Re:Donald MacKenzie on high-frequency trading
by bluecalm on Thursday September 04, 2014 @09:09AM
(#8267971)
Obligatory link: http://stackoverflow.com/questions/11227809/why-is-processin... see most upvoted answer for short lecture about cost of branching in modern CPU's.
Re:Donald MacKenzie on high-frequency trading
by kasey_junk on Thursday September 04, 2014 @09:07AM
(#8267959)
It is probably referring to the fact that code that branches allot will make the prefetcher/branch predictor fail more frequently which causes slower memory loads and less ops. Not sure why it mentions parallelism though.
"Don’t touch the kernel. Don’t touch main memory … Don’t branch.’ (That third commandment means don’t fill your program with ‘if’ statements – ones with the generic form ‘if A then do B else do C’ – because they get in the way of a modern microprocessor’s capacity to do several things in parallel.)"