can Vegas benefit from a dual processor system? at DVinfo.net

John Gaspain · December 8th, 2004, 04:59 PM

Im thinking of going dual, will my renders and previews be faster and higher FPS?

(im editing HDV)

Aaron Koolen · December 8th, 2004, 07:17 PM

Technically yes, practically no (In one instance of the software). Vegas uses multiple processors to spread the work out between rendering video and audio only. It won't give some of your video frames to one and some to the other - unfortunately. Of course having two processors is nice anway cause you could fire up another instance of Vegas and render another project out or if practical, have two instances each rendering half the project.

Aaron

John Gaspain · December 8th, 2004, 07:29 PM

wow, thats very interesting. thanks!

Rob Lohman · December 9th, 2004, 04:53 AM

Or have one (or part of a) project render and work on another
project (part) etc.

Kyle Ringin · December 9th, 2004, 04:50 PM

How does this work with a single CPU Hyper-Threading P4?

I've tried it with HT on and off and found I get better preview framerates with HT on, But faster render times with it off.

Do you guys run HT on or off?

I'm running with it on at the moment because I can deal with a couple of extra rendering minutes, but I'd rather have a better preview while editing.

Aaron Koolen · December 9th, 2004, 05:31 PM

I haven't done any tests myself but that doesn't surprise me. I have HT on (I would never turn it off as it has lots of other benefits like running other programs smoothly as I render etc) and when I render, I see that not all of my CPU is being used (I think it was 70% when I last looked). So I'd assume with it off, probably most of it would be and therefore you'd get a faster render time. How much faster were you getting with it off?

Cheers
Aaron

Kyle Ringin · December 9th, 2004, 07:46 PM

I didn't have time to do a comprehensive test but heres the results:

I rendered 1 min of PAL DV footage with heaps of effects applied. My 3GHz P4 was doing ~5.1 fps at Preview Full resolution With HT on. With it off it would only do ~4.8 fps (not much of a difference, but with less effects applied I think there'd be more noticably differences)

When rendering to mpeg2, IIRC it was about 2:30 vs 1:40 (a fair difference!) with HT on and off respectively. With HT on the CPU utilisation was ~80% and with it off it was 99%-100%.

Hmm. I think I need to do a more comprehensive test when I get time, I'll also write things down so I don't forget them...

Aaron Koolen · December 9th, 2004, 09:14 PM

Well those rendering times would coincide roughly to what I was thinking with HT using only 70% CPU to render and non HT using the full amount.

Aaron

Van Lam · December 10th, 2004, 07:29 PM

I heard somewhere that with dual processors, Vegas outputs the video to one and the audio to the other. I guess if you work extensivly with audio, 2 processors will help.

Glenn Chan · December 11th, 2004, 01:40 AM

Kyle's results seem to be what most people get.

MPEG2 rendering:
(from the sony forum)
Dual processors really speeds things up.
Hyperthreading speeds things up ~20%.

Vegas rendering (short/benchmark/non-real world projects):
By short projects I mean projects that barely requires rendering- i.e. video with 1 filter on it.
Dual processors really speeds things up.?
Hyperthreading speeds things no more than ~20%.

Vegas rendering (very complicated renders):
Hyperthreading makes about a 0% difference on rendering speed, if you compare HT on to HT off. When hyperthreading is disabled, I believe your CPU gets to use double the cache. This speeds things up to a small degree.

Vegas rendering for real world projects is somewhere in between.
Hyperthreading gives about a 3% speed bump. It depends on the number and rendering intensity of your filters.

Quote:

I heard somewhere that with dual processors, Vegas outputs the video to one and the audio to the other. I guess if you work extensivly with audio, 2 processors will help.

From a post on the Sony Vegas forum, I remember:
One processor handles the video rendering.
The other handles DV decoding and encoding, as well as audio processing.

It generally happens that DV encoding as well as audio processing take very little processing power compared to video rendering. A Pentium processor should be able to push a huge # of audio effects, and an AMD64 processor even more.

Dave Largent · December 16th, 2004, 11:40 PM

Real-world tests on medium-complex renders,
HT gives ~3% improvement.
Never seen anything render slower
with HT as compared to not HT.

Jim Lafferty · December 22nd, 2004, 09:32 PM

There's a dramatic difference in general workflow on an MP machine that a lot of people overlook -- I've got both a single CPU AthlonXP 2000+ machine and a dual AthlonMP 2000+ machine, both with a gig of RAM. I notice far fewer hangups, much lower launch times, and a generally snappier interface with the dual CPU machine than with the single. It's little things like these, and not render times, that make an SMP machine my favored platform -- and then there's multiple instances, and running other apps while rendering where an SMP machine really shines...

