Why is integrated video support give poor performance with Home Designer Software?

[David J. Potter]

I asked this question originally at Chief Talk for my own purposes, this was my initial question:

I answer posts daily at Home Talk ( a user forum for Home Designer End Users, consumer grade software published by Chief Architect Inc based upon Chief Architect Premier but with less features).

A common complaint by these people is about poor video card performance, usually integrated video hardware (as part of the person's mother board and not a stand-alone video card).

I would like to better understand the physics of why this is so. I know this has been discussed before, I know what to do so I obtain good results myself but just not why this is a common problem for people who buy economy type PC's or laptops in terms of the actual physics involved.

If any of you can refer me to good, well written explanations of this phenomena, I would appreciate it so I can better help others. I mean, why is it that Chief Inc software is so demanding over and above playing video files and simple video games, which most PC's can deliver reasonable performance.
I mean what is the "make-break" and why?

Read this article Please:

http://www.ehow.com/facts_6737699_in...ideo-card.html

http://www.chieftalk.com/images/styl...quote_icon.png Originally Posted by David J. Potter http://www.chieftalk.com/images/styl...post-right.png
I suppose there is not an "easy", pedestrian way to say to someone who just wants to create something on a $300.00 laptop with $100.00 software that they will just have to spend a little more treasure than that to arrive at the level expected or hoped for, no matter how naive.

DJP



True David.

In this case, the buyer of the software should not expect good performance if their system does not meet the posted minimums.

http://www.homedesignersoftware.com/...uirements.html

"As others have basically said, the major difference between real time graphics and video is that in real time graphics something, be it the CPU or GPU, must perform all of the mathematical calculations to take a set of 3D mathematical data and display it on a 2D screen. This means performing many computations for potentially hundreds of thousands or millions of mathematical vertices in (ideally) no more than 1/30th of a second.

A GPU does this significantly more efficiently than a CPU for a few basic reasons, notably:
1.) Video memory on discrete cards is typically much faster than system memory. Note in the case of integrated cards video memory IS system memory so this gain doesn't exist.

and more importantly

2.) GPU's are very good at doing these calculations massively in parallel. A video card might be able to perform hundreds or thousands of mathematical calculations at the same time, whereas a CPU can perform one (potentially one per core) of these calculations at a time. Higher end cards are likely to be able to do more in parallel than lower end cards. As an example of this, the Intel HD 4000 integrated card has 16 execution units whereas the GeForce GTX 680 has 1536."

Last edited by Ryan Matern; 01-14-2013 at 04:44 PM.

Ryan Matern
Software Engineer
Chief Architect

Copy-pasted from the Chief Talk forum by David J. Potter