0

I tested two systems, both running Linux Mint 18.1 by running the same program on them.

System 1 has the following features:

Architecture:          x86_64
CPU op-mode(s):        32-bit, 64-bit
Byte Order:            Little Endian
CPU(s):                2
On-line CPU(s) list:   0,1
Thread(s) per core:    1
Core(s) per socket:    2
Socket(s):             1
NUMA node(s):          1
Vendor ID:             GenuineIntel
CPU family:            6
Model:                 23
Model name:            Intel(R) Core(TM)2 Duo CPU     E7500  @ 2.93GHz
Stepping:              10
CPU MHz:               1603.000
CPU max MHz:           2936.0000
CPU min MHz:           1603.0000
BogoMIPS:              5866.45
L1d cache:             32K
L1i cache:             32K
L2 cache:              3072K
NUMA node0 CPU(s):     0,1
Flags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx lm constant_tsc arch_perfmon pebs bts rep_good nopl aperfmperf pni dtes64 monitor ds_cpl est tm2 ssse3 cx16 xtpr pdcm sse4_1 xsave lahf_lm dtherm
          `   total        used        free      shared  buff/cache`   available
Mem:           5967        1809         835          88        3321        3750
Low:           5967        5131         835

And system 2 :

Architecture:          x86_64
CPU op-mode(s):        32-bit, 64-bit
Byte Order:            Little Endian
CPU(s):                12
On-line CPU(s) list:   0-11
Thread(s) per core:    2
Core(s) per socket:    6
Socket(s):             1
NUMA node(s):          1
Vendor ID:             GenuineIntel
CPU family:            6
Model:                 63
Model name:            Intel(R) Core(TM) i7-5820K CPU @ 3.30GHz
Stepping:              2
CPU MHz:               3399.902
CPU max MHz:           3600.0000
CPU min MHz:           1200.0000
BogoMIPS:              6596.24
Virtualization:        VT-x
L1d cache:             32K
L1i cache:             32K
L2 cache:              256K
L3 cache:              15360K
NUMA node0 CPU(s):     0-11
Flags:                 fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf eagerfpu pni pclmulqdq dtes64 monitor ds_cpl vmx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm epb tpr_shadow vnmi flexpriority ept vpid fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid cqm xsaveopt cqm_llc cqm_occup_llc dtherm ida arat pln pts

              total        used        free      shared  buff/cache   available
Mem:          48203        5957         388         138       41857       41600
Low:          48203       47814         388

timing the program was as follows(real time):

system 1    2 cores          ~30  seconds
system 2    6 cores          ~120 seconds

I could not find an answer for why system 2 with better specs has lower performance. System 2 has 8 free cores. A few Matlab codes were running on 4 other cores. Also, during the process I checked that I have enough free RAM.

Edit: My program is a simple Euler integration paralleled by OpenMP. It doses not have lots of Input/Output. I limited the number of threads with export OMP_NUM_THREADS=6 on the terminal for system 2. Programs are running on my local systems. I have no idea what Matlab codes are doing. It's a shared machine and I use ssh to have access to that machine. I hope these details be helpful.

I did another timing with 2 and 6 thread:

system 1     2 threads   52 seconds 
system 2     2 threads   54 seconds
system 2     6 threads   4 minutes and 22.3492 seconds
system 3     6 threads   19 seconds

system 3 is :

Architecture:          x86_64
CPU op-mode(s):        32-bit, 64-bit
Byte Order:            Little Endian
CPU(s):                8
On-line CPU(s) list:   0-7
Thread(s) per core:    2
Core(s) per socket:    4
Socket(s):             1
NUMA node(s):          1
Vendor ID:             GenuineIntel
CPU family:            6
Model:                 60
Stepping:              3
CPU MHz:               800.000
BogoMIPS:              7982.20
Virtualization:        VT-x
L1d cache:             32K
L1i cache:             32K
L2 cache:              256K
L3 cache:              8192K
NUMA node0 CPU(s):     0-7

It seems that the program has no problem. system 2 is suspicious.

EDIT Recently we have add 2 16GB RAM to system 2. I guess that openmp allocate memory on different RAM slots and communication between RAMs has made the process very slow. Is there any way that force the program to allocate memory only from one slot of RAM?

  • We can't really answer this. It depends on what your program was doing and what else was going on at the time. The number of cores is not relevant unless your program can utilize more than a single core at a time, for example. For us to help you, you need to edit your question and explain what these processes do. Do they read a lot of data, for example? Are the data on a local drive or shared over a network? – terdon Jul 3 '17 at 11:15
  • What happens if you run the program with 2 threads on both systems? If the inter-thread communication and/or locking takes more time than the computations (due to a too small problem), then running with more threads will slow things down. – Kusalananda Jul 3 '17 at 11:29
  • @Kusalananda I updated the question... – Abolfazl Jul 3 '17 at 12:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.