Skip to content

Instantly share code, notes, and snippets.

@kazuki
Created September 16, 2017 05:44
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save kazuki/b76797c7ceeebb04efcdded4d251a271 to your computer and use it in GitHub Desktop.
Save kazuki/b76797c7ceeebb04efcdded4d251a271 to your computer and use it in GitHub Desktop.
AMD Ryzen Threadripper 1950X Benchmark (ASRock X399 Taichi, Crucial ECC Unbuffered DDR4-2400 16GB x 4)
# NUMA-mode (Memory Interleaving: Channel)
Intel(R) Memory Latency Checker - v3.4
Measuring idle latencies (in ns)...
Numa node
Numa node 0 1
0 86.3 142.3
1 140.9 86.0
Measuring Peak Injection Memory Bandwidths for the system
Bandwidths are in MB/sec (1 MB/sec = 1,000,000 Bytes/sec)
Using all the threads from each core if Hyper-threading is enabled
Using traffic with the following read-write ratios
ALL Reads : 63173.3
3:1 Reads-Writes : 58214.7
2:1 Reads-Writes : 58082.0
1:1 Reads-Writes : 56281.7
Stream-triad like: 58941.3
Measuring Memory Bandwidths between nodes within system
Bandwidths are in MB/sec (1 MB/sec = 1,000,000 Bytes/sec)
Using all the threads from each core if Hyper-threading is enabled
Using Read-only traffic type
Numa node
Numa node 0 1
0 31816.2 18881.6
1 18884.9 31694.5
Measuring Loaded Latencies for the system
Using all the threads from each core if Hyper-threading is enabled
Using Read-only traffic type
Inject Latency Bandwidth
Delay (ns) MB/sec
==========================
00000 570.65 63131.0
00002 562.00 63051.6
00008 524.69 63070.6
00015 517.48 63049.8
00050 476.51 62991.1
00100 446.76 63070.3
00200 158.61 61226.9
00300 110.62 43529.5
00400 103.94 33898.2
00500 99.80 27827.4
00700 95.66 20475.0
01000 93.05 14754.5
01300 91.98 11605.4
01700 90.60 9104.7
02500 89.36 6473.8
03500 88.94 4850.4
05000 88.40 3624.5
09000 87.90 2344.2
20000 86.94 1465.1
Measuring cache-to-cache transfer latency (in ns)...
Local Socket L2->L2 HIT latency 20.6
Local Socket L2->L2 HITM latency 33.0
Remote Socket L2->L2 HITM latency (data address homed in writer socket)
Reader Numa Node
Writer Numa Node 0 1
0 - 152.1
1 152.0 -
Remote Socket L2->L2 HITM latency (data address homed in reader socket)
Reader Numa Node
Writer Numa Node 0 1
0 - 149.5
1 150.6 -
# UMA-mode (Memory Interleaving: Auto)
Intel(R) Memory Latency Checker - v3.4
Measuring idle latencies (in ns)...
Memory node
Socket 0 1
0 115.2 115.1
1 112.7 112.9
Measuring Peak Injection Memory Bandwidths for the system
Bandwidths are in MB/sec (1 MB/sec = 1,000,000 Bytes/sec)
Using all the threads from each core if Hyper-threading is enabled
Using traffic with the following read-write ratios
ALL Reads : 57361.6
3:1 Reads-Writes : 52071.2
2:1 Reads-Writes : 52830.5
1:1 Reads-Writes : 54686.0
Stream-triad like: 52459.6
Measuring Memory Bandwidths between nodes within system
Bandwidths are in MB/sec (1 MB/sec = 1,000,000 Bytes/sec)
Using all the threads from each core if Hyper-threading is enabled
Using Read-only traffic type
Memory node
Socket 0 1
0 59872.6 59915.3
1 59909.9 59851.7
Measuring Loaded Latencies for the system
Using all the threads from each core if Hyper-threading is enabled
Using Read-only traffic type
Inject Latency Bandwidth
Delay (ns) MB/sec
==========================
00000 641.42 57368.4
00002 625.04 57447.5
00008 570.28 57428.4
00015 559.88 57405.1
00050 521.16 57368.6
00100 508.19 57377.6
00200 261.15 58330.8
00300 151.15 43428.7
00400 141.81 33722.3
00500 135.53 27640.1
00700 132.42 20314.0
01000 127.05 14588.6
01300 125.09 11442.9
01700 123.41 8942.1
02500 121.54 6296.7
03500 120.27 4672.3
05000 116.39 3456.6
09000 115.06 2176.4
20000 115.15 1286.7
Measuring cache-to-cache transfer latency (in ns)...
Local Socket L2->L2 HIT latency 20.0
Local Socket L2->L2 HITM latency 39.3
Remote Socket L2->L2 HITM latency (data address homed in writer socket)
Reader Socket
Writer Socket 0 1
0 - 38.4
1 38.3 -
Remote Socket L2->L2 HITM latency (data address homed in reader socket)
Reader Socket
Writer Socket 0 1
0 - 38.5
1 38.2 -
# UMA-mode
$ numactl -H
available: 1 nodes (0)
node 0 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
node 0 size: 64354 MB
node 0 free: 59475 MB
node distances:
node 0
0: 10
# NUMA-mode
$ numactl -H
available: 2 nodes (0-1)
node 0 cpus: 0 1 2 3 4 5 6 7 16 17 18 19 20 21 22 23
node 0 size: 32130 MB
node 0 free: 31387 MB
node 1 cpus: 8 9 10 11 12 13 14 15 24 25 26 27 28 29 30 31
node 1 size: 32235 MB
node 1 free: 31612 MB
node distances:
node 0 1
0: 10 16
1: 16 10
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment