freeing up memnory

i think that would depend on the VSTi, and the size of the resulting sample.
If it is a long sample, it will obv iously uise up more memory than a short one.
A complex VSTi will use moree memory than a simple one.
 
Memory management is more complex than it appears at first sight.

I'm just *starting* to get my head round the way it works in Linux and I presume it's similar in Windows.

I can see a real-time system working *much* faster with more memory, and it's all down to the amount of page swapping that's going on.

Basically - even though your system might be showing a fair amount of free memory, you could still have your processor doing lots of work paging the memory in and out for code and data. That adds latency, which is obviously *bad* in a real-time environment such as a DAW.

I can see some operations on this system having around 50 milliseconds added while a lot of page swapping occurs.

I was always sceptical about the more memory makes your machine go faster argument, until I started to look at this stuff.

I guess my problem was - I'd always worked on non-MMU based real-time systems previously and wasn't aware how much work needs to be done to work with virtual memory. I'm also coming to the conclusion that user mode / MMU isn't the ideal way forward for some embedded real-time systems. It makes it very hard to have deterministic responses. I guess a hybrid of the two is the best way: kernel mode for hard response requirements and user mode for non-critical stuff.
 
have ditched LINUX cos i can't be bothered faffing around with emulators to run cubase so am back on MS XP atm.....

also:

DAW & MMU = ??
 
DAW - Digital Audio Workstation (e.g. Cubase system)
MMU - Memory Management Unit (deals with your virtual memory)

BTW - the Linux system I'm working on is an embedded real-time system under development in my day-job.

I don't run Linux at home.

XP is the way to go for DAW at the moment.
 
Biggins2012 said:
probably a daft question but, does converting a line/sound/noise/line (whatever to a wav file and then importing it into cubase use up less memory that using the VSTi to make it??

Not necessarily, but it *will* free up cpu.
The amount of memory used by a running VSTi is unimportant compared with the CPU load required to do all those squillions of calculations.
So bouncing to audio is always a good thing to do once you're happy with a part.
 
Or these days, *freezing* the part effectively gives you a short cut bounce. Some would argue whether or not it's really a short-cut though.
 
Back
Top