cryoSPARC Installation- Is this configuration okay?

Hi!

If possible I would like to have a feedback regarding the architecture of my workstation and how I have done the CryoSPARC installation and how to arrange the project directories. The characteristics of the workstation:

  • 4xGPU Nvidia RTX-2880Ti (11GB)
  • 16x32GB DDR4 2933mHz ECC REG
  • 1xSSD Samsung 970 PRO 1TB M.2 NVMe 2280 PCIe 3.0 x4 (boot Centos7)
    Sequential Read: 3500 MB/s, Sequential Write: 2700 MB/s,
    Random Read (4KB): 500000 IOPS, Random Write (4KB): 500000 IOPS
  • 1xSSD Intel D3 S4610 Series 3.8TB 6Gb/s 2,5 7mm TLC (scratch)
    Sequential Read (up to) 560 MB/s Sequential Write (up to) 510 MB/s
    Random Read (100% Span) 96000 IOPS Random Write (100% Span) 42000 IOPS
  • 3xHD 14TB, SATA 6 Gb/s 7200rpm 3,5 256MB

I have installed the master and the worker in the 3.8TB SSD (/ssd/cryosparc) which I also use as ssd scratch. I usually have the raw cryoEM movies that I am processing in the 14TB HDD and planning to place the directory of the running project in the 3.8TB SSD (/ssd/cryosparc/projects), and then move it to the HDD once I am done with the processing.

Do you think this configuration is ok, or may I change something?

Thank you very much in advance.

Best,
Igor

Hi @igor,

This is definitely a great setup- installing cryoSPARC on the SSD will ensure the web interface is snappy, and having the project be located on the same directory will also make jobs run even faster. Just make sure to specify “SSD caching” to be disabled by default since your particle image files will already be on an SSD. For more info, see this page: https://cryosparc.com/docs/tutorials/ssd-caching#use

Also, good idea to offload your project directory once it’s done, as project directories can quickly fill up in later stages of processing. You have two options for offloading projects. One is to “export” the project (compress it) and move it over to the HDD (the project will not be able to be used in this state, and you’d have to restore the project directory if you want to continue processing data that is inside the project). See this tutorial for more information: https://cryosparc.com/docs/tutorials/data-management#use-case-archive

Your other option is to modify the cryoSPARC database to tell cryoSPARC that the project directory is now located in a different location. This method would allow you to still access the project and continue processing. To do this, see this guide: https://cryosparc.com/docs/tutorials/migration#use-case-b

Good luck!

Hi Stephan,

thanks a lot, I very much appreciate your feedback!

Best,
Igor