Matthias Petri

Results 21 issues of Matthias Petri

Added Intersection of k ranges over T[0..n]. based on "New Algorithms on Wavelet Trees and Applications to Information Retrieval" by Travis Gagie, Gonzalo Navarro, Simon J. Puglisi. works for both...

Hey, in my local fork of libcds I have some more features which I'm not sure are of interest as libcds2 is currently in development. The features are: - 64bit...

Could you elaborate on the space requirements on GPU. I'm trying to run a 2.9Mx300 with K=32000 instance on a 16GB RAM P100 GPU and I'm getting out of memory...

**Describe the bug** The docs at https://pisa.readthedocs.io/en/latest/compress_index.html#usage talk about using `create_freq_index` which is what the binary was called in `ds2i`. It looks like it was renamed to `compress_inverted_index` and additional...

bug
help wanted
documentation

**Describe the solution you'd like** The parse_collection command seems to have several potential areas of improvements: 1. "Remapping IDs" could be done in parallel 2. "Concatenating batches" and "Remapping IDs"...

enhancement

**Describe the solution you'd like** The bitfunnel paper contained some additions to the `partitioned_elias_fano` codebase which allowed measuring queries per second with multiple threads. see here: `https://github.com/BitFunnel/partitioned_elias_fano/blob/master/Runner/QueryLogRunner.cpp#L56` It would be...

enhancement

**Describe the bug** Ingesting plaintext records via stdin seems very slow (`12`MiB/sec) even though `60` worker threads are used. I suspect this has to do with https://en.cppreference.com/w/cpp/io/ios_base/sync_with_stdio Before I disabled...

bug
enhancement
performance
priority:high

Should be possible by 1. Performing interval_symbols on the class WT 2. For each of the classes perform interval_symbols on the offset WTs

enhancement

This pull request adds functionality to the int-vector-mapper (#232) such as [x] typedef for a `read_only_mapper` [x] support class which allows the creation of output buffer: `write_out_buffer` [x] support class...

The `csa_sada` should be decoupled form the underlying PSI encoding. This would allow `csa_sada` to be accessed from multiple threads.