Holee Smokes!

Spot|DSE wrote on 6/8/2005, 2:36 PM

Just finished building an AMD dual 275 machine on a Tyan K8WE mobo, 3 Barracuda 250's, with a PATA boot drive. 4 gigs of RAM running WinXP 32.

This mama screams!
I captured a fairly lengthy stream of HDV, and dropped it on the timeline to test speed, and thought I'd made an error when the m2t file played back at a solid 29.97fps. Rendering out to 4:2:2 and DV simultaneously, this thing is moving at a fast clip. I'm rendering 9 hours of HDV at one shot, and started the render early this afternoon so it would be done by late tonight. It's moving along so fast, I suspect it will be done with both renders in a quarter the expected time.
The power of the new AMD multicores is pretty obvious at this point. This is the first AMD machine I've worked with deeply in over a year. The motherboard has some configuration issues, but after spending several hours with it and the Tyan tech support yesterday, it's been kicking up dust very nicely.
So, if you're looking into a new system for HDV....I'm bowled over.
We bought most of the parts from Colfax International, I'm not associated with them in any way, except they were recommended to us by a person at AMD. (oddly enough, Intel recommended them as well)

Back to post

Comments

Spot|DSE wrote on 6/10/2005, 10:25 AM

Good, which is the default. Changing any of those settings (which a lot of folks do), invalidates the test

GlennChan wrote on 6/10/2005, 10:42 AM

Where is this rendertest?

2- If it's different from the original rendertest.veg, shouldn't we call it something else?

Spot|DSE wrote on 6/10/2005, 12:08 PM

It's noted in the download that this is different than the rendertest for V2, 3,4.
It needed to be changed for V5, because 3D had a significant impact on render times.

http://www.vasst.com/search.aspx?text=render%20test is where you'll find it.

GlennChan wrote on 6/12/2005, 11:38 AM

Spot, that link shows four different rendertests.
Which one is it exactly?

2- Do you know of any other results to that benchmark?

FrigidNDEditing wrote on 6/12/2005, 12:48 PM

it's the one that says "new" - he said that earlier, and I quit doing it when I hit 40 min. with my 3.2 HT 512 MB.

(it was getting close to done, but I think it would have gone a bit yet, lot of 3D in there.)

Dave

http://www.vasst.com/resource.aspx?id=aec79b30-5d7e-47dc-b5fa-eb02972deacc

That's the link to the direct file - it's the second one down; "Beta Test Render File - New"

Dave

GlennChan wrote on 6/12/2005, 3:45 PM

Oh ok. In practical terms, I don't think that benchmark is that useful because it will take at least two hours.

The old rendertest.veg would take around two minutes. Very practical for people who want to do a spot check (no pun intended) on their system to see if it's as fast as it should be. As well, there is at least many people reporting in results for it... so you can see which processors are fastest (unfortunately there are no results for the newest processors).

Anyways I will try that beta render test benchmark again. I stopped it when I noticed it was going to take a very long time. I though Spot got 2 minutes on his rig, which implied he was running a different benchmark.

2- Kind of related, but I think a lot of people would like to know how various processors stack up in Vegas performance:

Dual Xeons
Pentium D (dual core)
Pentium 6xx series
Pentium 5xx series
Pentium-M
(legacy) Pentium Northwood core

Dual core Opteron
Single core Opteron
AMD X2 (dual core)
AMD64 venice core
AMD64 san diego core
AMD64-mobile
(kind of legacy) those other AMD64 cores...
(legacy) Athlon XP

Of the list above, only a few of those processors have benchmarks for them.

johnmeyer wrote on 6/12/2005, 6:29 PM

I tried the test, but my computer hung (with Vegas 5.0d) about 25 minutes into the test. I let it run for another hour, but it didn't progress any further. I"ve never had problems rendering to a DV AVI file, so I assume it is a bug in Vegas or in the VEG.

Spot|DSE wrote on 6/12/2005, 7:01 PM

2 hours? I dunno what you mean by that, Glen. I just tested this on my VAIO laptop, and it rendered in 13 minutes.

That's a single, partitioned hard drive on a P4 3.06HT processor, 1 gig of RAM.
2 hours just doesn't make sense on any machine that's at least a 2GHz machine.

You're using the file found at:
http://www.vasst.com/resource.aspx?id=aec79b30-5d7e-47dc-b5fa-eb02972deacc ?

FrigidNDEditing wrote on 6/12/2005, 7:47 PM

How much does render time improve with RAM? is that a question that can be easily answered?

If not, you could just say substantial, or not too much.

Dave

Spot|DSE wrote on 6/12/2005, 8:27 PM

Not too much. Render time is decreased by having faster, separate drives or RAIDs, fast FSB, fast procs, and optimization of the application processes. Various combinations thereof can result in a wide variety of time increases/reductions.

