We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
1 parent 1509fd5 commit f53375aCopy full SHA for f53375a
1 file changed
benchmarks/bench_load_latency.py
@@ -261,7 +261,9 @@ def print_run_settings(
261
for destination_rank in range(num_ranks):
262
delta = mm_end_cpu[destination_rank, :] - mm_begin_cpu[destination_rank, :]
263
avg_cc = float(delta.sum() / max(1, delta.size) / max(1, niter))
264
- local_latency[destination_rank] = avg_cc * 10 # since the value is updated every 10ns (cc freq is 100MHZ in MI300)
+ local_latency[destination_rank] = (
265
+ avg_cc * 10
266
+ ) # since the value is updated every 10ns (cc freq is 100MHZ in MI300)
267
268
latency_matrix = mpi_allgather(local_latency.cpu())
269
0 commit comments