About
Modern computing architectures allow for unprecedented levels of parallelization, bringing a much-needed speedup to key scientific applications, such as ever improving numerical simulations and their post-processing, likewise increasingly taxing. We report on optimization techniques used on popular codes for computational astrophysics (FLASH and ECHO) and the performance gained on second-generation Intel Xeon Phi and Xeon Scalable Processors (code-named Knights Landing and Skylake, respectively). We also show how simulation post-processing can largely benefit from HPC methods. We focus specifically on yt (an open source Python package for data analysis and visualization), in which speedups as high as to 4x or 8x with respect to the code baseline can be easily achieved just through the use of cython and the Intel Distribution for Python..
0:00:00 Slide 1
0:00:12 Slide 2
0:00:16 Slide 3
0:00:20 Slide 4
0:01:39 Slide 5
0:01:49 Slide 6
0:02:06 Slide 7
0:03:20 Slide 8
0:05:00 Slide 9
0:05:26 Slide 10
0:06:11 Slide 11
The talk
Credits and licence
- Institution
- CERN
- Department
- other events or meetings
- Speaker
- Salvatore Cielo
- Year
- 2019
- Language
- English
- Subjects
- Mathematics
- Source
- CERN
- Licence
- CERN, as CERN states it. CERN's terms
- Attribution
- © CERN. CERN provides its audiovisual media free of charge for educational and informational use, credited to CERN; this lecture is shown in CERN's own player.
- Who can watch
- CERN gives no rating. Invisible Network shows it to Tweens profiles and older, judged from what the source says about it. How courses are placed