Swapping/paging memory is not really an issue either, unless you are running in 64 bit mode and making great use of the extended workspace addressing (e.g maybe for strategy backtesting, or having massive amounts of tick data loaded), in which case NT7 will show up with a vast active memory size in the Task Manager. I doubt you will see that.

In my case I run 8 workspaces with 30+ charts 40+ instruments and NT7 never uses more than 400k and c 3-9% cpu on a i3 laptop with inbuilt Intel graphics. Sure, it runs better on my main i7 box with ATI and NVIDIA cards and more ram, but not by much.

The reasons for your 30s delays are simply NT reloading all data on the instrument switching. All that can easily be avoided by using different workspaces for your instruments and switching between them. There are other ways to do it as well. Always think if you really need all the instrument data that you have loaded.

Cheers

All that can easily be avoided by using different workspaces for your instruments and switching between them.

I run 32 bit Ninja on 64 bit OS's, simply because I don't need to address more than 4Gb ram (most people don't either), and also in my particular case because NT7 doesn't support 64 bit eSignal.
Cheers

No answer is far-fetched if it solves your switching times. If you have Ninja reloading all your data and reloading the indicators then that's what you will have to put up with, so yes, have separate workspaces. Your machine sounds more than powerful enough (CPU, graphics and Ram) to do whatever you want with it.
Cheers

I use other tools for fast switching but I suggest you get your machine and choice of OS setup and stable using just workspaces first.

And do check that you are not one of those who loads years of tick data.
Cheers

So if I will be running 16 GB, is it wise to use 64bit architecture to access more then 4GB? I have heard its better to use 32 GB to address the paging issue. You mean that with a 32bit version this problem would not be addressed as you can only use 4GB for NT7 compared to a 64bit architecture that support 16GB or more?

You can always disable the reloading feature in Ninjatrader ->Options->Data->Historical Chart data-> disable "Get data from server", and then it will not reload the data through the datafeed, but reads it from cache instead.

Which tool/tools do you use to overcome this problem to get really fast switching? I need fast switching so I can trade more then one instrument when the possibility arises.

I built my own called FlipCharts, lives here

FlipCharts - fast chart group layering rfkFlipCharts provides fast mouse click (toolbar buttons) or keyboard (space bar +/- Ctrl/Shift) access to arbitary named groups of one or more charts, resulting in a highly flexible and easy to use chart layering facility for NT7. Simply add the …

I would wait for the version due out next week and a couple of introduction movies that will be there shortly if you are interested in exploring it.
You don't need it yet, just have a week or two with the separate workspaces running stable. On 16Gb they will switch quickly anyway; then, once you're happy and stable, FlipCharts can whisk charts from any workspace to the front without switching.
Cheers

If you want Ninja to be able to use more than 4Gb of your RAM then you have to use the 64bit version. With 16Gb you should not see any paging unless you are doing extremely large backtests or similar. Just check in the Task Manager or Resource Monitor to see how it is doing. You just don't need to use the bigger boy unless you really do need it. If you are using a lot of data then the SSD route is well worth going down, again this is easy to see in the Resource Monitor first.

Many folks think that performance is all about clock speed and cores and multi-threading. That is a small part of it compared to RAM performance and the rate at which the low level MMU stores get busted. The more stuff you are churning through in RAM the worse it gets and the CPU's are stalled more often than doing anything useful. This is not paging, this is just great systems being destroyed by large-scale sparse-access object-oriented software, no escape unless highly designed for, we want the functionality.
Cheers

I am already using 2 x SSD drives in raid 1 mirror level, for extra backup protection.
I will have a dedicated Dell Optiplex 9010. CPU i7-3770 (3.4 Ghz) with 16 MB ram machine with these 2 x 120 GB SDD drives entirely dedicated to NT7 on maybe 3-4 screens, as I recently bought a 6 screen stand. I hope the screen graphics will not take to much memory, and that my calculation heavy indicators can quickly transform between instruments and workspaces.

Would you recommend Win 7 or Win 10 for performance in NT7?

Sounds great, I wouldn't have thought the SSD's need to be dedicated just to NT but that's cool anyway.



I don't see any real performance differences, simply one of aesthetics, others might have different information, in which case I'm all ears. Personally I don't happen to like the way NT7 looks under Windows 10; I mean it's fine, but it wasn't built for it and it shows. Strictly speaking that is personal taste though, no big deal, it still runs alright, I don't mind either OS and do use both myself.

(As an aside I noticed recently with FlipCharts is that using double buffered graphics under Forms isn't so easily available under Windows 10, although I think this is likely to be more down to the choice of graphics Theme rather than OS, I have a little to explore there (switching off some of the layering/transparency nonsense, etcetera.))

So, that was an advanced lesson in useless fence sitting but, if forced, I prefer W7 for NT7, and either W10 or W7 for NT8. Not a big issue, if it's got four wheels then the missus is only interested in the colour.

Cheers