Archived from groups: rec.video,rec.video.production,rec.video.desktop (More info?)

My theory: that rendering to a thumb drive should be faster than writing
to a hard drive, so should decrease render time.

I did a rather unscientific test this afternoon but the initial results
seem rather interesting. I rendered the same project twice. The first
time, the rendered file was created on my 64 Mb thumb drive. The second
time, the rendered file was created on a hard drive, but not the same
file that the source files were on. The first file rendered in
approximately 34 minutes, while the second trial took approximately 51
minutes. This is a huge difference, but again this was not a very
scientific test. The results are so good as to make me feel somewhat
skeptical. This is a pretty simple test, so I'm hoping that some other
people will perform the test on their own systems and share the results.
I'll include some of the information on my test below.

James

System: 1.5 Ghz Pentium IV, 640 Mb RAM, Windows XP Home.
Project: 8:47 minute video rendered to WMV file at 512 kbps. Project
contains approximately 100 clips, plus audio tracks and transitions.

Trial #1: Rendered to 64 Mb thumb drive connected via USB 1.1 port
Trial #2: Rendered to primary hard drive. Source files on second
physical hard drive.
Size of file produced. Trial #1: 30,463 kb. Trial #2: 30,469 kb.

Archived from groups: rec.video,rec.video.production,rec.video.desktop (More info?)

A second test show no significant difference. Oops!

Archived from groups: rec.video,rec.video.production,rec.video.desktop (More info?)

On Sat, 26 Jun 2004 04:04:09 GMT, James Messick
<jmessick@triad.rr.com> wrote:

>My theory: that rendering to a thumb drive should be faster than writing
>to a hard drive, so should decrease render time.


What's a thumb drive? One of those USB dongle thingies?

What makes you think it would be faster than a hard drive? Or that
drive access time is a limiting factor when rendering video?

Archived from groups: rec.video,rec.video.production,rec.video.desktop (More info?)

"Laurence Payne" <l@laurenceDELETEpayne.freeserve.co.uk> wrote in message
news:hhmrd01p81tmlho0j3ci730blhjt13b430@4ax.com...
> On Sat, 26 Jun 2004 04:04:09 GMT, James Messick
> <jmessick@triad.rr.com> wrote:
>
> >My theory: that rendering to a thumb drive should be faster than writing
> >to a hard drive, so should decrease render time.
>
>
> What's a thumb drive? One of those USB dongle thingies?
>
> What makes you think it would be faster than a hard drive? Or that
> drive access time is a limiting factor when rendering video?

Right now, the majority of my rendering is in real-time. You can get faster
than that?

Archived from groups: rec.video,rec.video.production,rec.video.desktop (More info?)

On Sat, 26 Jun 2004 14:08:12 -0700, "Tony" <tony23@dslextreme.com>
wrote:

>> What makes you think it would be faster than a hard drive? Or that
>> drive access time is a limiting factor when rendering video?
>
>Right now, the majority of my rendering is in real-time. You can get faster
>than that?

Sure. Why not? IF you're dealing with 25fps video, where's the law
that says a frame can't render in LESS than 1/25 sec?

I routinely render effects to audio files in considerably less time
than it would take to play them through. Video takes more computer
power, but that's an ever-increasing commodity.

Archived from groups: rec.video,rec.video.production,rec.video.desktop (More info?)

"James Messick" <jmessick@triad.rr.com> wrote in message
news:ZA6Dc.67640$2o2.4149064@twister.southeast.rr.com...
> My theory: that rendering to a thumb drive should be faster than writing
> to a hard drive, so should decrease render time.
>
> I did a rather unscientific test this afternoon but the initial results
> seem rather interesting. I rendered the same project twice. The first
> time, the rendered file was created on my 64 Mb thumb drive. The second
> time, the rendered file was created on a hard drive, but not the same
> file that the source files were on. The first file rendered in
> approximately 34 minutes, while the second trial took approximately 51
> minutes. This is a huge difference, but again this was not a very
> scientific test. The results are so good as to make me feel somewhat
> skeptical. This is a pretty simple test, so I'm hoping that some other
> people will perform the test on their own systems and share the results.
> I'll include some of the information on my test below.
>
> James
>
> System: 1.5 Ghz Pentium IV, 640 Mb RAM, Windows XP Home.
> Project: 8:47 minute video rendered to WMV file at 512 kbps. Project
> contains approximately 100 clips, plus audio tracks and transitions.
>
> Trial #1: Rendered to 64 Mb thumb drive connected via USB 1.1 port
> Trial #2: Rendered to primary hard drive. Source files on second
> physical hard drive.
> Size of file produced. Trial #1: 30,463 kb. Trial #2: 30,469 kb.
>


