CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > General Forums > Hardware

Pc with min. 128 GB RAM for numerical computations

Register Blogs Community New Posts Updated Threads Search

Reply
 
LinkBack Thread Tools Search this Thread Display Modes
Old   January 4, 2021, 03:21
Question Pc with min. 128 GB RAM for numerical computations
  #1
New Member
 
Join Date: Jan 2021
Posts: 1
Rep Power: 0
Came is on a distinguished road
Hi everyone,

I'm planning to build a PC for numerical computations for science applications. One of the main tasks involves computing the FFT (Fast Fourier Transform) of large datasets using programs such as Matlab or Python. In order to handle such large datasets, the pc should have at least 128 GB RAM (maybe even 256 GB, but this seems to be tough with the given budget of around 2200 $)

Currently, I do have a build (https://de.pcpartpicker.com/list/cgdRnL) but just with mainstream components. This build should probably be fine, but there could still be room for improvement. Since we are going to perform numerical computations and handling huge datasets, I thought that it could make sense to ask hardware experts like you for some advice.

I would be very grateful for any advice. Thank you very much
Came is offline   Reply With Quote

Old   January 4, 2021, 06:42
Default
  #2
Member
 
dab bence
Join Date: Mar 2013
Posts: 48
Rep Power: 13
danbence is on a distinguished road
The build looks good.

A concern I have is that the ram is not ECC. With 256GB you will at some point get bit flips caused by cosmic rays. If the vast majority of the data stored is numbers to be FFTed , you will got get any error message, just a bit flip in one of your numbers.

Below is an old quote I pulled off the internet

“Generally speaking, cosmic ray soft errors occur in DRAM memory at a rate of ~10 to 100 FIT/MB (1 FIT = 1 device fail in 1 billion hours). So a system with 10 GB of memory should show an ECC event every 1,000 to 10,000 hours, and a system with 100 GB would show an event every 100 to 1,000 hours. However, this is a rough estimation that will change as a function of the effects outlined above.

With 255GB, you might get a bit flip every 50 hours. I personally don’t like those odds as it puts doubt in my mind whether an issue is a bug or a comic ray. Do I need to run the FFT twice to make sure ?
danbence is offline   Reply With Quote

Reply

Tags
numerical computation, pc build, ram


Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
y+ and u+ values with low-Re RANS turbulence models: utility + testcase florian_krause OpenFOAM 114 August 23, 2023 06:37
Converging Diverging Nozzle with dbnsTurbFoam Saleh Abuhanieh OpenFOAM Running, Solving & CFD 4 December 13, 2019 11:26
Near wall treatment in k-omega SST Arnoldinho OpenFOAM Running, Solving & CFD 38 March 8, 2017 14:48
Problem of simulating of small droplet with radius of 2mm liguifan OpenFOAM Running, Solving & CFD 5 June 3, 2014 03:53
On the damBreak4phaseFine cases paean OpenFOAM Running, Solving & CFD 0 November 14, 2008 22:14


All times are GMT -4. The time now is 09:41.