-
Notifications
You must be signed in to change notification settings - Fork 13
Intel Haswell S2 M2 C14
- Processor: Intel(R) Xeon(R) CPU E5-2695 v3
- Base frequency: 2.3 GHz
- Number of sockets: 2
- Number of memory domains per socket: 2
- Memory domain specs: 2-channel DDR4-2133
- Number of cores per socket: 14
- Number of HWThreads per core: 2
- MachineState output: json
+----------+---------------------------------+
| Compiler | icc (ICC) |
|----------|---------------------------------|
| Version | icc (ICC) 19.0.5.281 20190815 |
+----------+---------------------------------+
Optimizing flags: -fast -xHost -qopt-streaming-stores=always -std=c99 -ffreestanding -qopenmp
All results are in GB/s
.
Summary results:
+---------------------------------------------+
| Single core | 21.17 (Copy) |
| Memory domain | 32.36 (Sum with 7 cores) |
| Socket | 64.41 (Sum with 7 cores) |
| Node | 128.16 (Sum with 7 cores) |
+---------------------------------------------+
Results for scaling within a memory domain:
#nt Init Sum Copy Update Triad Daxpy STriad SDaxpy
1 19.14 15.25 21.17 17.21 21.05 19.29 20.77 20.01
2 22.03 25.74 20.64 25.26 25.59 25.97 25.99 25.77
3 22.35 31.43 25.07 27.66 28.02 26.92 29.41 28.08
4 22.46 31.78 25.98 26.00 28.26 26.88 29.28 27.54
5 22.22 32.15 25.81 25.10 28.43 28.08 29.07 28.97
6 22.37 32.28 26.06 25.20 28.49 28.22 29.18 28.98
7 22.41 32.36 26.15 25.08 28.36 28.08 28.91 28.87
Results for scaling across memory domains. Shown are the results for the number of memory domains used (nm) with columns number of cores used per memory domain.
Init:
#nm 1 2 3 4
1 19.14 33.36 57.18 76.08
2 22.03 34.90 64.37 81.96
3 22.35 35.29 63.10 78.42
4 22.46 34.20 62.48 81.15
5 22.22 34.37 61.78 78.54
6 22.37 34.44 61.45 79.33
7 22.41 34.50 60.31 79.20
Sum:
#nm 1 2 3 4
1 15.25 29.44 44.71 59.54
2 25.74 49.83 74.28 96.40
3 31.43 60.00 89.45 118.17
4 31.78 63.04 94.16 125.02
5 32.15 64.06 95.66 127.02
6 32.28 64.39 96.40 127.83
7 32.36 64.41 96.29 128.16
Copy
#nm 1 2 3 4
1 21.17 40.67 59.68 79.50
2 20.64 47.57 63.21 80.07
3 25.07 48.07 76.21 99.92
4 25.98 50.35 75.91 101.44
5 25.81 50.39 76.19 99.73
6 26.06 50.03 75.55 98.62
7 26.15 50.00 74.08 97.31
Update
#nm 1 2 3 4
1 17.21 33.20 49.85 66.58
2 25.26 48.99 73.10 98.10
3 27.66 51.34 75.66 98.72
4 26.00 50.32 74.31 99.71
5 25.10 50.63 74.39 96.78
6 25.20 50.36 74.57 98.09
7 25.08 50.12 70.06 94.07
Triad
#nm 1 2 3 4
1 21.05 41.22 61.06 81.34
2 25.59 50.43 77.22 102.06
3 28.02 55.33 82.43 109.78
4 28.26 55.50 83.24 111.28
5 28.43 55.76 83.36 109.63
6 28.49 55.26 83.07 108.99
7 28.36 54.93 80.92 108.01
Memory bandwidth scaling within one memory domain:
The following plots illustrate the the performance scaling over multiple memory domains using different number of cores per memory domain.
Memory bandwidth scaling across memory domains for init:
Memory bandwidth scaling across memory domains for sum
Memory bandwidth scaling across memory domains for copy
Memory bandwidth scaling across memory domains for Triad