From subsecret
Jump to: navigation, search
No edit summary
No edit summary
Line 6: Line 6:
**Compile BWA with command: ./compile.sh
**Compile BWA with command: ./compile.sh


*Retrieve ftp://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/000/005/845/GCF_000005845.2_ASM584v2/GCF_000005845.2_ASM584v2_genomic.fna.gz
*Retrieve https://files.subsecret.dk/benchmark/GCF_000005845.2_ASM584v2_genomic.fna.gz
*Retrieve https://s3.amazonaws.com/public.ged.msu.edu/ecoli_ref-5m-trim.fastq.gz
*Retrieve https://files.subsecret.dk/benchmark/ecolireads.fastq.gz
*Run ./bwa index GCF_000005845.2_ASM584v2_genomic.fna.gz
*Run ./bwa index GCF_000005845.2_ASM584v2_genomic.fna.gz


Line 13: Line 13:
Benchmark command
Benchmark command
*1 thread
*1 thread
**./bwa aln GCF_000005845.2_ASM584v2_genomic.fna.gz ecoli_ref-5m-trim.fastq.gz > /dev/null
**./bwa aln GCF_000005845.2_ASM584v2_genomic.fna.gz ecolireads.fastq.gz > /dev/null
*4 threads  
*4 threads  
**./bwa aln -t4 GCF_000005845.2_ASM584v2_genomic.fna.gz ecoli_ref-5m-trim.fastq.gz > /dev/null
**./bwa aln -t4 GCF_000005845.2_ASM584v2_genomic.fna.gz ecolireads.fastq.gz > /dev/null
 


==Results==
==Results==
{| class="wikitable sortable"
!Device
!ARCH
!Kernel
!OS
!GCC
!1 thread walltime (sec.)
!4 thread walltime (sec.)
!8 thread walltime (sec.)
!16 thread walltime (sec.)
|-
|
|
|
|
|
|
|
|
|
|}


==Results (old data)==
These benchmarks are generated with a now unavailable dataset.
{| class="wikitable sortable"
{| class="wikitable sortable"
!Device
!Device

Revision as of 11:36, 10 September 2018

BWA Benchmark

BWA is a tool for mapping reads against a genome. This is nice for doing a non-synthetic benchmark, as meaningful work is done, and IO, Memory and CPU is all used.

BWA Setup instructions

Running Benchmark

Benchmark command

  • 1 thread
    • ./bwa aln GCF_000005845.2_ASM584v2_genomic.fna.gz ecolireads.fastq.gz > /dev/null
  • 4 threads
    • ./bwa aln -t4 GCF_000005845.2_ASM584v2_genomic.fna.gz ecolireads.fastq.gz > /dev/null

Results

Device ARCH Kernel OS GCC 1 thread walltime (sec.) 4 thread walltime (sec.) 8 thread walltime (sec.) 16 thread walltime (sec.)

Results (old data)

These benchmarks are generated with a now unavailable dataset.

Device ARCH Kernel OS GCC 1 thread walltime (sec.) 4 thread walltime (sec.) 8 thread walltime (sec.) 16 thread walltime (sec.)
PC (i5-750@2.67GHz) x64 4.4.0-53-generic x86_64 Ubuntu 16.04 5.4.0 162.644 54.980
PC (R7 1700@3.8GHz) x64 4.4.0-75-generic x86_64 Ubuntu 16.04 97.116 32.375 21.908 18.246
PC (Xeon Gold 6148@2.4GHz) x64 3.10.0-514.16.1.el7.x86_64 CentOS 7 6.3.1 102.149 34.774 22.088 16.100
PC (i7-6700k@4.5 GHz) x86 4.6.4-301.fc24.x86_64 Fedora 24 6.1.1 144.221 53.833
PC (i7-6700k@4.5 GHz) x64 4.6.4-301.fc24.x86_64 Fedora 24 6.1.1 77.836 27.840
PC (i7-980x@4.0 GHz) x64 4.6.4-301.fc24.x86_64 Fedora 24 6.1.1 130.605 45.549
ODROID U2 arm7l 3.8.13.30 Ubuntu 14.04.4 LTS 4.8.4 939.300 266.262
ODROID C2 arm64 3.14.65-73 Ubuntu 16.04 LTS 5.4.0 797.158 235.510
Raspberry Pi 3 armv7l 4.4.13-v7+ Raspbian Jessie 4.9.2 1224.290 355.897
ASUS Tinker Board armv7l 4.4.16 Debian 4.9.2-10 4.9.2 786.833 239.936
Raspberry Pi 3 arm64 4.9.26-v8 Raspbian Jessie (Debian 64bit chroot) 6.3.0 924.110 297.149
Orange PI one armv7l 3.4.112 Armbian Jessie 4.9.2 1677.26 486.256
Raspberry Pi zero armv6l 4.4.13+ Raspbian Jessie 4.9.2 2801.69 2812.427
Raspberry Pi 1 armv6l 4.4.13+ Raspbian Jessie 4.9.2 4031.349 4072.279
Raspberry Pi 2 armv7l 4.4.13-v7+ Raspbian Jessie 4.9.2 2009.622 592.694
Retrieved from "https://subsecret.dk/index.php?title=BWA_Benchmark_(ODROID_C2,RPI3)&oldid=525"