![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
I do also, always ran it with NUMA enabled. Though, since my July PCIe BIOS trouble were solved, I have been pretty reluctant to reboot. I would be grateful if you or Patriot could carry out the experiment suggested by AndyE
This is what I meant.Or they just have the setting reversed
horde@fah-003048cbc23e:~/stream-5.10$ taskset -c 1 ./stream_c.exe
-------------------------------------------------------------
STREAM version $Revision: 5.10 $
-------------------------------------------------------------
This system uses 8 bytes per array element.
-------------------------------------------------------------
Array size = 10000000 (elements), Offset = 0 (elements)
Memory per array = 76.3 MiB (= 0.1 GiB).
Total memory required = 228.9 MiB (= 0.2 GiB).
Each kernel will be executed 10 times.
The *best* time for each kernel (excluding the first iteration)
will be used to compute the reported bandwidth.
-------------------------------------------------------------
Your clock granularity/precision appears to be 1 microseconds.
Each test below will take on the order of 15724 microseconds.
(= 15724 clock ticks)
Increase the size of the arrays if this shows that
you are not getting at least 20 clock ticks per test.
-------------------------------------------------------------
WARNING -- The above is only a rough guideline.
For best results, please be sure you know the
precision of your system timer.
-------------------------------------------------------------
Function Best Rate MB/s Avg time Min time Max time
Copy: 8008.0 0.019991 0.019980 0.020009
Scale: 7781.4 0.020586 0.020562 0.020606
Add: 8117.2 0.029594 0.029567 0.029611
Triad: 8076.2 0.029746 0.029717 0.029762
-------------------------------------------------------------
Solution Validates: avg error less than 1.000000e-13 on all three arrays
-------------------------------------------------------------
horde@fah-003048cbc23e:~/stream-5.10$horde@fah-003048cbc23e:~/stream-5.10$ taskset -c 1 ./stream_c.exe
-------------------------------------------------------------
STREAM version $Revision: 5.10 $
-------------------------------------------------------------
This system uses 8 bytes per array element.
-------------------------------------------------------------
Array size = 10000000 (elements), Offset = 0 (elements)
Memory per array = 76.3 MiB (= 0.1 GiB).
Total memory required = 228.9 MiB (= 0.2 GiB).
Each kernel will be executed 10 times.
The *best* time for each kernel (excluding the first iteration)
will be used to compute the reported bandwidth.
-------------------------------------------------------------
Your clock granularity/precision appears to be 1 microseconds.
Each test below will take on the order of 15354 microseconds.
(= 15354 clock ticks)
Increase the size of the arrays if this shows that
you are not getting at least 20 clock ticks per test.
-------------------------------------------------------------
WARNING -- The above is only a rough guideline.
For best results, please be sure you know the
precision of your system timer.
-------------------------------------------------------------
Function Best Rate MB/s Avg time Min time Max time
Copy: 8700.9 0.018419 0.018389 0.018431
Scale: 8523.3 0.018804 0.018772 0.018899
Add: 9375.4 0.025628 0.025599 0.025652
Triad: 9317.1 0.025784 0.025759 0.025823
-------------------------------------------------------------
Solution Validates: avg error less than 1.000000e-13 on all three arrays
-------------------------------------------------------------
horde@fah-003048cbc23e:~/stream-5.10$horde@fah-003048cbc23e:~/stream-5.10$ taskset -c 1 ./stream_c.exe
-------------------------------------------------------------
STREAM version $Revision: 5.10 $
-------------------------------------------------------------
This system uses 8 bytes per array element.
-------------------------------------------------------------
Array size = 80000000 (elements), Offset = 0 (elements)
Memory per array = 610.4 MiB (= 0.6 GiB).
Total memory required = 1831.1 MiB (= 1.8 GiB).
Each kernel will be executed 100 times.
The *best* time for each kernel (excluding the first iteration)
will be used to compute the reported bandwidth.
-------------------------------------------------------------
Your clock granularity/precision appears to be 1 microseconds.
Each test below will take on the order of 107101 microseconds.
(= 107101 clock ticks)
Increase the size of the arrays if this shows that
you are not getting at least 20 clock ticks per test.
-------------------------------------------------------------
WARNING -- The above is only a rough guideline.
For best results, please be sure you know the
precision of your system timer.
-------------------------------------------------------------
Function Best Rate MB/s Avg time Min time Max time
Copy: 8126.6 0.157614 0.157507 0.161598
Scale: 8116.7 0.157810 0.157699 0.159920
Add: 8261.2 0.232485 0.232412 0.232542
Triad: 8246.4 0.232900 0.232828 0.232990
-------------------------------------------------------------
Solution Validates: avg error less than 1.000000e-13 on all three arrays
-------------------------------------------------------------
horde@fah-003048cbc23e:~/stream-5.10$horde@fah-003048cbc23e:~/stream-5.10$ taskset -c 1 ./stream_c.exe
-------------------------------------------------------------
STREAM version $Revision: 5.10 $
-------------------------------------------------------------
This system uses 8 bytes per array element.
-------------------------------------------------------------
Array size = 80000000 (elements), Offset = 0 (elements)
Memory per array = 610.4 MiB (= 0.6 GiB).
Total memory required = 1831.1 MiB (= 1.8 GiB).
Each kernel will be executed 100 times.
The *best* time for each kernel (excluding the first iteration)
will be used to compute the reported bandwidth.
-------------------------------------------------------------
Your clock granularity/precision appears to be 1 microseconds.
Each test below will take on the order of 101459 microseconds.
(= 101459 clock ticks)
Increase the size of the arrays if this shows that
you are not getting at least 20 clock ticks per test.
-------------------------------------------------------------
WARNING -- The above is only a rough guideline.
For best results, please be sure you know the
precision of your system timer.
-------------------------------------------------------------
Function Best Rate MB/s Avg time Min time Max time
Copy: 8851.6 0.144708 0.144606 0.148615
Scale: 8663.4 0.147841 0.147748 0.151805
Add: 9272.9 0.207140 0.207055 0.209344
Triad: 9209.8 0.208530 0.208473 0.209325
-------------------------------------------------------------
Solution Validates: avg error less than 1.000000e-13 on all three arrays
-------------------------------------------------------------
horde@fah-003048cbc23e:~/stream-5.10$indrek@2419x2:~$ numastat
node0 node1
numa_hit 266457046 215030938
numa_miss 36019018 97972016
numa_foreign 97972016 36019018
interleave_hit 3292 3182
local_node 266457007 215027747
other_node 36019057 97975207
indrek@2419x2:~$ numactl --hardware
available: 2 nodes (0-1)
node 0 cpus: 0 1 2 3 4 5
node 0 size: 2946 MB
node 0 free: 91 MB
node 1 cpus: 6 7 8 9 10 11
node 1 size: 3022 MB
node 1 free: 85 MB
node distances:
node 0 1
0: 10 20
1: 20 10local_node 266457007 215027747
other_node 36019057 97975207
Those would be with "reduced" support from manufacturer ? When I checked cErtain Sources I saw Quite Some differences in price. But I'm also concerned in power drawing/heat production as it suppose to fold 24/7; seems those pull 130watt each. Ivys run a bit cooler. My GPU rig keep my room already nice warm