ikawrakow/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-04-30 19:31:48 +00:00
Code Issues Packages Projects Releases Wiki Activity

llama-bench: add ability to turn off warmup runs

Browse Source 
So we don't need to wait forever on, e.g., benchmarks involving
long contexts.
This commit is contained in:
Iwan Kawrakow
2024-08-20 14:44:01 +02:00
parent ad456dc25b
commit 4257ea1e92
1 changed files with 18 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

24
examples/llama-bench/llama-bench.cpp
View File

@@ -237,6 +237,7 @@ struct cmd_params {
ggml_numa_strategy numa; ggml_numa_strategy numa;
int reps; int reps;
bool verbose; bool verbose;
bool warmup;
output_formats output_format; output_formats output_format;
output_formats output_format_stderr; output_formats output_format_stderr;
}; };
@@ -263,6 +264,7 @@ static const cmd_params cmd_params_defaults = {
/* numa */ GGML_NUMA_STRATEGY_DISABLED, /* numa */ GGML_NUMA_STRATEGY_DISABLED,
/* reps */ 5, /* reps */ 5,
/* verbose */ false, /* verbose */ false,
/* warmup */ true,
/* output_format */ MARKDOWN, /* output_format */ MARKDOWN,
/* output_format_stderr */ NONE, /* output_format_stderr */ NONE,
}; };
@@ -295,6 +297,7 @@ static void print_usage(int /* argc */, char ** argv) {
printf(" -o, --output <csv|json|md|sql> (default: %s)\n", output_format_str(cmd_params_defaults.output_format)); printf(" -o, --output <csv|json|md|sql> (default: %s)\n", output_format_str(cmd_params_defaults.output_format));
printf(" -oe, --output-err <csv|json|md|sql> (default: %s)\n", output_format_str(cmd_params_defaults.output_format_stderr)); printf(" -oe, --output-err <csv|json|md|sql> (default: %s)\n", output_format_str(cmd_params_defaults.output_format_stderr));
printf(" -v, --verbose (default: %s)\n", cmd_params_defaults.verbose ? "1" : "0"); printf(" -v, --verbose (default: %s)\n", cmd_params_defaults.verbose ? "1" : "0");
printf(" -w, --warmup <0|1> (default: %s)\n", cmd_params_defaults.warmup ? "1" : "0");
printf("\n"); printf("\n");
printf("Multiple values can be given for each parameter by separating them with ',' or by specifying the parameter multiple times.\n"); printf("Multiple values can be given for each parameter by separating them with ',' or by specifying the parameter multiple times.\n");
} }
@@ -338,6 +341,7 @@ static cmd_params parse_cmd_params(int argc, char ** argv) {
params.output_format_stderr = cmd_params_defaults.output_format_stderr; params.output_format_stderr = cmd_params_defaults.output_format_stderr;
params.reps = cmd_params_defaults.reps; params.reps = cmd_params_defaults.reps;
params.numa = cmd_params_defaults.numa; params.numa = cmd_params_defaults.numa;
params.warmup = cmd_params_defaults.warmup;
for (int i = 1; i < argc; i++) { for (int i = 1; i < argc; i++) {
arg = argv[i]; arg = argv[i];
@@ -555,6 +559,12 @@ static cmd_params parse_cmd_params(int argc, char ** argv) {
invalid_param = !output_format_from_str(argv[i], params.output_format_stderr); invalid_param = !output_format_from_str(argv[i], params.output_format_stderr);
} else if (arg == "-v" || arg == "--verbose") { } else if (arg == "-v" || arg == "--verbose") {
params.verbose = true; params.verbose = true;
} else if (arg == "-w" || arg == "--warmup") {
if (++i >= argc) {
invalid_param = true;
break;
}
params.warmup = std::stoi(argv[i]);
} else { } else {
invalid_param = true; invalid_param = true;
break; break;
@@ -1429,12 +1439,14 @@ int main(int argc, char ** argv) {
llama_kv_cache_clear(ctx); llama_kv_cache_clear(ctx);
// warmup run // warmup run
if (t.n_prompt > 0) { if (params.warmup) {
//test_prompt(ctx, std::min(t.n_batch, std::min(t.n_prompt, 32)), 0, t.n_batch, t.n_threads); if (t.n_prompt > 0) {
test_prompt(ctx, t.n_prompt, 0, t.n_batch, t.n_threads); //test_prompt(ctx, std::min(t.n_batch, std::min(t.n_prompt, 32)), 0, t.n_batch, t.n_threads);
} test_prompt(ctx, t.n_prompt, 0, t.n_batch, t.n_threads);
if (t.n_gen > 0) { }
test_gen(ctx, 1, 0, t.n_threads); if (t.n_gen > 0) {
test_gen(ctx, 1, 0, t.n_threads);
}
} }
for (int i = 0; i < params.reps; i++) { for (int i = 0; i < params.reps; i++) {