Hi Chris,
Further to this, our group is using camelot 1.5.8, running with Linux (Ubuntu 18) on a server (multiple people access the database remotely using a VPN and their web browser). I have Java64 bit installed, have allocated 18GB RAM to java process (the server has 64 cores and 128 GB RAM in total --- it's a high-performance machine).
I just checked the latest camelot speed (there were no other people using it at this time):
- From Survey page to Library load of first 150 images (160K images in database) = 115 seconds
- To load a new page of 150 images = 110 seconds
- Search for specific terms (e.g. isolate to a species) = 115 seconds
- To load a full individual image after clicking on thumbnail = <1 sec (i.e. this part is OK)
- Return to Organisation page, or doing anything other than accessing the library = instantaneous
I am current using the "Other" (i.e., server version), not Linux, but this only seems to be 32 bit? Should I be using the Linux version instead? Multiple people access the database remotely.
I note that in your documentation, you say:
"To serve as a guide for how Camelot may perform for large datasets, below is the timing for working with a dataset of 2 million images using a high-end laptop in 2017 (Dell Precision 5510; 16GB RAM; Intel i7-6820HQ; SSD) with a maximum JVM heap size of 14GB:
Full Export report: 368 seconds
Library load time: 13 seconds
Library search time (basic search): 6.5 seconds
Library search time (full-text search): 28 seconds"
I'm clearly not getting that sort of performance! Any advice?
cheers,
Barry