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
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
|
# vcflib performance
## vcfwave
The early edition of WFA2-lib did not automatically compile for parallel. In the 'build' dir:
```
/usr/bin/time -v ./vcfwave ../samples/grch38#chr8_36353854-36453166-bcftools-normalised.vcf > /dev/null
Percent of CPU this job got: 99%
Elapsed (wall clock) time (h:mm:ss or m:ss): 0:09.18
```
```
make BUILD_WFA_PARALLEL=1 BUILD_TOOLS=0 BUILD_TOOLS=0 BUILD_EXAMPLES=0 CC=gcc CC_FLAGS=-fPIC setup asan lib_wfa
```
I made sure to build vcflib with OPENMP=ON and libgomp.so.1 is linked. Also WFA builds with openmp:
```
CMakeFiles/wfa-EXT.dir/build.make: cd /export/local/home/wrk/iwrk/opensource/code/pangenome/vcflib/contrib/WFA2-lib && $(MAKE) clean BUILD_WFA_PARALLEL=1 BUILD_TOOLS=0 BUILD_EXAMPLES=0 CC=gcc CC_FLAGS=-fPIC setup lib_wfa
```
This happens when running `cmake -DOPENMP=ON ..`, but, still, no multi-core `vcfwave`!
When setting
```C
wavefront_aligner_set_max_num_threads(wf_aligner,4);
```
We got a speedup:
Percent of CPU this job got: 242%
Elapsed (wall clock) time (h:mm:ss or m:ss): 0:08.24
### Big files
This is a 30Mb file after compression:
```
tux02:~/tmp/vcflib/build$ /usr/bin/time -v ./vcfwave ../samples/chr18.grch38.vcfbub.a100k.vcf.gz > test.vcf
Percent of CPU this job got: 99%
Elapsed (wall clock) time (h:mm:ss or m:ss): 1:17:16
```
Trying 16 threads
```
tux02:~/tmp/vcflib/build$ /usr/bin/time -v ./vcfwave ../samples/chr18.grch38.vcfbub.a100k.vcf.gz -t 16 > test16.vcf
Percent of CPU this job got: 464%
Elapsed (wall clock) time (h:mm:ss or m:ss): 1:16:30
```
hmm. Not so exciting. Looks like we need to parallelize at a different level. Memory used was about 6Gb, and that is not bad.
The good news is that the output is identical.
## vcfcreatemulti
This program is actually slower than vcfwave.
|
