Creative Communities of the World Forums

The peer to peer support community for media production professionals.

Activity Forums Apple Final Cut Pro Legacy Cluster rendering: what’s wrong?

  • Cluster rendering: what’s wrong?

    Posted by Paulo Jan on May 16, 2007 at 11:55 am

    [First of all, sorry for the 2 posts in a row. Is there any netiquette rule in this board about not posting too many separate messages, even if they are about completely unrelated subjects?]

    We did the other day a cluster rendering test. We have several dual G5s with 1.5 Gb. of RAM, linked together with 100Mbp Ethernet, and we configured 3 of them in an unmanaged QuickCluster (without touching anything, just leaving the default options). We sent a job to the cluster (a conversion from a FCP sequence to MPEG2 using Compressor) that takes usually 40 min. for one computer to render… and the 3 computers together took 51 minutes. Is this normal? Could the bottleneck be in the 100 Mbps. Ethernet?

    (Some additional notes: we’re using FCP 5.0.4 and Qmaster 2. The 3 computers weren’t running anything else that was CPU-intensive; in fact, according to the Activity Monitor, none of the 2 other computers (the ones that weren’t the master) were having more than 10% load).

    John Schroter replied 17 years, 12 months ago 5 Members · 5 Replies
  • 5 Replies
  • Jerry Hofmann

    May 16, 2007 at 12:12 pm

    You did exactly the right thing by posting separate questions in two threads, and you’re welcome to post as many questions as you’d like.

    Take a ride on the reading railroad! https://www.macworld.com/2006/01/secrets/marcchreateside/index.php is a dandy article on how to set up the cluster… Gigabit would be better, but not mandatory…

    Jerry

    Apple Certified Trainer

    Author: “Jerry Hofmann on Final Cut Pro 4” Click here

    Dual 2 gig G5, AJA Kona SD, AJA Kona 2, Huge Systems Array UL3D

  • Walter Biscardi

    May 16, 2007 at 12:16 pm

    [PaulJBis] “[First of all, sorry for the 2 posts in a row. Is there any netiquette rule in this board about not posting too many separate messages, even if they are about completely unrelated subjects?]”

    Nope, just rude to post multiple times about the same issue over a short period of time. 🙂

    Walter Biscardi, Jr.
    https://www.biscardicreative.com
    HD Editorial & Animation for Broadcast and independent productions.

    All Things Apple Podcast! https://cowcast.creativecow.net/all_things_apple/index.html

    Read my blog! https://blogs.creativecow.net/WalterBiscardi

  • Paulo Jan

    May 16, 2007 at 12:53 pm

    Thanks for the answer. The instructions in your link are for setting up a cluster manually (as opposed to a QuickCluster), which is what I did at first: configure a master computer, “services only” nodes, etc.; and the resuts I got were so disappointing (a 17 sec. clip took 2 minutes to export, vs. 40 sec. in a single computer) that I decided to set everything up as a QuickCluster and leave the default options. Apparently, it hasn’t worked too well either…

    Those of you using render clusters: are you all using Gigabit Ethernet? Is there any other trick I’m missing?

  • David Bogie

    May 16, 2007 at 2:46 pm

    > Is this normal? Could the bottleneck be in the 100 Mbps. Ethernet?< There is a QMaster forum over at apple.com. It's as useless as the software is worthless, though. There are only a few people in the world who have successfully gained anything from Qmaster and they do not post. Qmaster (like distributed rendering in After Effects, Aperture v1.0 and OpenGL) is a cruel joke. bogiesan This is my standard sigfile so do not take it personally: “For crying out loud, read the freakin’ manual.”

  • John Schroter

    May 30, 2007 at 2:44 am

    We had some changes to a piece I compressed last Friday and had to recompress the updated version in a hurry today, so I set up a quick cluster with a dual 2.6GHz G5 controller node and 2 QuadCore G5s. I followed the manual and it was set up in about 5 minutes.

    I compressed a 4 minute DVCPRO HD FCP clip to a 1920×1080 MPEG-2 multiplex’ed file using compressor and it took 20 minutes. When I compressed the same file on Friday from the dual 2.6 G5 it took 45 minutes. I don’t know what I did right or differently, but it worked as advertised. I set the networked computers as ‘service only’. We may be gigabit ethernet, but maybe not, since we’re tied into the corporate LAN as well. I mounted a common shared drive drive with the original media on all three systems – but I don’t think it will work anyway unless you do this.

    I initially thought it was taking longer than the original local compression because the status read 3% complete with nearly 2 hours remaining – until I switched the batch monitor window to show the cluster I had set up and realized it had burned through 90% of the compression in 15 minutes (there’s a drop-down window on the top right of the batch monitor window in compressor where you can check the status of each computer in the cluster including the local controller). Thinking back, I think I paused the compression on the controller machine and double-checked the service machines and then came back to the controller and started it again. Maybe this kicked the other processors up a notch or something, but I really don’t know. It’s discouraging to hear about the mixed results you’re all having, though. I thought I was really on to something.

    I tested the file and it looks great – and done in almost a third of the time. I was pretty impressed. I’m now looking for a way to harness Qmaster for AE renders. Any pointers? I don’t want to write custom command lines. I just want to send the render to the Quick Cluster I’ve set up. Probably wishful thinking, though.

We use anonymous cookies to give you the best experience we can.
Our Privacy policy | GDPR Policy