Alf, My initial thoughts on the issues you raise are:
1. (Post-processor compatibility.) A new format could be implemented in both OPM Flow and Resinsight, making it useful rather quickly. 2. (Write only at end of run.) I assume this can be resolved by wasting a little space: initially allocating "enough" space in the file (to fit expected number of steps) and re-allocating/re-writing only if required (i.e. expected number of steps too small). 3. (Not suited for parallel.) That is a bummer. But on the other hand, I do not see how reading summary-type (time series) output could be a bottleneck for a large case, compared to the time-evolution data for pressure, saturation etc. in the restart file? Atgeirr ________________________________ Frå: Opm <[email protected]> på vegne av Alf Birger Rustad <[email protected]> Sendt: måndag 18. mai 2020 17:51 Til: '[email protected]' <[email protected]> Emne: [Opm] File formats Dear community, We are at a cross roads with respect to file formats, and I hope you are motivated to help us arrive at the best solution. We need better load-on-demand performance for summary files than what is currently possible with the default Eclipse format for summary files. Currently you will find an implementation in opm-common that simply transposes the summary vectors, while still using the same Fortran77 binary format. That approach has mainly three drawbacks. One is that it is not supported by any post-processing application (yet). The second is that it can only be created from a finished simulation, so you need to wait for simulations to finish before you get the performant result file. The third being that it is not suited for parallel processing, so forget about each process writing out it's part. We have started looking at HDF5 to see if it fits all the needs, as we have seen that in use for reservoir simulation already. We see that there are alternatives to HDF5, so any experiences on that is highly appreciated. Any and all input is highly appreciated at this point in time. What do you think? Alf _______________________________________________ Opm mailing list [email protected] https://opm-project.org/cgi-bin/mailman/listinfo/opm
