(1.37.109.15/16.2) id AA021095360; Thu, 17 Aug 1995 14:36:01 -0700
Return-Path: <owner-lightwave@mail.webcom.com>
Received: from kitten.mcs.com by webcom.webcom.com with ESMTP
(1.37.109.15/16.2) id AA020505320; Thu, 17 Aug 1995 14:35:21 -0700
Received: from mailbox.mcs.com (Mailbox.mcs.com [192.160.127.87]) by kitten.mcs.com (8.6.10/8.6.9) with SMTP id QAA04060 for <lightwave@webcom.com>; Thu, 17 Aug 1995 16:27:06 -0500
Received: by mailbox.mcs.com (/\==/\ Smail3.1.28.1 #28.5)
id <m0sjBsn-000jsaC@mailbox.mcs.com>; Thu, 17 Aug 95 15:49 CDT
Date: Thu, 17 Aug 95 15:49 CDT
From: johnc@mcs.com (John Crookshank)
To: lightwave@webcom.webcom.com (lightwave)
Subject: Re: Yet another PAR question...
Message-Id: <2977.6436T600T113@mcs.com>
X-Mailer: THOR 2.1� (TCP/IP)
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Lines: 53
Sender: owner-lightwave@webcom.com
Precedence: bulk
on 15-Aug-95 22:58:48, Jeric (Jeric@cup.portal.com) Emailed:
>>By the way, for better results, do not render frames 1-100 on the first
>>computer and frames 101-200 on the second. What works better is to tell
>>computer #1 to do frames 1-199, frame step = 2, so it does all the odd
>>frames, and tell computer #2 to do frames 2-200, frame step = 2, so it
>>does all the even frames. That way, the two computers stay fairly
>>neck-and-neck as far as the rendering go.
> Is this superior somehow to having them start at both ends and meet
> in the middle?
Well, I don't know if "superior" applies to _any_ of the methods, they will
all work OK. But I like to do it this way so that if I stop in the middle
to check things out, or if it's a long one that you have to do in stages,
the chances are quite high that you'll have a contiguous range of frames
finished, so you can record them and view them, etc.
Using the 1-100 and 101-200 method, if you stop part way through, you'll
have perhaps 1-50 and 101-150, my way you'll have 1-100. The other benefit
of doing it my way is when you have some parts of the animation take a lot
longer than others. A flying logo could easily go this way, with frame 1
taking 30 seconds, and frame 500 taking 30 minutes, since frame 1 is
basically empty, and frame 500 has a bunch of stuff that's moved on-screen.
Since both (or three or four) computers are all working on frames that are
not too far apart, they stay pretty "even".
Whenever I had a couple machines rendering using a 1-100, 101-200, 201-300
numbering scheme, I always had one or more machines that would get ahead of
the others because it's particular range of frames were easier than the
other groups. So they would end up sitting idle while the others worked.
Most irritating to come in the morning and find 3 computers sitting idle,
while the last one still had 50 frames to go...
Now of course, if your computers aren't evenly matched horsepower-wise,
you'll always have to guesstimate... "Let's see, this one is an 040, so
I'll give it 500 frames, this one is an 030, so I'll give it 200 frames."
You have problems in this scenario no matter _how_ you assign frame ranges.
That's why I eventually bought the Plantation software. It keeps track of
which computer is doing which frame, and whenever one machine finishes, the
Plantation software assigns it the next available frame to render. Once I
started using Plantation, I never again came into work and found 2 or more
computers sitting idle, with one or two still having a bunch of frames to
