3.3 KiB
Connect to CryoSPARC
Once the CryoSPARC installation is complete and the instance is running, open a web browser on your local machine and navigate to:
http://service03.merlin7.psi.ch:<CRYOSPARC_BASE_PORT>
The URL is displayed at the end of a successful installation. You can also list the running CryoSPARC instances with:
cryosparc-merlin instances
The URL will open the CryoSPARC web interface, where you can log in with your CryoSPARC credentials.
Port conflicts
If a CryoSPARC instance is stopped and the ports assigned to it are subsequently occupied by another process, restarting the instance may fail.
To restart the instance, run:
cryosparc-merlin cryosparcm restart
If the restart fails because of a port conflict, you can assign a new base port with:
cryosparc-merlin changeport
This command automatically searches for available ports on the remote host and updates the CryoSPARC configuration accordingly.
SSD particle caching
For classification, refinement, and reconstruction jobs that process large particle datasets, using local SSD storage on worker nodes can improve performance. Many cryo-EM algorithms perform random data access and multiple passes over particle data, making fast local storage particularly beneficial compared with repeated access to shared storage.
When Cache particle images on SSD is enabled for a job, CryoSPARC copies the required particle data to the configured SSD cache and subsequently reads the data from the cache. If multiple jobs within the same project use the same particles, the cached data can be reused, avoiding redundant data transfers.
When the available cache space is exhausted, CryoSPARC automatically removes older cached data to make room for new data.
Configure the SSD cache path
To allow multiple jobs across worker nodes to access a shared cache, configure a shared SSD cache directory.
Add the following line to the worker configuration file:
/data/user/$USER/cryosparc/cryosparc_worker/config.sh
export CRYOSPARC_SSD_PATH=/data/scratch/shared/$USER
!!! warning "Worker Architecture" Make sure you access the correct worker installation. For installations with version 5 or above, the path to the worker installation has the architecture in the directory path.
```bash
/data/user/$USER/cryosparc/cryosparc_worker_x86_64
```
or
```bash
/data/user/$USER/cryosparc/cryosparc_worker_aarch64
```
Adjust the cache copy threads
Starting with CryoSPARC v4.3.0, multiple threads are used to copy particle data from the project directory to the SSD cache. The default is 2 threads.
To increase or decrease the number of copy threads, add the following line to:
/data/user/$USER/cryosparc/cryosparc_worker/config.sh
For example, to use 8 threads:
export CRYOSPARC_CACHE_NUM_THREADS=8
Configure the cache lifetime
CryoSPARC automatically removes cached particle data that has not been accessed for a specified period. The default cache lifetime is 30 days.
You can customize this value by setting the number of days after which unused cached data is removed.
Add the following line to the master configuration file:
/data/user/$USER/cryosparc/cryosparc_master/config.sh
For example, to retain cached data for 15 days:
export CRYOSPARC_SSD_CACHE_LIFETIME_DAYS=15