Difference between revisions of "PREMA Shared Memory"
From crtc.cs.odu.edu
Pthomadakis (talk | contribs) (Created page with "{| class="wikitable" |- | !colspan="3" |Time |- | #Threads || OpenMP || Argobots || TBB |- | 1 || 9.50265 || 9.50447 || 9.5061 |- | 2 || 4.75915 |...") |
Pthomadakis (talk | contribs) |
||
(5 intermediate revisions by the same user not shown) | |||
Line 1: | Line 1: | ||
+ | Details and code about the benchmarks can be found here [https://github.com/bsc-pm/bots/tree/master/omp-tasks]. | ||
+ | |||
+ | == Alignment Benchmark == | ||
+ | [[File:Aling_img.png|none|link=|800px]] | ||
+ | |||
{| class="wikitable" | {| class="wikitable" | ||
|- | |- | ||
− | | | + | |100 Sequences |
!colspan="3" |Time | !colspan="3" |Time | ||
|- | |- | ||
Line 19: | Line 24: | ||
|- | |- | ||
| 32 || 0.399665 || 0.510632 || 0.402276 | | 32 || 0.399665 || 0.510632 || 0.402276 | ||
+ | |} | ||
+ | |||
+ | == FFT Benchmark == | ||
+ | |||
+ | [[File:fft_img.png|none|link=|800px]] | ||
+ | {| class="wikitable" | ||
+ | |- | ||
+ | |Array size=1024*1024*1024 | ||
+ | !colspan="3" |Time | ||
+ | |- | ||
+ | |Threads ||OpenMP||Argobots||TBB | ||
+ | |- | ||
+ | |1||349.996||332.732||325.001 | ||
+ | |- | ||
+ | |2||230.326||180.618||173.621 | ||
+ | |- | ||
+ | |4||147.278||92.8936||87.9898 | ||
+ | |- | ||
+ | |8||205.424||50.7829||47.7683 | ||
+ | |- | ||
+ | |16||413.116||23.8058||19.3438 | ||
+ | |- | ||
+ | |24||363.685||30.5584||26.7309 | ||
+ | |- | ||
+ | |32||507.409||23.5382||17.0179 | ||
+ | |} | ||
+ | |||
+ | == Cilk Sort Benchmark == | ||
+ | |||
+ | [[File:sort_img.png|none|link=|800px]] | ||
+ | {| class="wikitable" | ||
+ | |- | ||
+ | |Array size=1024*1024*1024 | ||
+ | !colspan="3" |Time | ||
+ | |- | ||
+ | |Threads ||OpenMP||Argobots||TBB | ||
+ | |- | ||
+ | |1||105.983||106.415||106.25 | ||
+ | |- | ||
+ | |2||56.7085||53.9414||53.7241 | ||
+ | |- | ||
+ | |4||30.2197||27.4225||27.2155 | ||
+ | |- | ||
+ | |8||19.1937||15.4158||15.274 | ||
+ | |- | ||
+ | |16||28.1843||9.80334||9.6809 | ||
+ | |- | ||
+ | |24||39.7613||8.03864||7.84578 | ||
+ | |- | ||
+ | |32||55.8831||7.4676||7.19926 | ||
|} | |} |
Latest revision as of 15:19, 8 April 2019
Details and code about the benchmarks can be found here [1].
Alignment Benchmark
100 Sequences | Time | ||
---|---|---|---|
#Threads | OpenMP | Argobots | TBB |
1 | 9.50265 | 9.50447 | 9.5061 |
2 | 4.75915 | 4.76397 | 4.75832 |
4 | 2.39993 | 2.58783 | 2.39404 |
8 | 1.30484 | 1.42106 | 1.31969 |
16 | 0.74355 | 0.854036 | 0.758118 |
24 | 0.51322 | 0.621946 | 0.523768 |
32 | 0.399665 | 0.510632 | 0.402276 |
FFT Benchmark
Array size=1024*1024*1024 | Time | ||
---|---|---|---|
Threads | OpenMP | Argobots | TBB |
1 | 349.996 | 332.732 | 325.001 |
2 | 230.326 | 180.618 | 173.621 |
4 | 147.278 | 92.8936 | 87.9898 |
8 | 205.424 | 50.7829 | 47.7683 |
16 | 413.116 | 23.8058 | 19.3438 |
24 | 363.685 | 30.5584 | 26.7309 |
32 | 507.409 | 23.5382 | 17.0179 |
Cilk Sort Benchmark
Array size=1024*1024*1024 | Time | ||
---|---|---|---|
Threads | OpenMP | Argobots | TBB |
1 | 105.983 | 106.415 | 106.25 |
2 | 56.7085 | 53.9414 | 53.7241 |
4 | 30.2197 | 27.4225 | 27.2155 |
8 | 19.1937 | 15.4158 | 15.274 |
16 | 28.1843 | 9.80334 | 9.6809 |
24 | 39.7613 | 8.03864 | 7.84578 |
32 | 55.8831 | 7.4676 | 7.19926 |