I use NT7 heavily with scripts as I have built my own custom ground-up trading and backtesting system in Ninjascript (over 6000 lines of code so far and growing).
I am keen on NT8 primarly for its multithreading capabilities (which I am still not sure about). With many CPU cores available and huge parameter spaces to search I have to parallel back-test; with NT7 the only way I can do it is to simultaneously run multiple instances of NT7 on multiple virtual machines (up to 20 at a time).
I have checked the code-breaking section of the NT8 guide and it appears that to convert my scripts for NT8 would be fairly simple; I think the vast majority of my code will run without any changes at all.
In order to properly beta test NT8 for my use scenario (and it is in my interest to do so because as a beta tester I can influence what gets fixed) I feel I should move all of my research and development to NT8 already. I don't have time to set up and work with a parallel NT8 test bed. Either I work with it now as my daily driver, warts and all, or I wait until the release candidate.
But is this advisable? Just how "finished" is NT8 at this point, and could I trust it to handle my scripts for actual production work? From what I can tell just starting it up, loading charts, and fiddling around with it it looks pretty finished to me. But I cannot tell what awaits lurking beneath the surface...
I am keen on NT8 primarly for its multithreading capabilities (which I am still not sure about). With many CPU cores available and huge parameter spaces to search I have to parallel back-test; with NT7 the only way I can do it is to simultaneously run multiple instances of NT7 on multiple virtual machines (up to 20 at a time).
I have checked the code-breaking section of the NT8 guide and it appears that to convert my scripts for NT8 would be fairly simple; I think the vast majority of my code will run without any changes at all.
In order to properly beta test NT8 for my use scenario (and it is in my interest to do so because as a beta tester I can influence what gets fixed) I feel I should move all of my research and development to NT8 already. I don't have time to set up and work with a parallel NT8 test bed. Either I work with it now as my daily driver, warts and all, or I wait until the release candidate.
But is this advisable? Just how "finished" is NT8 at this point, and could I trust it to handle my scripts for actual production work? From what I can tell just starting it up, loading charts, and fiddling around with it it looks pretty finished to me. But I cannot tell what awaits lurking beneath the surface...
Comment