Hi,
I am a researcher, and I want to try Nvidia Parabricks toolkit on my system. I have previously executed GATK4 on my system, but now want to try it using GPU. Kindly help me to start with it. System configuration is:
I do not want to run full scale Genome sequence right now. For the time being, I need to familiarise myself with the basics. So, help will be highly appreciated.
Thanks in advance.
Unfortunately you’re out of luck with a GTX 1660 Ti, or in fact any GeForce card at this stage. You only really have two options as far as I know:
either buy compatible GPUs - so Tesla P100, V100, T4, or I’m assuming the A100 now as well.
using GPU enabled cloud images (i.e. via AWS)
I’m hoping we’ll see Nvidia provide the option for non-Tesla/A100 grade GPUs to be used in the future, especially since the 3000 series that is selling now (or sold out everywhere :-) ) is such a leap in ‘power’. We have a couple of Tesla V100 cards which we use for Parabricks and Nanopore base calling, and I’m very interested to see that the 3080/3090 actually have more CUDA cores than our super expensive server grade cards.
Hey miles.benton84,
Thanks for your reply. I can understand that it is not possible to execute Parabricks on my native machine. If you could help, then could you please tell me if it is possible to execute the pipeline on Google Colab Pro, as they also provide high end NVIDIA GPUs and in a much affordable cost.
Thanks