Differences
This shows you the differences between two versions of the page.
| Both sides previous revision Previous revision Next revision | Previous revision | ||
| dragnet:benchmarks_of_the_lotaas_pipelines [2016-11-01 19:37] – [Dedispersion strategies for Cartesius] Sotiris Sanidas | dragnet:benchmarks_of_the_lotaas_pipelines [2017-03-08 15:27] (current) – external edit 127.0.0.1 | ||
|---|---|---|---|
| Line 105: | Line 105: | ||
| From the above, we can conclude that using 1 or 2 disks does not make a big difference. Also, hyperthreading works smoothly, and indeed the best strategy is to have the maximum possible number of rfifind instances running in parallel. | From the above, we can conclude that using 1 or 2 disks does not make a big difference. Also, hyperthreading works smoothly, and indeed the best strategy is to have the maximum possible number of rfifind instances running in parallel. | ||
| - | ==== Dedispersion strategies for Cartesius ==== | + | ==== Cartesius |
| - | The max number of timeseries (.dat files) that can be stored in 32GB of /dev/shm.\\ | + | Processing 1 full pointing on cartesius using either |
| - | The dedispersion plan is split at two points, the first at DM~45 without downsampling, | + | |
| - | Part1: 480<file number< | + | {{dragnet:benchmarks:cartesius_bm1.png? |
| - | Part2: 984<file number< | + | |
| - | Part3: 1944< | + | |