I just noticed that we should contact you via github, because i recently sent an E-Mail.
Maybe i have been doing something wrong, but i could not find a way to reproduce the results as stated in the paper, only slower results... The compression ratio is high, just the speed does not quit match.. The quantizer code doesn't run on the current python so i made some modifications, e.g. replace deprecated method calls to the matching new ones etc.
Other than that im using the provided lzbench via python, so the results generated are plotted nicely.
..I would like to ask you, if you can provide code and datasets (of the success cases) already quantized and ready for use? so I can just run it and reproduce the results. That would be very helpful.
I am a research assistant at the University of Marburg - Germany, and I have been working on compression for our time-series database system ChronicleDB.
I look forward to hearing from you!
I just noticed that we should contact you via github, because i recently sent an E-Mail.
Maybe i have been doing something wrong, but i could not find a way to reproduce the results as stated in the paper, only slower results... The compression ratio is high, just the speed does not quit match.. The quantizer code doesn't run on the current python so i made some modifications, e.g. replace deprecated method calls to the matching new ones etc.
Other than that im using the provided lzbench via python, so the results generated are plotted nicely.
..I would like to ask you, if you can provide code and datasets (of the success cases) already quantized and ready for use? so I can just run it and reproduce the results. That would be very helpful.
I am a research assistant at the University of Marburg - Germany, and I have been working on compression for our time-series database system ChronicleDB.
I look forward to hearing from you!