(migrated from Jason's issue NLeSC/PattyGenerate/issue#27)
Currently we use CMVS to generate a single PMVS2 config that processes all files. On a laptop this may require too much memory. For example, SITE_10 with 26 used images requires about 6GB of memory. (Note: issue #26 may also be a cause of this).
One of the goals of CMVS is too be able to split these image sets smaller in subsets that can be processed separately after which the result can be merged. Although this is originally used for very large datasets, using it to run on very small machines may also be useful.
(migrated from Jason's issue NLeSC/PattyGenerate/issue#27)
Currently we use CMVS to generate a single PMVS2 config that processes all files. On a laptop this may require too much memory. For example, SITE_10 with 26 used images requires about 6GB of memory. (Note: issue #26 may also be a cause of this).
One of the goals of CMVS is too be able to split these image sets smaller in subsets that can be processed separately after which the result can be merged. Although this is originally used for very large datasets, using it to run on very small machines may also be useful.