Login with username, password and session length
Pages: [1] 2   Go Down
  Print  
Author Topic: ArtistX 64 bit?  (Read 2010 times)
3dz
Sr. Member
****

Cookies: 14
Offline Offline

Posts: 175



« on: July 12, 2009, 12:52:01 PM »

Does anybody know if ArtistX has a 64-bit version in the workings?  Huh
Logged
PGTips91
Sr. Member
****

Cookies: 36
Offline Offline

Posts: 238


When God spilled the paint.


WWW
« Reply #1 on: July 12, 2009, 08:06:54 PM »

Does anybody know if ArtistX has a 64-bit version in the workings?  Huh

Hi 3dz,

There does not seem to be much activity on the ArtistX forums and I cannot see anything about a 64-bit version in the future.

Of course, you could roll your own quite easily ...

I use debfoster to keep track on what is on my system(s) and it is fairly simple to use. You could start with Ubuntu, or Kubuntu 9.04, 64-bit version, add debfoster then update the 'keepers' file to the same as the latest ArtistX and run 'debfoster -f' to force compliance with the keepers file. Apart from the graphics, you would essentially have a 64-bit version of ArtistX.

I have attached the keepers file, compressed, from my ArtistX 0.6 installed. It may contain a few extras over and above what comes on the DVD, if you want to try going down this road. You will just need to drop off the .txt suffix to use it. Copy it and put it in as ' /var/lib/debfoster/keepers'.

Occasionally I output the current state of the keepers file with something like : --
Code:
cp /var/lib/debfoster/keepers  ~/Documents/DebfosterLists/KeepersList-2009-07-13

Paul
Logged

3dz
Sr. Member
****

Cookies: 14
Offline Offline

Posts: 175



« Reply #2 on: July 13, 2009, 01:42:33 PM »

Thanks Paul, Grin  this is something new.  Do you need to load the OS first, and place this keepers file in the ' /var/lib/debfoster/keepers' directory?

I just ordered the 64-bit version of Kubuntu 9.04.  I was going to try Remastery, and make a Blender Head Version. 
I heard about it from a post you had written.  I was thinking, Blender of course, The Gimp, MakeHuman,Voodoo, and Wings3d for now.
 
I have a feeling I'm forgetting something.  I'll have to wait till Kubuntu 9.04 cd comes, to see what I'm missing. 
But this may be a good alternative.  Cool
Logged
PGTips91
Sr. Member
****

Cookies: 36
Offline Offline

Posts: 238


When God spilled the paint.


WWW
« Reply #3 on: July 13, 2009, 02:45:32 PM »

Thanks Paul, Grin  this is something new.  Do you need to load the OS first, and place this keepers file in the ' /var/lib/debfoster/keepers' directory?

I just ordered the 64-bit version of Kubuntu 9.04.  I was going to try Remastery, and make a Blender Head Version.  
I heard about it from a post you had written.  I was thinking, Blender of course, The Gimp, MakeHuman, Voodoo, and Wings3d for now.
  
I have a feeling I'm forgetting something.  I'll have to wait till Kubuntu 9.04 cd comes, to see what I'm missing.  
But this may be a good alternative.  Cool

Well, with Linux, there are often many ways to reach your goal. I just happen to find debfoster simple and intuitive although I am led to believe that it is now depreciated in favour of apt-get and the like. Any way, here's how you can use debfoster to make any version of Linux conform to your preferred set of applications.

Starting with what you've been given by the distro, add debfoster. You can do this in Synaptic, or if you are happy with the CLI just do
Code:
sudo apt-get update && apt-get install debfoster
.

Next you run debfoster in quiet mode to produce the keepers file of the existing set-up.
Code:
sudo debfoster -q

This will produce the keepers list, which is found in the '/var/lib/debfoster/' directory. Since I have a memory problem and keep forgetting where the keepers file is stored, I usually end up finding it with
Code:
sudo find / -name keepers
which reminds me!

