Machinist not a programmer

Parallel processing was a stated goal in the 2009 release of CAELinux. The last sentance in the first paragraph of CAELinux 2009 release stated: With these software, Open-Source FEA & CFD is continuing its progression towards ease of use and parallel computing power to get closer to commercial FEA & CFD codes!

caelinux.com/CMS/index.php?option=com_co...9&catid=69&Itemid=41

AND
The page describing New release: CAELinux 2013 Bullet #6 reads: CAE, FEA, CFD & multiphysics simulation: Salome_Meca 2013.1, Code-Aster 11.3, Code-Saturne 3.0, OpenFOAM 2.1.1, Elmer 6.2, Calculix 2.6, Impact FEM, MBDyn (all parallel codes)

caelinux.com/CMS/index.php?option=com_co...tid=1:news&Itemid=29

I was hoping that others using these programs through CAELinux would have posted their coding fixes when running in parallel but hey have not. So it looks like I will be starting a new thread on Parallel Processing and search the forums of the individual pieces of software so I can copy over their research & experiences. I was figuring on six months to learn the several pieces of software I will need to use for my project; this could add another six months to it. I still don't know if the computer I am looking at will work with the software I need.

Thanks! I am buying multi core server processors and if I understand what you are saying I won't need a GPU unless there is code written for it because THAT is way outside of my abilities.... unless I wanted to become a programmer but I'd rather stare at the sun, at least its alive...

I appreciate the people who write programs because the little I have done tells me I wouldn't have a computer if that was up to me.

My friend Pat who is a programmer doesn't like integrated graphics because the chip changes your settings to use IG instead of your graphics card even when the display runs slower on the chip. I have read that some programmers try to keep the RAM down to levels that will not overburden the chip's IG. My fear of that approach would be finding the program difficult to run on a processor different than the one used to create the program. That kind of technical detail is not always available.

Its been difficult to find information on this subject {weird since computer info used to be the easiest subject to research online} I did see someone say that some programs are accelerated only on the GPU it was programmed for. That made me hesitant to buy one very expensive 24G GPU in favor of several smaller ones of different types as another place had a suggestion that the program will use the GPU with the best configuration available. I had decided to pass buying a $5000 GPU for $1200 on e-bay and your comment only reinforced my original plan:

BUY NOTHING UNTIL YOU KNOW THAT IT WILL ALL WORK TOGETHER.

Thanks again claws !