Inquiring about Computer HW Specifications for Time-Intensive Calculations

Click For Summary

Discussion Overview

The discussion revolves around the hardware specifications necessary for running time-intensive computational quantum chemistry programs. Participants explore various factors influencing hardware choices, including processors, memory, storage, and the potential use of graphics cards for calculations.

Discussion Character

  • Exploratory, Technical explanation, Debate/contested

Main Points Raised

  • One participant suggests consulting software developers if existing software is to be used, or conducting systems analysis for custom software.
  • Another participant emphasizes the importance of dedicated cores for chemistry applications.
  • There is a proposal that using graphics cards could significantly reduce the number of computers needed for calculations.
  • A later reply challenges the reliability of GPU-optimized code, noting that experiences can vary widely, with performance improvements or degradations observed.

Areas of Agreement / Disagreement

Participants express differing views on the effectiveness of using graphics cards for computational tasks, indicating a lack of consensus on this approach.

Contextual Notes

Participants do not provide specific calculations or benchmarks for hardware requirements, and there is an acknowledgment of variability in performance outcomes when using GPUs.

sams
Gold Member
Messages
84
Reaction score
2
Dear Everyone,

We are a research group doing research in computational sciences. We are considering ordering new computers for our research. I was wondering if anyone could advise us regarding the hardware specifications needed to run heavy and time-expensive programs, specifically in computational quantum chemistry research. Although it is well-known to consider the number of processors, the memory (RAM), the Hard Disk (capacity of data storage), etc. But is there a more specific way I can calculate the running jobs in terms of the needed computer hardware? Are there other things I should also be considering? Would recommend any book that would be helpful regarding this matter?

I really appreciate your shared experience.
Thank you so much for your help and for your responses
 
Computer science news on Phys.org
If the software you will be running already exists, you need to talk to the software developer.
If it is software that you will be writing, you need to do some systems analysis.
These types of applications (chemistry) often use dedicated cores.
 
  • Like
Likes   Reactions: sams
Thank you Sir for your help and advice...
 
DrZoidberg said:
You can greatly reduce the number of computers needed by doing the calculations on the graphics card.

You have said this several times in different threads. Have you ever actually done this? I have, and my experience is that sometimes it helps - a lot - and sometimes it doesn't - a lot. I have seen GPU-ized code run 10x faster and I have seen it run 10x slower. I would never suggest to someone that a GPU is the right direction without knowing a lot more about their code than one would get from a paragraph or two.
 

Similar threads

Replies
1
Views
3K
  • · Replies 6 ·
Replies
6
Views
3K
  • · Replies 1 ·
Replies
1
Views
2K
Replies
29
Views
6K
  • · Replies 2 ·
Replies
2
Views
4K
  • · Replies 7 ·
Replies
7
Views
3K
  • · Replies 4 ·
Replies
4
Views
2K
  • · Replies 7 ·
Replies
7
Views
2K
  • · Replies 2 ·
Replies
2
Views
4K