the thumb drive will never , ever run as fast as a hard disk. it is NAND
flash and requires more time to store and retrieve large data than a hard
disk. I suspect some of what you were seeing was because you had limited
your file size to 64MB.Nonetheless.. there are many other things that can
affect the test the way you have done it..

largely irrelevant I suppose since those devices hold so little...

Archived from groups: rec.video,rec.video.production,rec.video.desktop (More info?)

nappy wrote:
> "James Messick" <jmessick@triad.rr.com> wrote in message
> news:ZA6Dc.67640$2o2.4149064@twister.southeast.rr.com...
>
>>My theory: that rendering to a thumb drive should be faster than writing
>>to a hard drive, so should decrease render time.
>>
>>I did a rather unscientific test this afternoon but the initial results
>>seem rather interesting. I rendered the same project twice. The first
>>time, the rendered file was created on my 64 Mb thumb drive. The second
>>time, the rendered file was created on a hard drive, but not the same
>>file that the source files were on. The first file rendered in
>>approximately 34 minutes, while the second trial took approximately 51
>>minutes. This is a huge difference, but again this was not a very
>>scientific test. The results are so good as to make me feel somewhat
>>skeptical. This is a pretty simple test, so I'm hoping that some other
>>people will perform the test on their own systems and share the results.
>>I'll include some of the information on my test below.
>>
>>James
>>
>>System: 1.5 Ghz Pentium IV, 640 Mb RAM, Windows XP Home.
>>Project: 8:47 minute video rendered to WMV file at 512 kbps. Project
>>contains approximately 100 clips, plus audio tracks and transitions.
>>
>>Trial #1: Rendered to 64 Mb thumb drive connected via USB 1.1 port
>>Trial #2: Rendered to primary hard drive. Source files on second
>>physical hard drive.
>>Size of file produced. Trial #1: 30,463 kb. Trial #2: 30,469 kb.
>>
>
>
>
> the thumb drive will never , ever run as fast as a hard disk. it is NAND
> flash and requires more time to store and retrieve large data than a hard
> disk. I suspect some of what you were seeing was because you had limited
> your file size to 64MB.Nonetheless.. there are many other things that can
> affect the test the way you have done it..
>
> largely irrelevant I suppose since those devices hold so little...
>
>
Yep, I have large amounts of egg on my face for posting so quickly in my
excitement. This real-time rendering sounds pretty exciting, is that
done with a rendering card or something? My renders usually take 3 or 4
times real-time, minimum.

Archived from groups: rec.video,rec.video.production,rec.video.desktop (More info?)

On Sun, 27 Jun 2004 04:53:00 GMT, James Messick
<jmessick@triad.rr.com> wrote:

>Yep, I have large amounts of egg on my face for posting so quickly in my
>excitement. This real-time rendering sounds pretty exciting, is that
>done with a rendering card or something? My renders usually take 3 or 4
>times real-time, minimum.

Real-time rendering is nice, 'cos you can watch your result in
....er...real time :-)

If you're rendering off-line, considerably FASTER than real-time is
the aim. Why sit around for 90 minutes while a full-length movie
renders? If there's enough power to do it quicker, yes please!

Archived from groups: rec.video,rec.video.production,rec.video.desktop (More info?)

"James Messick" <jmessick@triad.rr.com> wrote in message
news:MosDc.75821$tH1.3581232@twister.southeast.rr.com...

>This real-time rendering sounds pretty exciting,

Faster than real time is better. You just need the horsepower. I've been
in editing rooms with extensive renderfarms, and I've run cinnerella with a
few zippy processors behind it.

There was a supercomputing project put on by a California college (I think
it was in CA, anyway) a few weeks ago where they asked people to bring in
thier computers. They ended up with about a thousand computers working on
the same project at a few terraflops of cpu power. One of the goals of the
software author was to enable people to do mpeg encoding on home networks.
The more machines you own - the faster the encoding.

