NOTE I now post my TRADING ALERTS into my personal FACEBOOK ACCOUNT and TWITTER. Don't worry as I don't post stupid cat videos or what I eat!
http://quantlabs.net/membership.htm
Learn how I may implement this combo
NOTE I now post my TRADING ALERTS into my personal FACEBOOK ACCOUNT and TWITTER. Don't worry as I don't post stupid cat videos or what I eat!LMAX Disruptor model gets less than 1 ms latency for quant development and HFT? Video using open source Java, JBOSS, PostgresSQL
LMAX – How to Do 100K TPS at Less than 1ms Latency
Summary
Martin Thompson and Michael Barker talk about building a HPC financial system handling over 100K tps at less than 1ms latency by having a new approach to infrastructure and software. Some of the tips include: understand the platform, model the domain, create a clear separation of concerns, choose data structures wisely, and run business logic on a single thread.
http://www.infoq.com/presentations/LMAXgen
Also, the one thing I like about this talk is how they pre-allocated their buffers to optimize and minimize garbage collection. They put everything in a ‘ring buffer’ to minimize GC and optimize their caching. This may eliminate those who doubt Java because of garbage collection.
This has been open sourced as well http://code.google.com/p/disruptor/
After researching, this may become a game change for my solutions?
NOTE I now post my TRADING ALERTS into my personal FACEBOOK ACCOUNT and TWITTER. Don't worry as I don't post stupid cat videos or what I eat!How LMAX uses standard Java Hotspot compiler and vanilla Java code using Disruptor pattern making their HFT lightening fast!
This was part of a Linked In discussion:
No, we use the standard hotspot compiler. We just looked hard at how it worked and then wrote standard Java that we knew would compile down to high performance code. We did have to do some minor things around creating cache line padding so we didn’t have false sharing. But other than that the Disruptor is vanilla Java written with an understanding of CPU architecture and the cost of contention. It’s basically a lock-free algorithm which is why it performs so fast.
NOTE I now post my TRADING ALERTS into my personal FACEBOOK ACCOUNT and TWITTER. Don't worry as I don't post stupid cat videos or what I eat!LMAX has a unique Java disruptor for multithreading latency for quant development. Now C++ and C# versions exists
Here the links you should know about this:
http://stackoverflow.com/questions/6943569/lmaxs-disruptor-pattern-is-there-a-port-to-c
http://www.2robots.com/2011/08/13/a-c-disruptor/
From LMAX In Java:
http://code.google.com/p/disruptor/
C++ version:
http://code.google.com/p/disruptor-cpp/
http://www.infoq.com/presentations/LMAX
NOTE I now post my TRADING ALERTS into my personal FACEBOOK ACCOUNT and TWITTER. Don't worry as I don't post stupid cat videos or what I eat!