- jim

December 8th, 2004, 04:59 PM	#1
John Gaspain Major Player Join Date: Oct 2003 Location: WA-USA Posts: 371	can Vegas benefit from a dual processor system? Im thinking of going dual, will my renders and previews be faster and higher FPS? (im editing HDV) __________________ The glory of the World passes by.

December 8th, 2004, 07:17 PM	#2
Aaron Koolen Trustee Join Date: Apr 2002 Location: Auckland, New Zealand Posts: 1,727	Technically yes, practically no (In one instance of the software). Vegas uses multiple processors to spread the work out between rendering video and audio only. It won't give some of your video frames to one and some to the other - unfortunately. Of course having two processors is nice anway cause you could fire up another instance of Vegas and render another project out or if practical, have two instances each rendering half the project. Aaron __________________ My Website Meat Free Media

December 8th, 2004, 07:29 PM	#3
John Gaspain Major Player Join Date: Oct 2003 Location: WA-USA Posts: 371	wow, thats very interesting. thanks! __________________ The glory of the World passes by.

December 9th, 2004, 04:53 AM	#4
Rob Lohman RED Code Chef Join Date: Oct 2001 Location: Holland Posts: 12,514	Or have one (or part of a) project render and work on another project (part) etc. __________________ Rob Lohman, visuar@iname.com DV Info Wrangler & RED Code Chef Join the DV Challenge \| Lady X Search DVinfo.net for quick answers \| Buy from the best: DVinfo.net sponsors

December 9th, 2004, 05:31 PM	#6
Aaron Koolen Trustee Join Date: Apr 2002 Location: Auckland, New Zealand Posts: 1,727	I haven't done any tests myself but that doesn't surprise me. I have HT on (I would never turn it off as it has lots of other benefits like running other programs smoothly as I render etc) and when I render, I see that not all of my CPU is being used (I think it was 70% when I last looked). So I'd assume with it off, probably most of it would be and therefore you'd get a faster render time. How much faster were you getting with it off? Cheers Aaron __________________ My Website Meat Free Media

December 9th, 2004, 04:50 PM	#5
Kyle Ringin Major Player Join Date: Oct 2004 Location: Melbourne, Australia Posts: 241	How does this work with a single CPU Hyper-Threading P4? I've tried it with HT on and off and found I get better preview framerates with HT on, But faster render times with it off. Do you guys run HT on or off? I'm running with it on at the moment because I can deal with a couple of extra rendering minutes, but I'd rather have a better preview while editing.

December 9th, 2004, 07:46 PM	#7
Kyle Ringin Major Player Join Date: Oct 2004 Location: Melbourne, Australia Posts: 241	I didn't have time to do a comprehensive test but heres the results: I rendered 1 min of PAL DV footage with heaps of effects applied. My 3GHz P4 was doing ~5.1 fps at Preview Full resolution With HT on. With it off it would only do ~4.8 fps (not much of a difference, but with less effects applied I think there'd be more noticably differences) When rendering to mpeg2, IIRC it was about 2:30 vs 1:40 (a fair difference!) with HT on and off respectively. With HT on the CPU utilisation was ~80% and with it off it was 99%-100%. Hmm. I think I need to do a more comprehensive test when I get time, I'll also write things down so I don't forget them...

December 9th, 2004, 09:14 PM	#8
Aaron Koolen Trustee Join Date: Apr 2002 Location: Auckland, New Zealand Posts: 1,727	Well those rendering times would coincide roughly to what I was thinking with HT using only 70% CPU to render and non HT using the full amount. Aaron __________________ My Website Meat Free Media

December 10th, 2004, 07:29 PM	#9
Van Lam Regular Crew Join Date: Jun 2004 Location: Marin, CA Posts: 35	I heard somewhere that with dual processors, Vegas outputs the video to one and the audio to the other. I guess if you work extensivly with audio, 2 processors will help.

December 16th, 2004, 11:40 PM	#11
Dave Largent Trustee Join Date: Aug 2003 Posts: 1,483	Real-world tests on medium-complex renders, HT gives ~3% improvement. Never seen anything render slower with HT as compared to not HT.

December 22nd, 2004, 09:32 PM	#12
Jim Lafferty Major Player Join Date: Dec 2003 Location: Brooklyn, NY Posts: 636	There's a dramatic difference in general workflow on an MP machine that a lot of people overlook -- I've got both a single CPU AthlonXP 2000+ machine and a dual AthlonMP 2000+ machine, both with a gig of RAM. I notice far fewer hangups, much lower launch times, and a generally snappier interface with the dual CPU machine than with the single. It's little things like these, and not render times, that make an SMP machine my favored platform -- and then there's multiple instances, and running other apps while rendering where an SMP machine really shines... - jim __________________ Realism, anyway, is never exactly the same as reality, and in the cinema it is of necessity faked. -- J-L G