JForex JVM heap size MAY prevent your system from crashing
This solution prevented my JForex from crashing. It was originally set at 512mb but it is now set at the max which is 4 gb approx. As this automated has survived the night, it appears this was the best solution so far. Also it has been running for well over 12 hours where the amount of heap spaced used is 14. gb out of 3.6 gb internally of JForex. I would recommend attempting this fix before another server with mega amounts of RAM like 32 gb for instance.
This was a comment for one of my videos. I need to do this for my Linux environment
Interactive Brokers TWS can be quite slow, especially if you follow a lot of instruments at the same time. This article provides some useful tips for increasing Java memory size for TWS, which significantly improves performance. http://www.tradinggeeks.net/2014/05/interactive-brokers-performance-optmization/
Popular videos of uses of Redis including Twitter with use of Java JVM vs C and C++
NOTE: As I watch these videos and the advancements of Redis, is there a real reason to have even Erlang as part of this proposed system? Since Redis has cluster techniques, it gets redundant pretty well.
Why does Lua keep coming up? Twitter video says it is not production ready. She says Redis PUB/SUB could the better service than Storm for pipelining. See below
Check these out:
This is one of the best presentation on the use of Redis and use of C/C++ vs Java. A total insight on how Twitter has a very fast service:
1. Oracle, as a corporation, has no incentive to fix security issues. It doesn’t generate profit.
2. Taking a PR beating eventually provided enough incentive – it finally lit enough of a bonfire under their nuts to fix the issue.
After removing the latest Java, I can safely I say I don’t get mysterious Windows update daily.
In a nutshell, Java is outta here!
No need to flame me after working 8 years with it.
IS JRockit the way to modify Java or JVM for high speed quant development and HFT platform?
After looking for how people are ‘optimizing’ Java and the JVM, it sounds like JRockit was a standard JVM option. I also heard IBM had a custom JVM as well. Some info sounds like Jrockit could be used from Oracle which is ‘open source’ as well. It sounds like it is part of the standard WebLogic app server product.
I am nore sure if these methods are still good enough to use for a HFT platform as compared to C++ or C. Anyone want to convince the general community on this? This is especially true with Java’s garbage collector.