What you do next will depend a lot on how much change you wish to make to the installed applications base. If you just want to add a few applications, all you need to do is
Code:
sudo debfoster application1 application2 application3 application4 etc
and to remove applications simply append a trailing minus sign to an application name [with no space - like 'applicationx-' without the quotes].

You can get help by typing

Code:
debfoster --help
Usage: debfoster [-ck FILE] [-adefhinopqrsvV] package1 package2-
Installs package1, deinstalls package2

-v, --verbose                  Be a loudmouth
-V, --version                  Show version and copyright information
-h, --help                     Show this message
-q, --quiet                    Silently build keeper file
-f, --force                    Force system to conform to keeper file
-m, --mark-only                Do not install or delete packages
-u, --upgrade                  Try to upgrade dependencies
-c, --config FILE              Specify configuration file
-k, --keeperfile FILE          Specify keeper file
-n, --no-keeperfile            Don't read keeper file
-i, --ignore-default-rules     Ignore default rules
-a, --show-keepers             Show packages on keeper list
-s, --show-orphans             Show orphaned packages
-d, --show-depends PACKAGE     Show all depends of PACKAGE
-e, --show-dependents PACKAGE  Show dependents of PACKAGE
-p, --show-providers PACKAGE   Show packages providing PACKAGE
-r, --show-related PACKAGE     Show packages brought in by PACKAGE
-t, --use-tasks                Make tasks visible as packages
-o, --option OPT=VAL           Override any configuration option

See also: debfoster(8)
debfoster 2.7 -- Copyright (C) 2000,2001 Wessel Dankers.
Distributed under the GNU General Public License.

and a full manual by typing
Code:
man debfoster
which will produce a 313 line manual with more information than you will ever want to know about the application!

Oh, and I nearly forgot, when you have your keepers file set-up the way you want it, which you can do by directly editing the keepers file in place with your preferred text editor such as kate, nano, vi, vim or gedit, you issue the command for debfoster to force compliance with the keepers file and let it rip!
Code:
sudo debfoster -f
I must confess that I haven't yet tried that, but it would be fun to see!

The best way to learn is by doing, so you can try all this out in a LiveDVD session where mistakes are of no consequence. I have been fortunate enough to be able to build a test-bed computer on which to experiment without risk to my usual computing environment, so if you have or can build a cheap computer like that, that is a good way to learn, too. I am amazed at the freedom this has given me.

Hope this clarifies things enough. Just keep posting if you want more help.

By the way, there are some over on the #! forum who do this sort of thing with a script that they have customised to their own requirements. You may like to look at that method too. Here's a link http://crunchbanglinux.org/forums/post/26108/#p26108

Paul

PS I have ArtistX 0.7 installed on my test-bed machine so I can experiment with that and report back if that would help.

PPS Well I started down the track of an upgrade to Ubuntu 9.04, but stopped at the point where it said that there would be 2406 MB to download! I think I have remembered that number right.
« Last Edit: July 15, 2009, 12:11:54 AM by PGTips91 » Logged

3dz
Sr. Member
****

Cookies: 14
Offline Offline

Posts: 175



« Reply #4 on: July 16, 2009, 11:52:01 AM »

Thanks!  I got my Kubuntu disk today, but I think I'm going to wait to see what the market is going to do.  I'm leaning towards an AMD processor, but the market is sending out some weird vibes.

I think I'm going to exercise a little patience. 

It's kind of a shame, I have the case and power supply.
I was looking at a motherboard, but saw the market was pretty unstable.  Sad

Logged
PGTips91
Sr. Member
****

Cookies: 36
Offline Offline

Posts: 238


When God spilled the paint.


WWW
« Reply #5 on: July 16, 2009, 02:34:35 PM »

Thanks!  I got my Kubuntu disk today, but I think I'm going to wait to see what the market is going to do.  I'm leaning towards an AMD processor, but the market is sending out some weird vibes.

I think I'm going to exercise a little patience. 

It's kind of a shame, I have the case and power supply.
I was looking at a motherboard, but saw the market was pretty unstable.  Sad



Where are you getting your information from? I'm not great on the hardware side, but found this from a Google search.

http://www.motherboards.org/

If you have time on your side, this might be a good place to research.

