weighting equipment importance for backtesting

Discussion in 'App Development' started by cdcaveman, Jan 31, 2013.

  1. i'm sure its obvious that processsing power is important.. but i'm sure there are bottlenecks ... theoretically what is the best machine to do alot of backtesting on.. CPU is obviously important.. but how much does Ram play a role.. will SSD hard drives help..

    anyone care to give there opinion.. and or speculation as to what is the best machine or even what they think would be the best machine for extenstive backtesting..
     
  2. misaki

    misaki

    Don't think anyone can answer your question. It completely depends on your software implementation.

    Without knowing what you are doing and how you are doing it, currently, the best machine that you can use for practical applications is a 1.5 million core Bluegene/Q system with 1.6 petabytes of memory, although there is a K20x system that has achieved higher TFlops at peak.
     
  3. umm hmm.. out of my budget.. haha... more of the personal computer for backtesting in excel or ninjatrader ...
     
  4. misaki

    misaki

    All right, I use neither Excel nor NT. But I've played around a little with NT and can give a semi-qualified opinion.

    1. What is more important, CPU or memory? CPU, by a huge margin. It varies with user case. RAM is extremely cheap nowadays, so if you get 16 GB, I'd imagine 99.9% of NT's users don't have to worry about memory any further. Personally, I've never pushed NT past 800 MB in memory, so applying a huge safety factor, I came round with the 99.9% figure. (I've pushed other things well past 800 MB, though.)

    2. NT is notoriously single-threaded. There are workarounds, but chances are, if you are asking this question, you don't have to worry about this. The implications for you are that you should focus on clock frequency of your processor instead of number of threads for maximum value. Get an Ivy Bridge processor, of course. But note that the performance gains per dollar spent deteriorate very quickly after $300.

    3. Will a SSD help? Yes. NT relies a lot on your secondary storage (which it shouldn't doing). See: http://www.bigmiketrading.com/274968-post1.html

    Hope this helps. :)
     
  5. nice.. i have to RMA my memory ... i've got Vengance 8gigs corsair.. right now.. . and i'm sick of RMAing ssd's from OCA..... gonna buy a intel.. i've had these OCZ's crapping out on me alot... its like all of a sudden they just disappear.. been trading options .. looking to do some backtesting to see if i can get a little more of an edge direction wise.. C# seems possible for me.. with alot of patients.. and the wisywig.. helps get a feel for what the code is doing.. you can kind of develop indicators on indicators.. and then view the code.. and get a idea how the script is working... this isn't high frequency trading stuff.. its more of looking for a satistical edge with such things as regime switching, TSI.. Rsi(atr) i wish i could develop a relative skew indicator between the index and baskets or even just single members.. but whatever one day at a time.. i can barely get through the errors i get through now..
    personally i wish i would have gotton 2 xeons, intel SSD's , 32 gigs of ram.. and called it a day.. i've had so many problems with the SSD's i can't tell you .. i have a few friends that are that work in feilds relative to server clustering/ virtualization, etc etc.. friends are truely THE ultimate resource in life..
     
  6. interesting article.. basically.. the guy is doing the reverse of what a swap file is for on a hard disk.. .when memory is being used up.. it swaps a section of the hard disk to ram for usage..

    swapping the ram for the hard drive in this case is a very good idea haha.. NT is very very very slow at switching charts and such.. its so so so annoying..

    there is no one more intolerate of low speed.. i remember learning about bottlenecks.. and hard disk are the ultimate one.. the first thing i did when i learned about it .. was i spent the money bought two SSD's and raid zeroed them together.. haha double the speed.. more then half the reliability is my experience hahah..
     
  7. the other thing.. have a local DB to access historical data seems so much more efficient.. storage is cheap.. SSd's are cheap.. why suck the data from a feed?
     
  8. misaki

    misaki

    I've had problems with SSDs too, but it gives the best value as of the moment. The global production of HDDs hasn't recovered since the 2011 flood in Thailand, otherwise it would have been cheap to use a HDD platter.

    From your description, you just need a <$300 Ivy Bridge processor, 8 GB RAM and a SSD. Make incremental improvements when the need is called for.

    NinjaTrader is a business study in software development. Their issue tracking wiki must be larger than their Kinetick server capacity... ):
     
    #10     Jan 31, 2013