The next wave in trading speed

Posted: February 10, 2010 in GPUs, Uncategorized

With the rise in the need for speed in trading over the last decade, it appears the next step in trading is going to be the migration to massive parallel computing with graphical processing units (GPUs).  A GPU differs from a CPU in the fact that it has a parallel architecture allowing it to run many threads simultaneously.  The main provider of GPUs is NVIDIA and the language they are utilizing is called CUDA (Compute Unified Device Architecture) which is basically C.

With that being said, there is a new text that looks to be quite good.  It is written by David Kirk.

I believe over the coming years this will be the source of many small edges in trading.  Mostly, I think it will occur from optimization of older option quoting (mass quoting is painfully intensive) and market making algorithms.

Additionally, I ran across this which has quite a bit of information for this textbook on it (http://sites.google.com/site/cudaiap2009/materials-1/cuda-textbook).  I believe it is the draft version of the first few chapters of this textbook.

Finally, there are some cool college courses on GPU programming coming out; here (http://www.evl.uic.edu/aej/525/) is one such course.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s