1 Compilers ctd

1. I installed Nvidia's compiler suite on parallel into /opt/nvidia

2. Make them available to you by sourcing /opt/nvidia/initenvs

3. These are the recommended compilers.

4. They work with Nvidia GPUs better.

5. However g++ produced faster OpenMP code for the Xeon on one example.

2 OpenACC ctd

1. https://www.openacc.org/events/openacc-online-course-2018 weeks 2 and 3.

They also introduce you to Nvidia.

3 Parallel.ecse available

You may use spare time on parallel.ecse for any legal ethical purpose, even unrelated to this course. No running businesses or trying to make money. No mining. However research and having fun are fine.

I was running boinc for https://milkyway.cs.rpi.edu/milkyway/ since 2018, and am user 359 in total credit.

4 New teaching tool

Today's new gimmick is mirroring my ipad to a window on my laptop, so I can use the ipad as a graphics tablet. Relevant SW: uxplay on the laptop; see https://rodrigoribeiro.site/2020/08/09/mirroring-ipad-iphone-screen-on-linux/ and any notetaking tool on the ipad.

Earlier I tried using a laptop with a touch screen. That didn't work well, apparently because of deficiencies with linux SW controlling the screen.

5 Applied EE

My home Tesla powerwalls (27KHW capacity) and solar panels (8KW peak) are finally working. Over the year, my net electrical consumption will be close to zero. I can survive perhaps a 2-day blackout. However the real reason for getting them is that I like gadgets.

6 Changes in this course from 2020 to 2021

1. Drop the attempt to use docker for compilers. It's, complicated, not necessary, and is a security risk.

   However, docker is an important industrial tool. If anyone would like some class time on it, just ask.

2. Use nvc++ not pgc++. nvc++ didn't exist last year.

7 Nvidia GPU and accelerated computing

This is from https://developer.nvidia.com/teaching-kits-downloads

My local copy of what I'm using is in /parclass/2021/files/nvidia/GPU-Teaching-Kit-Accelerated-Computing.zip