Difference between revisions of "CSC352 Mini Lab 1 2013"
(Created page with "--~~~~ ---- <bluebox> This mini lab is intended to get you to run serial and multithreaded applications on the various platforms available to us. You are welcome to work in p...") |
|||
Line 3: | Line 3: | ||
<bluebox> | <bluebox> | ||
This mini lab is intended to get you to run serial and multithreaded applications on the various platforms available to us. You are welcome to work in pairs. No lab report needed for this lab. | This mini lab is intended to get you to run serial and multithreaded applications on the various platforms available to us. You are welcome to work in pairs. No lab report needed for this lab. | ||
− | + | </bluebox> | |
+ | <br /> | ||
=PiSerial and PiParallel1= | =PiSerial and PiParallel1= | ||
==Question 1: Getting a sense of speed of execution== | ==Question 1: Getting a sense of speed of execution== | ||
− | + | <br /> | |
Copy/Paste the java programs from the [[Computing_Pi| Computing Pi]] page and time their execution on different machines | Copy/Paste the java programs from the [[Computing_Pi| Computing Pi]] page and time their execution on different machines | ||
* beowulf | * beowulf | ||
Line 14: | Line 15: | ||
* the Fedora machines in FH345 | * the Fedora machines in FH345 | ||
* the MacMinis in FH345 | * the MacMinis in FH345 | ||
− | + | <br /> | |
==Question 2: Better Performance?== | ==Question 2: Better Performance?== | ||
− | + | <br /> | |
Figure out a way to ''potentially'' increase the parallelism of your application, modify the program if necessary, and test whether your approach can improve the performance of the computation of Pi (approximation, really). | Figure out a way to ''potentially'' increase the parallelism of your application, modify the program if necessary, and test whether your approach can improve the performance of the computation of Pi (approximation, really). | ||
Latest revision as of 11:05, 3 September 2013
--D. Thiebaut (talk) 12:05, 3 September 2013 (EDT)
This mini lab is intended to get you to run serial and multithreaded applications on the various platforms available to us. You are welcome to work in pairs. No lab report needed for this lab.
PiSerial and PiParallel1
Question 1: Getting a sense of speed of execution
Copy/Paste the java programs from the Computing Pi page and time their execution on different machines
- beowulf
- beowulf2
- your laptop
- the Fedora machines in FH345
- the MacMinis in FH345
Question 2: Better Performance?
Figure out a way to potentially increase the parallelism of your application, modify the program if necessary, and test whether your approach can improve the performance of the computation of Pi (approximation, really).