# Next generation Intel MPI product for next generation systems. The latest Intel MPI features and optimization techniques

- Institution: CERN
- Department: other events or meetings
- Speaker: Klaus-Dieter Oertel
- Year: 2019
- Language: English
- Source: [CERN](https://videos.cern.ch/record/3002573)
- Licence: [CERN](http://copyright.web.cern.ch)
- Who can watch: Teens profiles and older (sources give no ratings; placed by Invisible Network from the source's own words)

Main goal of the presentation/tutorial is to provide audience an information about key aspects of new generation of Intel MPI product and the way the library may help to HPC/ML workloads.

You will learn about: - Main difference between old and new generations of the product and rationale behind the changes - New unique features for multithreading like multiple endpoint support and new design of asynchronous progress engine - New MPI library tuning approach and the way it may help to your HPC/ML applications - The latest features for intranode communication optimization.

0:00:00 Slide 1

0:00:36 Slide 2

0:00:44 Slide 3

0:01:33 Slide 4

0:02:23 Slide 5

0:04:48 Slide 6

0:06:34 Slide 7

0:07:16 Slide 8

0:08:20 Slide 9

0:10:10 Slide 10

0:10:21 Slide 11

0:11:30 Slide 12

0:11:34 Slide 13

0:13:08 Slide 14

0:15:17 Slide 15

0:19:00 Slide 16

0:19:28 Slide 17

0:20:18 Slide 18

0:21:34 Slide 19

0:22:17 Slide 20

0:22:40 Slide 21

0:22:49 Slide 22

0:23:08 Slide 23

0:23:31 Slide 24

0:28:40 Slide 25

0:30:40 Slide 26

0:31:24 Slide 27

0:32:50 Slide 28

0:32:55 Slide 29

0:33:04 Slide 30

0:34:26 Slide 31

0:34:31 Slide 32

0:35:03 Slide 33

0:35:32 Slide 34

0:36:27 Slide 35

0:37:13 Slide 36

0:38:06 Slide 37

0:39:10 Slide 38

0:39:34 Slide 39

0:39:50 Slide 40

## Lectures

1. [Next generation Intel MPI product for next generation systems. The latest Intel MPI features and optimization techniques](https://invisiblenetwork.ai/learn/cern/3002573/cab0ede68106), 41 min (CERN's player)