Help with performance improvements: allocating resources for tools

Hello,
Can anyone of you help me config my galaxy server. I gave my server 16 virtual cores in HyperV and found some informations that i have to change some params in galaxy.yml file(galaxy/config). Whats the best way to do my galaxy faster. Now i’ve got
buffer-size: 16384
processes:3
threads:15
offload-threads:2

Thanks in advance
Maciej

Hi @Maciej,

what do you mean by “faster”? Is the web interface slow? Or are the tools that you execute only using one core? Galaxy on its own is just a web server, that is handling API requests, if you don’t have 100s simultaneous users, the web server should not need 16 cores. However, the tools that Galaxy will execute can use many cores and a lot of memory. This depends obviously on the tool you are running and you can give every single tool its own resources.

We have covered all those aspects in our training material here: Galaxy Training!

1 Like

Hi @bjoern.gruening,
Thank you for respond.

I just want to improve performance for my lab members(it will be maximally 30 users, but maybe 10 at the same time). I have a virtual machine on Hyper V(16 cores and 32 GB’s of RAM) and I just want to make calculations faster for each user. So the question is how to boost performance for calculations ? Do I need to add more cores or ram for my VM? Do I need to change some config files for it ?

Thanks in advance
Maciej

Yes, in that case, you need to change the job_conf.xml file to give specific resources to your tools. Like BWA should get 6 cores and 16GB of memory for example.

1 Like