GlennChan wrote on 6/12/2005, 8:55 PM

Ok now I am trying the proper file.
http://www.vasst.com/resource.aspx?id=aec79b30-5d7e-47dc-b5fa-eb02972deacc

I seem to have the same results as john is having... the render seems to hang. Mine kinda gets stuck at 80% then crawls along at a snail's pace.

2- Operons seem to be really, really fast at this render test. If you multiply your result by four, it's still faster than a 3.06ghz Pentium!

philfort wrote on 6/12/2005, 8:57 PM

Same here... hangs at 80%. I tried to render a piece of it after the 80% mark, and I don't think its actually hanging. Just rendering really slowly. About 1 frame every 5 minutes or so.

I have a 1.7Ghz P4.

GlennChan wrote on 6/12/2005, 9:16 PM

WTF?!

Ok tried this some more... Vegas 5 and 6 seems to choke in the parts after marker 3. Rendering crawls to a snail's pace.

I popped open Windows Task Manager:
Available RAM drops to a very low amount
PF size shoots up around 150MB.
the Vegas process is using 283MB RAM... RAM preview is set to 120MB, so it shouldn't be using that much.
It seems like Vegas gets stuck in some sort of bug, which is causing some page file or memory issues.
In draft mode, everything is fast fast fast.

Tried changing # of rendering threads to 1, doesn't seem to help.

FrigidNDEditing wrote on 6/12/2005, 9:29 PM

Mine crawled to a snails pace, but kept rendering. Don't know what the deal is. for you.

Dave

DavidMcKnight wrote on 6/12/2005, 11:09 PM

Snail's pace, indeed. Mine ground down at 80% but continued. I'm now at 82% after an hour and 17 mins. Using a 2.4 gHz P4, 512MB ram, Vegas 5.0d

EDIT
I let it run overnight and it's now at 94% after 6 hours 35 mins. Yikes!

Liam_Vegas wrote on 6/12/2005, 11:39 PM

Now at 85% after 3 hours... on a 3.2Ghz with 2GB RAM.

EDIT
After 3hrs 16min it was at 86% and the time to render dropped to zero... but it was still working hard. I cancelled the render (need to get on with other stuff). I loaded the partially complete AVI file onto the timeline and looking at where it got to that is part way in to a motion blur enevlope on the video bus track.

If it's not supposed to take 3 hours on a 3.2Ghz box.... it would be nice to know what's going on here. I don't think my box is running slow as far as rendering on other projects.

RNLVideo wrote on 6/13/2005, 4:10 AM

Mine too started to crawl at about 80%. I let it run over night and it took 5 hours and 39 minutes to complete. The first 80% was complete in 6 minutes.

This is on a 3.6G HT, 1gig of ram. Something doesn't seem right....

Rick

GlennChan wrote on 6/13/2005, 8:25 AM

Ok I ran that test overnight... 10:13:07. That's over 10 hours.

Pentium 2.6C, 512MB RAM.

jkrepner wrote on 6/13/2005, 8:27 AM

I think that is part of the Vegas 6 render bugs. I've not done any long renders since upgrading to 6.0b, but with Vegas 6 I would render to the 80% mark in a few minutes, then it would take about 5 or 6 hours hour to finish the rest.

Awesome!

Liam_Vegas wrote on 6/13/2005, 9:16 AM

80% is definitely right where the motion blur envelope begins.

Spot|DSE wrote on 6/13/2005, 1:53 PM

The Motion blur is where it's killed outright, depending on the system, how it's configured, etc.
For ease of comparisons, le's use the

I just tried the "new" test on a different machine here, and it froze at the Blur envelope, and while it's rendering on that machine, it's rendering about 1 frame every 30 seconds.

So, use the old for comparisons. I'll run the old test on the duallie machine tonight, and post the speed results both here, and in the Feedback form on the download page.

johnmeyer wrote on 6/13/2005, 2:19 PM

My results on the old render test, using a 2.8 GHz P4, no hyperthreading:

Vegas 6.0b 1:46
Vegas 5.0d 1:43

This was rendered to an NTSC DV AVI file.

Interesting that the older version of Vegas was the same (actually very slightly faster).

GlennChan wrote on 6/13/2005, 6:53 PM

I get the incredibly slow speed in Vegas 5 too.

It is very interesting that some systems are so incredibly slow on that benchmark.

rmack350 wrote on 6/13/2005, 7:01 PM

I was trying the test out this morning and decided to do it as prerenders because Vegas was using up around 800 MB with the straight render.

It seemed like, by doing it as prerenders, V6 would use less than 300MB for the first three segments. When it got to the last segment, where the Blur is, the memory usage started climbing. I stopped the render there since I could see it had slowed way down. Memory usage was around 500MB at that point.

I tried this to see if I could minimize the swap file usage. Seems to me like that works but I don't think it made anything any faster. Face it, if you've slowed down to 2 frames/minute, a swap file isn't going to slow you down. At that speed you could write each frame to a color printer and make a flip-book.

Rob Mack