_size, num_shards, shared_memory_size, stream>>>(
depthwise_conv_op_gpu.cu.cc 624 const int shared_memory_size =
628 num_outputs, device, kernel, shared_memory_size,
630 kernel<<<config.block_count, block_dim, shared_memory_size,
    [all...]

Completed in 197 milliseconds