Open orioltinto opened 3 months ago
Created a new figure for the paper which computes the throughput with the different compressors. However, it looks like I found a bug in SZ3 which makes things super-slow with the ERA5 data used for figure 2. Must investigate that.
Made the small benchmark, both in a single process and a scalability experiment. Should send the reproducible case to SZ3 asking if the slow-down is normal.
The reviewer asked information about the compressors speed. Prepare a benchmark to test it.