speeding up simulations on linux

Toggle sidebar Toggle sidebar
I

iceblitzed

Distinguished
May 12, 2014
1,541
0
20,460
Hello, I am using a program for physics simulations.
I have read tips on how to optimize performance on linux mint.
Here is a short summary of what I have done so far.


/etc/sysctl.conf

vm.swappiness=10 in config file
vm.vfs.cache.pressure=50 in config file


# Move /tmp to RAM
sudo nano /etc/fstab
tmpfs /tmp tmpfs defaults,noexec,nosuid 0 0

Is there anything else I can do to specifically speed up simulations in linux
 
Solution
skittle
Speeding up programs, and specifically physics simulations, depends LARGELY on the algorithm used and the programming language. For example if you are trying to use a PDE algorithm like Eulers method or Runge-Kuta you are more much likely to be bound by CPU limits than IO limits. Thus you would only benefit by using a faster algorithm, faster programming language or buy better CPU. Adjusting swapiness (this has no impact unless you are running out of RAM) and moving data to /tmp may have no impact at all if you are CPU bounded!

More specifically I don't know of any way to parralize these kinds of algorithms so they don't take advantage of multi-core processors or GPU processing. Maybe the kinds of simulations you are doing can...
Sort by date Sort by votes
skittle

skittle

Splendid
Dec 27, 2006
4,036
3
23,960
Speeding up programs, and specifically physics simulations, depends LARGELY on the algorithm used and the programming language. For example if you are trying to use a PDE algorithm like Eulers method or Runge-Kuta you are more much likely to be bound by CPU limits than IO limits. Thus you would only benefit by using a faster algorithm, faster programming language or buy better CPU. Adjusting swapiness (this has no impact unless you are running out of RAM) and moving data to /tmp may have no impact at all if you are CPU bounded!

More specifically I don't know of any way to parralize these kinds of algorithms so they don't take advantage of multi-core processors or GPU processing. Maybe the kinds of simulations you are doing can be parralized in some way. For example I have written a Runge-Kutta implentation in C, but it can only run on 1 core of my quadcore processor :( This is where I would start looking for ways to speed up your calculations.
 
Upvote 0 Downvote
Solution
I

iceblitzed

Distinguished
May 12, 2014
1,541
0
20,460


ok thanks im glad I have already optimized my system for multiple core usage.
 
Upvote 0 Downvote
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom