From da1889834a036a63ead2b0ca5c9ed8967712568c Mon Sep 17 00:00:00 2001 From: slaren Date: Tue, 25 Jul 2023 14:32:20 +0200 Subject: ggml : improve graph build time via hash table lookup (#2329) * improve graph build time * ggml_tensor : use 1 bit per flag * use a hash table instead --- ggml.h | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) (limited to 'ggml.h') diff --git a/ggml.h b/ggml.h index 1870b62e..c309f136 100644 --- a/ggml.h +++ b/ggml.h @@ -442,7 +442,7 @@ extern "C" { void * extra; // extra things e.g. for ggml-cuda.cu - char padding[8]; + char padding[4]; }; static const size_t GGML_TENSOR_SIZE = sizeof(struct ggml_tensor); @@ -463,6 +463,11 @@ extern "C" { void * abort_callback_data; }; + // next prime after GGML_MAX_NODES + // #define GGML_GRAPH_HASHTABLE_SIZE 4099 + // next prime after GGML_MAX_NODES * 2 (nodes + leafs) + #define GGML_GRAPH_HASHTABLE_SIZE 8273 + // computation graph struct ggml_cgraph { int n_nodes; @@ -472,6 +477,8 @@ extern "C" { struct ggml_tensor * grads[GGML_MAX_NODES]; struct ggml_tensor * leafs[GGML_MAX_NODES]; + void * visited_hash_table[GGML_GRAPH_HASHTABLE_SIZE]; + // performance int perf_runs; int64_t perf_cycles; -- cgit v1.2.3