OpenCl and the tools for it. Reviews and impressions. - page 29

New comment
 
Renat:
You are retelling a theory that all interested people already know.

The realities are that the cpu is faster in general purpose tasks due to a combination of factors. This has now become clear. The gpu silver bullet categorically fails to reach the target.

I suppose this is too crude a generalization of GPU program development results. After all, there are people reading this forum who are not interested in speeding up "general purpose tasks" but in speeding up OPTIMIZATION and TESTING of complex numerical algorithms within a trading terminal.

Renat, sometimes you react so nervously to the slightest disagreement with your opinion.....

And it is such a pleasure to see - how MT4 terminal quickly and calmly finds an optimum for a trading system, working on a VIDEO CHARD, while you are quietly doing any other job on the same computer. While doing it on a host processor, even in multithreaded mode, would be much longer and more expensive. Moreover, if you have for example 3-4 graphics cards, you can run 4 terminals and simultaneously optimize 4 currency pairs, almost without noticing it.


Many thanks to the developers of Metatrader 4 for the great development tool.


Especially valuable is the portability of programs from MQL4 to classical C and back. This saves a lot of time.

Another value is a clean interface of MQL4 with DLL inside the terminal without any unnecessary complications and nonsense. Without this, CUDA program development would be downright difficult.

 

(wiping away a man's tear) ....

... And so I forgive all the moderators of this forum for banning me several times over the past few years.

I forgive everyone for everything.

 
Use mt5 - opencl is native there.
 
Renat:
Use mt5 - opencl is native there.
But you banned me from the fifth forum a long time ago. How am I supposed to get advice?
 
As far as I can see, there is no ban on five
 

I checked AlexEro, but not AlexEros.

Try again, ban removed.

 
Renat:

I checked AlexEro, but not AlexEros.

Try again, ban removed.

Ok, it's working.
 

An example of using GPU acceleration for trading (derivatives).

Mark Joshi - famous for his books on financial mathematics, in particular on derivatives and options trading, has reported here about his work:

http://ssrn.com/abstract=2388415

He translated his OOP-style work to CUDA GPU. He started it in 2010, then had a break, and from 2011 until summer 2014 he made it to working version 0.3. He managed to achieve acceleration of 100X... 137X times - and that's on a CONNECTED algorithm, which is difficult.

The work used the QuantLib library in C++, which he himself admits he had to rework along the lines of "OOP ->-> procedural approach" - in order to make it all work on the CUDA GPU.

He writes:

"I have implemented Monte Carlo pricing of IRD with the LMM on the GPU with least-squares for early exercise features.

You can get the code from kooderive.sourceforge.net in both C++ and CUDA. The paper is at ......

I used a completely different code for CUDA than I had previously used for C++. In essence, I treat data as the central concept and use the code to act on the data. The style is very functional. It did took a lot of work because my previous C++ implementations had been object oriented."

His project itself is open source:

http://sourceforge.net/projects/kooderive/

Files:
ssrn-id2388415.zip  288 kb
1...2223242526272829
New comment