Paul
Logged

3dz
Sr. Member
****

Cookies: 14
Offline Offline

Posts: 175



« Reply #6 on: July 16, 2009, 08:38:30 PM »


 Wink Thanks, nice site.  I like how they explain how things are going in the industry, as they go over the features.

I went through the site for about an hour and a half.  And I still like the motherboard that I have been eyeing up:

http://www.mwave.com/mwave/SkuSearch_v3.asp?SCriteria=BA31207

I guess I'm waiting for that miracle, that the i7 will drop in price.
It would be nice to run 8 threads in Blender.  Shocked
Logged
PGTips91
Sr. Member
****

Cookies: 36
Offline Offline

Posts: 238


When God spilled the paint.


WWW
« Reply #7 on: July 17, 2009, 01:55:21 AM »

I guess I'm waiting for that miracle, that the i7 will drop in price.
It would be nice to run 8 threads in Blender.  Shocked

I reckon that you need a stack of these in a cluster!


ServerPro

And here is another URL where you can compare blender speeds between systems : --
Blender Render Benchmark v0.2

Paul
« Last Edit: July 17, 2009, 01:58:53 AM by PGTips91 » Logged

3dz
Sr. Member
****

Cookies: 14
Offline Offline

Posts: 175



« Reply #8 on: July 17, 2009, 01:28:31 PM »

lol, that's what I am trying to stay away from.
I'm leaning towards more of the GPU. Nvidia has the right idea, just too expensive.
By comparison, ATi GPU's have more stream processors.

I've seen those benchmarks before.  Some of the results are little bit confusing.
compare  #102, to #108, and you see what I mean.

#102, by the looks of it is overclocking. He's only getting 4 threads. He's also indicating that he's using a 64-bit system, although he's got that data in the Blender version.  Huh

#108 by the looks of it is not overclocking, but has 8 threads. Also, he's not indicating whether he uses 64-bit or not.
 Huh  Huh 

If you go through those benchmarks, you'll find a lot of those inconsistencies.
Logged
PGTips91
Sr. Member
****

Cookies: 36
Offline Offline

Posts: 238


When God spilled the paint.


WWW
« Reply #9 on: July 17, 2009, 02:49:00 PM »

lol, that's what I am trying to stay away from.
I'm leaning towards more of the GPU. Nvidia has the right idea, just too expensive.
By comparison, ATi GPU's have more stream processors.

I've seen those benchmarks before.  Some of the results are little bit confusing.
compare  #102, to #108, and you see what I mean.

#102, by the looks of it is overclocking. He's only getting 4 threads. He's also indicating that he's using a 64-bit system, although he's got that data in the Blender version.  Huh

#108 by the looks of it is not overclocking, but has 8 threads. Also, he's not indicating whether he uses 64-bit or not.
 Huh  Huh 

If you go through those benchmarks, you'll find a lot of those inconsistencies.

True, I did find the results rather confusing and asking more questions than they answered.

What I would like to know, though, is whether clustering a group of low-cost computers does the job better than one very high-powered computer. The end result you are looking for is the fastest rendering for the least dollars. I don't know the answer to that, but, reading about the relative speeds and the corresponding costs, there does seem to be a diminishing return for dollars spent. To get a 10% improvement in rendering speed costs more than 10% in dollars spent. So, logically, massively increasing the parallel processing by low-cost systems, linked by software, seems to me to be the way to go to optimise the cost/benefit.

Of course, your ultimate goal has to be defined, first, before the design of the system to accomplish that goal can be derived. I am assuming that you want to render a motion picture, with its huge demand on rendering. That is where rendering farms have been successfully used. Any way, I am way out of my depth already ...

Paul
Logged

3dz
Sr. Member
****

Cookies: 14
Offline Offline

Posts: 175



« Reply #10 on: July 19, 2009, 09:35:06 AM »

Quote
What I would like to know, though, is whether clustering a group of low-cost computers does the job better than one very high-powered computer. The end result you are looking for is the fastest rendering for the least dollars. I don't know the answer to that, but, reading about the relative speeds and the corresponding costs, there does seem to be a diminishing return for dollars spent. To get a 10% improvement in rendering speed costs more than 10% in dollars spent. So, logically, massively increasing the parallel processing by low-cost systems, linked by software, seems to me to be the way to go to optimise the cost/benefit.

