Parallella: Supercomputer for Embedded Systems

Parallella is about the size of a Raspberry Pi, but with 18 cores instead of two. This presentation will show how to get up and running with Parallella, and why developers might care about supercomputing. Why does parallel computing matter? How can developers use parallel computing to deliver better results for clients? Let’s explore these questions together.

Details: http://rayhightower.com/blog/2015/12/09/chicago-acm-and-parallella/

Curious about the Internet of Things? Check out http://WindyCityThings.com.

Transcript

1. Chicago ACM Chicago, IL, USA December 9, 2015 Parallella: Supercomputer for Embedded Systems
2. Moore’s Law: 2x every 18 months
3. Moore’s Law: 2x every 18 months Moore’s Law 1993 - 2013 http://www.washingtonpost.com/blogs/innovations/wp/2015/04/14/10-images-that-explain-the- incredible-power-of-moores-law/ https://www.parallella.org/board/
5. RayHightower.com
6. IBM Blue Gene https://commons.wikimedia.org Cray X-MP
7. RJ-45 Power μUSB μHDMI μSD
8. RISC ARM 18 cores: 2 ARM + 16 RISC
12. Watts => Dollars.
14. 17.8 million watts $17.8 million per year
15. 5 watts for Parallella?
17. http://rayhightower.com/blog/2014/09/09/solar-powered-parallella/
18. http://rayhightower.com/blog/2014/09/09/solar-powered-parallella/
19. http://rayhightower.com/blog/2014/09/09/solar-powered-parallella/ 5 volts 1 amp 5 watts Solar!
20. If one ox could not do the job they did not try to grow a bigger ox, but used two oxen. When we need greater computer power, the answer is not to get a bigger computer, but to build systems of computers and operate them in parallel. -Grace Hopper
21. Concurrency is not Parallelism. https://www.youtube.com/watch?v=cN_DpYBzKso&list=PLOnWKC1gI_OPU8SDIBnCLHsgzNLSbnPJQ&index=3 -Rob Pike, Go
22. Reduced Instruction Set Computer RISC
23. ARM A9
24. RISC ARM 18 cores: 2 ARM + 16 RISC
26. Hello Epiphany
27. host.c epiphany.c
29. Find all primes up to 16,000,000. Serial on Parallella.
30. 1 #include <stdio.h> 2 #include <stdlib.h> 3 #include <math.h> 4 #include <time.h> 5 6 #define DEFAULT_MAX_TESTS 16000000 7 8 inline int isprime(unsigned long number) 9 { 10 unsigned long i; 11 unsigned long s = sqrt(number); 12 for(i=3;i<=s;i+=2) 13 { 14 if(number % i == 0) 15 return 0; 16 } 17 return 1; 18 } /* Copyright (c) Adapteva, contributed by M. Thompson with modifications by T. Malthouse. */
31. 8 inline int isprime(unsigned long number) 9 { 10 unsigned long i; 11 unsigned long s = sqrt(number); 12 for(i=3;i<=s;i+=2) 13 { 14 if(number % i == 0) 15 return 0; 16 } 17 return 1; 18 } /* Copyright (c) Adapteva, contributed by M. Thompson with modifications by T. Malthouse. */
33. 237.1 sec
34. Find all primes up to 16,000,000. Serial on Mac OS X.
35. Same serial code, written in C. Build it on OS X.
37. 14.4 sec
38. Find all primes up to 16,000,000. Parallel on Parallella.
39. 27 #include <e-hal.h> 28 29 // Default max number of prime tests per core 30 // Used if a limit it not provided in argv[1] 31 #define DEFAULT_MAX_TESTS 500000 32 33 int main(int argc, char *argv[]) 34 { 35 unsigned row, col, coreid, i, j; 36 e_platform_t platform; 37 e_epiphany_t dev; 38 /* Copyright (c) Adapteva, contributed by M. Thompson with modifications by T. Malthouse.*/
41. 18.6 sec
42. Summary: Finding Primes 0" 50" 100" 150" 200" 250" Parallel Parallella Serial Mac Serial Parallella 18.6 sec 14.4 sec 237.1 sec ($2,000.00 Apple MacBook Pro) ($150.00 Parallella)
43. Embarrassingly parallel problem.
44. Mandelbrot Set
46. Why?
47. Finite Element Analysis http://www.ce.ncsu.edu/news/article/21550/making-bridges-more-robust-to-earthquakes/
48. break? wear out? work? Finite Element Analysis Will it {
49. Free Body Diagram F F applied friction F gravity F normal
50. Finite Element Analysis http://www.ce.ncsu.edu/news/article/21550/making-bridges-more-robust-to-earthquakes/
51. https://www.e-education.psu.edu/worldofweather/node/2029 Grid spacing influences accuracy. Weather Prediction
52. Barriers to Entry
53. Barriers to Entry Removing
54. $ ssh $ Xvfb rayhightower.com/blog/2015/12/08/how-to-run-headless-parallella/
55. Parallella vs GPU
56. 384 cores MacBook Pro Video
57. Specialized vs. General
58. Embedded Systems
61. https://www.parallella.org/2015/06/01/the-open-camera-project-1000-bounty-for-open-firmwaredrivers-for-raspberry-pi-camera-module/
63. Thanks! RayHightower.com