superbenchmark/tests/data/nccl_reduce.log

54 строки
4.9 KiB
Plaintext

# nThread 1 nGpus 8 minBytes 1 maxBytes 8589934592 step: 2(factor) warmup iters: 5 iters: 20 validation: 0
#
# Using devices
# Rank 0 Pid 112476 on localhost device 0 [0x00] A100-SXM4-40GB
# Rank 1 Pid 112476 on localhost device 1 [0x00] A100-SXM4-40GB
# Rank 2 Pid 112476 on localhost device 2 [0x00] A100-SXM4-40GB
# Rank 3 Pid 112476 on localhost device 3 [0x00] A100-SXM4-40GB
# Rank 4 Pid 112476 on localhost device 4 [0x00] A100-SXM4-40GB
# Rank 5 Pid 112476 on localhost device 5 [0x00] A100-SXM4-40GB
# Rank 6 Pid 112476 on localhost device 6 [0x00] A100-SXM4-40GB
# Rank 7 Pid 112476 on localhost device 7 [0x00] A100-SXM4-40GB
#
# out-of-place in-place
# size count type redop root time algbw busbw error time algbw busbw error
# (B) (elements) (us) (GB/s) (GB/s) (us) (GB/s) (GB/s)
hostname:3442:3442 [0] NCCL INFO Launch mode Parallel
0 0 float sum 0 36.90 0.00 0.00 N/A 36.47 0.00 0.00 N/A
0 0 float sum 0 34.18 0.00 0.00 N/A 35.70 0.00 0.00 N/A
4 1 float sum 0 35.40 0.00 0.00 N/A 35.59 0.00 0.00 N/A
8 2 float sum 0 36.35 0.00 0.00 N/A 35.74 0.00 0.00 N/A
16 4 float sum 0 35.47 0.00 0.00 N/A 34.27 0.00 0.00 N/A
32 8 float sum 0 36.16 0.00 0.00 N/A 36.19 0.00 0.00 N/A
64 16 float sum 0 35.61 0.00 0.00 N/A 35.45 0.00 0.00 N/A
128 32 float sum 0 34.78 0.00 0.00 N/A 35.80 0.00 0.00 N/A
256 64 float sum 0 35.37 0.01 0.01 N/A 35.89 0.01 0.01 N/A
512 128 float sum 0 35.49 0.01 0.01 N/A 35.53 0.01 0.01 N/A
1024 256 float sum 0 35.38 0.03 0.03 N/A 35.52 0.03 0.03 N/A
2048 512 float sum 0 35.97 0.06 0.06 N/A 35.13 0.06 0.06 N/A
4096 1024 float sum 0 36.03 0.11 0.11 N/A 35.82 0.11 0.11 N/A
8192 2048 float sum 0 36.80 0.22 0.22 N/A 36.71 0.22 0.22 N/A
16384 4096 float sum 0 35.37 0.46 0.46 N/A 36.79 0.45 0.45 N/A
32768 8192 float sum 0 35.16 0.93 0.93 N/A 35.72 0.92 0.92 N/A
65536 16384 float sum 0 38.08 1.72 1.72 N/A 37.74 1.74 1.74 N/A
131072 32768 float sum 0 43.07 3.04 3.04 N/A 41.59 3.15 3.15 N/A
262144 65536 float sum 0 52.16 5.03 5.03 N/A 50.49 5.19 5.19 N/A
524288 131072 float sum 0 67.58 7.76 7.76 N/A 66.57 7.88 7.88 N/A
1048576 262144 float sum 0 76.74 13.66 13.66 N/A 80.47 13.03 13.03 N/A
2097152 524288 float sum 0 78.51 26.71 26.71 N/A 78.76 26.63 26.63 N/A
4194304 1048576 float sum 0 81.47 51.48 51.48 N/A 80.30 52.23 52.23 N/A
8388608 2097152 float sum 0 94.72 88.57 88.57 N/A 94.06 89.19 89.19 N/A
16777216 4194304 float sum 0 137.7 121.83 121.83 N/A 139.6 120.17 120.17 N/A
33554432 8388608 float sum 0 218.3 153.70 153.70 N/A 218.1 153.83 153.83 N/A
67108864 16777216 float sum 0 370.8 180.96 180.96 N/A 369.8 181.49 181.49 N/A
134217728 33554432 float sum 0 661.0 203.06 203.06 N/A 659.9 203.39 203.39 N/A
268435456 67108864 float sum 0 1251.4 214.52 214.52 N/A 1268.1 211.68 211.68 N/A
536870912 134217728 float sum 0 2421.6 221.70 221.70 N/A 2413.4 222.45 222.45 N/A
1073741824 268435456 float sum 0 4736.0 226.72 226.72 N/A 4757.9 225.68 225.68 N/A
2147483648 536870912 float sum 0 9323.5 230.33 230.33 N/A 9354.0 229.58 229.58 N/A
4294967296 1073741824 float sum 0 18594 230.99 230.99 N/A 18570 231.28 231.28 N/A
8589934592 2147483648 float sum 0 37613 228.38 228.38 N/A 37539 228.83 228.83 N/A
# Out of bounds values : 0 OK
# Avg bus bandwidth : 65.018
#