I would like to know the same thing, but there are some considerations before making a cluster.
One of the considerations that seems to go by unnoticed is area space.  Not to mention efficient powering.  Then there are the racks, to organize your cluster.  Then there are networking cards, and cables. Then you need to decide the speed of your network.  For instance, let's say you want a gigabit Lan. Well, then you need to go to category 6 cable.

So low cost systems, can add up to big money.
Even though you're using old computers that have been given to you.
  
Of course, if you have the space, and efficient powering.  (So you don't have extension cords all over the place.)
This can be very dangerous, it can be a fire hazard.
If this is all up to snuff.
Then this may be the way to go.
  
One of the things that I had been thinking about is: the GPU.

Another thing I've been thinking about is: single core laptops.
With the introduction of dual core and quad core. Single core laptops have become obsolete, and their prices have been falling.
Do a search on eBay for (refurbished laptops), and you'll see what I mean.

Quote
Of course, your ultimate goal has to be defined, first, before the design of the system to accomplish that goal can be derived. I am assuming that you want to render a motion picture, with its huge demand on rendering. That is where rendering farms have been successfully used. Any way, I am way out of my depth already.

Ultimately; I would like to confine a cluster to the smallest space possible, but still achieve the most bang for the buck.

For now, I would be happy to run a physics or a particle simulation.  The problem with learning or experimenting, is that render time it takes.  I find that it does not take that long to set up a physics or particle simulation.  It just seems it takes for ever to render.
When you're learning something new, you don't want to stop the flow.
I would like to have the power to do movie shorts.  

Motion pictures is pretty much cornered by the industry.
Besides, who has time to sit down and watch a two-hour movie anymore.  Not me.

I am out of my depth also, but I'm trying to learn.  Huh
« Last Edit: July 19, 2009, 09:38:07 AM by 3dz » Logged
PGTips91
Sr. Member
****

Cookies: 36
Offline Offline

Posts: 238


When God spilled the paint.


WWW
« Reply #11 on: July 19, 2009, 05:24:33 PM »

I would like to know the same thing, but there are some considerations before making a cluster ...
  
One of the things that I had been thinking about is: the GPU.

Ultimately; I would like to confine a cluster to the smallest space possible, but still achieve the most bang for the buck.

For now, I would be happy to run a physics or a particle simulation.  The problem with learning or experimenting, is that render time it takes.  I find that it does not take that long to set up a physics or particle simulation.  It just seems it takes for ever to render.
When you're learning something new, you don't want to stop the flow.
I would like to have the power to do movie shorts.  

I am out of my depth also, but I'm trying to learn.  Huh

Hi 3dz,

I'm thinking back to the original thread about a small-form cluster that someone built.
A linux cluster in a IKEA Helmer cabinet.

The author concludes his account of building this cluster with this quote : --
Quote
The most amazing is that this machine just cost as a better standard PC, but has 24 cores that run each at 2.4 Ghz, a total of 48GB ram, and just need 400W of power!! This means that it hardly gets warm, and make less noise then my desktop pc.

Render jobs that took all night, now gets done in 10-12 min.


I think that this guy is onto a second-generation project now, so it might be worth your while to get in touch with him and see what he can tell you about building a cost-effective cluster, both in initial outlay, and in terms of ongoing running costs. I think that his system was low-energy, small form-factor but with the computing power needed and at a moderate cost.

The story of Helmer II
Quote
This is the story of Helmer II
"The Renderer"
(hoping to do:)

Home Page of SFE - Svensk Film Effekt

My understanding about the GPU is that it confines itself to the display, so it is not going to have all that much effect on the ability of the system to do rendering, which would be entirely a CPU function. It would help by not making demands on the CPU for the display, but that would be all, I think. Of course, I could be quite wrong on that, and reading about Helmer II seems to give a different impression ...

