Francesc Alted
Francesc Alted
I concur that this is a nice goal to do. Thanks for you AVX2-unshuffle for 12-bytes. Other contributions for different type sizes are welcome indeed.
That would be great, yes. Would you mind contributing a pull request?
If you want to accelerate HDF5 I/O with Blosc2, the direct chunking approach is best as can be seen in e.g. https://www.blosc.org/posts/blosc2-pytables-perf/. BTW, we have made a package that allows...
FWIW, here there is info on the expected Blosc2 info for a plugin in HDF5: https://github.com/Blosc/HDF5-Blosc2/blob/main/src/blosc2_plugin.c
Well, I don't remember the details, but I think blosc2/b2h5py plugin accepts both single chunks and super-chunks too. @ivilata may bring some light here.
The plugin for blosc1 is around here: https://github.com/Blosc/hdf5-blosc. For what I can see in your example, you try to use blosc1 from hdf5plugin, and then use b2h5py that is using...
Interesting. You may want to try setting the `BLOSC_TRACE` environment variable to something. This usually provides some debugging info on errors.
Thanks! Would you like to contribute a PR?
Yes, I can reproduce this. If you can find the root of the issue, shout!
That looks interesting. Would you want to contribute a PR?