Benchmark functions that returns total space, mem, cpu given input size and parameters for the CWL workflows
Project description
The repo contains a benchmarking script for some of the CWL workflows used by 4DN-DCIC (https://github.com/4dn-dcic/pipelines-cwl), that returns total space, mem and CPUs required per given input size and a recommended AWS EC2 instance type.
Example usage of benchmarking script
- importing the module
from Benchmark import run as B
- md5
app_name = 'md5'
input_json = {'input_size_in_bytes': {'input_file': 20000}}
B.benchmark(app_name, input_json)
{'aws': {'recommended_instance_type': 't2.xlarge', 'EBS_optimized': False, 'cost_in_usd': 0.188, 'EBS_optimization_surcharge': None, 'mem_in_gb': 16.0, 'cpu': 4}, 'total_size_in_GB': 14.855186462402344, 'total_mem_in_MB': 13142.84375, 'min_CPU': 4}
- fastqc-0-11-4-1
app_name = 'fastqc-0-11-4-1'
input_json = {'input_size_in_bytes': {'input_fastq':20000},
'parameters': {'threads': 2}}
B.benchmark(app_name, input_json)
{'recommended_instance_type': 't2.nano', 'EBS_optimized': False, 'cost_in_usd': 0.006, 'EBS_optimization_surcharge': None, 'mem_in_gb': 0.5, 'cpu': 1}
- bwa-mem
app_name = 'bwa-mem'
input_json = {'input_size_in_bytes': {'fastq1':93520000,
'fastq2':97604000,
'bwa_index':3364568000},
'parameters': {'nThreads': 4}}
B.benchmark(app_name, input_json)
{'aws': {'cost_in_usd': 0.188, 'EBS_optimization_surcharge': None, 'EBS_optimized': False, 'cpu': 4, 'mem_in_gb': 16.0, 'recommended_instance_type': 't2.xlarge'}, 'total_mem_in_MB': 12834.808349609375, 'total_size_in_GB': 15.502477258443832, 'min_CPU': 4}
To use Benchmark in from other places, install it as below.
pip install Benchmark-4dn
or
pip install git+git://github.com/SooLee/Benchmark.git
Note: From 0.5.3
we have a new function that takes in cpu and memory and returns a sorted list of instance dictionaries.
get_instance_types(cpu=1, mem_in_gb=0.5, instances=instance_list(), top=10, rank='cost_in_usd')
Keys in each instance dictionary:
'cost_in_usd', 'mem_in_gb', 'cpu', 'instance_type', 'EBS_optimized', 'EBS_optimization_surcharge'
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Benchmark-4dn-0.5.5.tar.gz
(23.6 kB
view hashes)
Built Distributions
Close
Hashes for Benchmark_4dn-0.5.5-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 704defeacb61ed9a4dc63b38f36c58b08a1dd6eae47e08ace631cfecd4925d2b |
|
MD5 | b9994f1aa455cb1db659c4fc81652422 |
|
BLAKE2b-256 | 4376395e7b6cba7eb24e64beef61a905dce49c4d65eda3f6574a5338479a95ec |
Close
Hashes for Benchmark_4dn-0.5.5-py2-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | b37b5eee7816bd39636e60bd333b7dea8fdfef3f8991604faeb32e71eae1a2ee |
|
MD5 | 210c18e185607a87b56650966d0f12e3 |
|
BLAKE2b-256 | 9f02ab7d23d6aca1e6ac9b7941247171c24d5404b0c0e0d03ec4747eba4d41f1 |