You might benefit from taking a look at Dynebolic, too, as it has special tools for setting up and using a cluster of all computers on a LAN. It is also geared towards creative stuff, such as films or broadcasting, etc. I think the main base of users come from countries where they don't have access to lots of money, so they are geared towards using the lowest cost solutions.

Here is their how-to : --
Cluster computer farms

Not that theirs is the only way to set up a cluster, but it may be one of the easier ways to begin with.

Paul
« Last Edit: July 19, 2009, 05:48:44 PM by PGTips91 » Logged

lunoob
Hero Member
*****

Cookies: 44
Offline Offline

Posts: 1196


Old 'LGU' Guys Rule.


« Reply #12 on: July 19, 2009, 08:16:13 PM »

Paul, you're pretty amazing - finding all that.  Thanks for sharing it!

+1   Smiley
Logged

Ssh!  I'm trying to concentrate!!
PGTips91
Sr. Member
****

Cookies: 36
Offline Offline

Posts: 238


When God spilled the paint.


WWW
« Reply #13 on: July 20, 2009, 01:45:48 AM »

Paul, you're pretty amazing - finding all that.  Thanks for sharing it!

+1   Smiley

And thank you for your appreciation! I ain't really amazing, I am just a reporter of what others have done and discovered.

I noticed that my 'sig' had an incomplete reference to the Linux Documentation Project, so I fixed that and while I was there I did a search on 'clusters' which brought up this detailed and helpful account of how one person/group have done this, with some really interesting ideas about how to set it all up [such as making a bootable disk that will clone Linux onto each and every node of the cluster automatically while you walk away and do something else!!], selection of hardware, etc. I've only just started to read it so can't add much more, but anyone wanting to set up a Linux cluster should probably read it as a compilation of lots of research that would be hard to find scattered all over the Net.

http://tldp.org/HOWTO/Cluster-HOWTO-1.html

Paul

PS Am looking at another HowTo, and, like the one above, the information is rather dated. [http://tldp.org/HOWTO/Parallel-Processing-HOWTO-3.html]

PPS Both bandwidth and latency are important considerations with a cluster. Multiple, load-sharing networks can vastly improve networking at marginal costs.

PPPS Clustering is better supported by something like Myrinet than Ethernet. [High performance, High cost, too!]
Quote
Myrinet is a cost-effective, high-performance, packet-communication and switching technology that is widely used to interconnect clusters of workstations, PCs, servers, blade servers, or single-board computers.
Clusters provide an economical way of achieving:

    * high performance, by distributing demanding computations across an array of cost-effective hosts.
          For "tightly coupled" distributed computations, the interconnect must provide high-data-rate and low-latency communication between host processes.
    * high availability, by allowing a computation to proceed with a subset of the hosts.
          The interconnect should be capable of detecting and isolating faults, and of using alternative communication paths.

PPPPS The site that seems to have inherited the mantle for many older sites is http://www.linuxhpc.org/
Quote
The #1 Site for News & Information Related to Linux High Performance Technical Computing, Linux High Availability and Linux Parallel Clustering

PPPPPS Looks as if I was completely wrong about GPUs. These apparently allow much denser clusters than CPU-only based clusters. See here
          Colfax International Unveils NVIDIA Tesla GPU Computing Clusters
« Last Edit: July 20, 2009, 05:44:46 AM by PGTips91 » Logged

rji
Hero Member
*****

Cookies: 38
Offline Offline

Posts: 405


« Reply #14 on: July 21, 2009, 10:59:52 AM »

It just seems it takes for ever to render.

Even Pixar has this problem:  http://www.byteandswitch.com/storage/infrastructure/pixar.php

Quote
Motion pictures is pretty much cornered by the industry.

Not really, the "Killer Bean" series of animated films is written, modelled, animated, rendered, produced, by one guy.  http://www.killerbeanforever.com/ one person can do it all.
« Last Edit: July 21, 2009, 11:24:50 AM by rji » Logged
Pages: [1] 2   Go Up
  Print  
 
Jump to: