After Effects Memory on Snow Leopard
by Manny Rellesiva on Sep 4, 2009 at 5:57:59 pm
Before I was using Mac OS 10.5 and everytime I open After Effects CS4 on startup, it says "4% used of 3GHz". I have 4G installed in my mac (2.53GHZ Intel core 2 duo, 4G of Ram). I understand that it only sees 3GHz because mac os 10.5 is only 32bit.
Now Ive upgraded to snow leopard Mac OS 10.6 which is a 64bit environment. I was shocked that when i opened After effects it still says "4% use of 3G". I thought snow leopard is 64Bit how come it only sees 3G when I have 4G in my system. Im using Windows XP on my other machine and its 64bit. But it says everytime i open After effects that im using 3.9G and i have 4G installed on my windows machine so it's ok. But with my mac, even though i have 4G installed, After effects only sees 3G. Is there any setting/preferences in After effects that i should be tweaking? Please help.
Re: After Effects Memory on Snow Leopard by Kevin Camp on Sep 4, 2009 at 6:25:48 pm
actually os x has been able to address over 4gb since it was created.
your problem is that ae is only 32-bit. and as such it con only address a maximum of 4gb. in os x, a 32-bit foreground application can only use 3gb, in windows i had heard that it was 3.5gb, but it may be higher now. those are the values you see when ae starts up.
32-bit background processes (such as the ae render engines that get launched when you enable 'render multiple frames simultaneously') on both platforms can use a bit more ram. osx can use 3.5gb and windows can use nearly the full 4gb.
with 'render multiple frames simultaneously' enabled in the multiprocessing prefs, ae will be able to use nearly 4gb of ram for each processing core that is available regardless of platform (assuming the windows version is 64-bit). so if you have a 4-core machine, ae can use up to 16gb... 8-cores, up to 32gb... 8 core i7 cores with hyper-threading could theoretically use up to 64gb (hyper-threading makes each fore look and act like 2), but i've not heard of anybody using that much ram.