Archived from groups: rec.video,rec.video.production,rec.video.desktop (More info?)

"twobirds" <notareal@eaddy.com> wrote in message
news:bcidnQO97diJ40LdRVn-gg@bresnan.com...
>
> "James Messick" <jmessick@triad.rr.com> wrote in message
> news:MosDc.75821$tH1.3581232@twister.southeast.rr.com...
>
> >This real-time rendering sounds pretty exciting,
>
> Faster than real time is better. You just need the horsepower. I've been
> in editing rooms with extensive renderfarms, and I've run cinnerella with
a
> few zippy processors behind it.
>
> There was a supercomputing project put on by a California college (I think
> it was in CA, anyway) a few weeks ago where they asked people to bring in
> thier computers. They ended up with about a thousand computers working on
> the same project at a few terraflops of cpu power. One of the goals of
the
> software author was to enable people to do mpeg encoding on home networks.
> The more machines you own - the faster the encoding.
>
>

Go figure... I sometimes wish I was in college so I could do things that
have already been done and claim some achievement...

S.E.T.I. already does distributed computing over thousands of desktop
computers..

REAT TIME MPeg rendering is a damn site cheaper than a cluster of PCs. Just
buy a hardware card and pipe video into it..

Archived from groups: rec.video,rec.video.production,rec.video.desktop (More info?)

"nappy" <no_spam_@sorry.com> wrote in message
news:8bVDc.638$YO.0@newssvr25.news.prodigy.com...

> Go figure... I sometimes wish I was in college so I could do things that
> have already been done and claim some achievement...

This is a little different

>
> S.E.T.I. already does distributed computing over thousands of desktop
> computers..

SETI is actually on a few million computers. United Devices is also on
several million (including all of mine)

>
> REAT TIME MPeg rendering is a damn site cheaper than a cluster of PCs.
Just
> buy a hardware card and pipe video into it..

Faster than real time is still better. :-) This project shows a little
promise for enabling people to do things like renderfarming.

Here it is. http://www.flashmobcomputing.org/ It is a bit different than
distributed computing like SETI and United Devices

Archived from groups: rec.video,rec.video.production,rec.video.desktop (More info?)

"twobirds" <notareal@eaddy.com> wrote in message
news:qvCdnb6H19DAwn3dRVn-vw@bresnan.com...
>
> "nappy" <no_spam_@sorry.com> wrote in message
> news:8bVDc.638$YO.0@newssvr25.news.prodigy.com...
>
> > Go figure... I sometimes wish I was in college so I could do things that
> > have already been done and claim some achievement...
>
> This is a little different
>
> >
> > S.E.T.I. already does distributed computing over thousands of desktop
> > computers..
>
> SETI is actually on a few million computers. United Devices is also on
> several million (including all of mine)
>
> >
> > REAT TIME MPeg rendering is a damn site cheaper than a cluster of PCs.
> Just
> > buy a hardware card and pipe video into it..
>
> Faster than real time is still better. :-) This project shows a little
> promise for enabling people to do things like renderfarming.
>
> Here it is. http://www.flashmobcomputing.org/ It is a bit different than
> distributed computing like SETI and United Devices
>
>

Thanks for the link.. I looked at it.. it's yak yak... noise.

Archived from groups: rec.video,rec.video.production,rec.video.desktop (More info?)

twobirds wrote:

> "James Messick" <jmessick@triad.rr.com> wrote in message
> news:MosDc.75821$tH1.3581232@twister.southeast.rr.com...
>
>
>>This real-time rendering sounds pretty exciting,
>
>
> Faster than real time is better. You just need the horsepower. I've been
> in editing rooms with extensive renderfarms, and I've run cinnerella with a
> few zippy processors behind it.
>
> There was a supercomputing project put on by a California college (I think
> it was in CA, anyway) a few weeks ago where they asked people to bring in
> thier computers. They ended up with about a thousand computers working on
> the same project at a few terraflops of cpu power. One of the goals of the
> software author was to enable people to do mpeg encoding on home networks.
> The more machines you own - the faster the encoding.
>
>
It would be nice if people could share other peoples computers over the
internet for rendering, but I think the bandwidth would be a limiting
factor right now.