Searched refs:latency (Results 1 – 6 of 6) sorted by relevance
207 for latency in latencies:208 mins.append(latency.time_freq_start_sec)209 to_add = len(latency.time_freq_sec) * latency.time_freq_step_sec210 maxs.append(latency.time_freq_start_sec + to_add)368 def generate_avg_ms(baseline, latency): argument370 if latency is None:371 latency = baseline373 result_avg_ms = (latency.total_time_sec / latency.iterations)*1000.0374 if latency is baseline:400 def generate_latency_graph_entry(tag, latency, tmin, tmax): argument[all …]
64 for latency, confidence in zip(latencies, confidences):67 latencies.remove(latency)69 total_latency += latency84 for latency in latencies:85 sd += (latency - avg_latency) * (latency - avg_latency)
10 1. Average round-trip latency.11 2. Standard deviation of round-trip latency.
275 full_api, latency = line.rsplit(":", 1)278 profiling_data.values[api_name].append(long(latency))280 profiling_data.values[api_name] = [long(latency)]
293 int64_t latency = end_timestamp - start_timestamp; in ProcessTraceForLatencyProfiling() local295 if (latency < 0) { in ProcessTraceForLatencyProfiling()300 cout << full_api_name << ":" << latency << endl; in ProcessTraceForLatencyProfiling()
17 Test Suite). MLTS allows to evaluate NNAPI acceleration latency and accuracy