Problem using the two cards concurrently Using 2 cards as the only one

Accelerated Computing CUDA CUDA Programming and Performance
#1

Dear al,
I have the problem like this. My machine have two graphic card, Geforce 9800 GT with the memory for each is 512Mb. My program needs to use a very large struture called A, as much as poosible. It runs well with just using only one card.
Now I want to improve the performance. I want to use 2 cards like described follows, one part of the structure A in card 1, one part in card 2. And when single threads runs, it will access the A just like with only one cards (actually it needs some modification for condition).
And my question is,

  1. Is it possible to do something like this
  2. And if possible, is there any recommendations.

Thanks u so much,
I hope that my question is clear to everyone.

#2

No. Each GPU is a completely separate device, and actually has to be controlled by its own CPU thread. Even if what you wanted were possible, performance over the PCIe bus would be horrible.

#3

Thanks for your answer.

I see. But in my case, when we need to allocate a very large memory in GPU, that exceeds the size of one card, maybe the size of 2 cards is enough. I cannot know any solutions for this.

My alogorithm is like this, we have a fixed and very large memory for the result, called structure A. Every time, one threads runs, it gets some information from A, computes something and updates the A.

Could u have any recommendations for the problem like this.

Can we do something like this, we allocate the memory for A in two cards 1 and 2. Every time, one threads run, depends on the information it has, it knows (in fact it must know) the information it needs is in card 1 or card 2. And it takes what its wants, computes and updates A.