CGTalk > Software Specific Forums > Maxon Cinema 4D
To minimize the ads you see on this page create a CGTalk account and log in HERE
Thread Closed share thread « Previous Thread | Next Thread »
 
Thread Tools Search this Thread Display Modes
Old 08-24-2013, 01:31 AM   #46
brasco
Polygone
 
brasco's Avatar
portfolio
Jon Bosley
United Kingdom
 
Join Date: Jun 2003
Posts: 1,725
Quote:
Originally Posted by ThePriest
The exponential curve that technological evolution progresses on


I'm fully in agreement, Moore's law is solid for another 15 or so years, I maybe should have clarified - it's the 475Ghz transistor that sounds great, but that doesn't translate to a 475Ghz CPU (sadly).

I think there'll be some really impressive progress in Quantum computing that'll supersede all the current tech before we see a 475Ghz graphene or a even SiGe CPU. Ofcourse, this is just the ramblings of an unenlightened muggle

cheers
brasc
__________________
brasco on vimeo
 
Old 08-26-2013, 09:03 PM   #47
penboack
Penboack
portfolio
Charles Kirk
United Kingdom
 
Join Date: Sep 2011
Posts: 103
Most technologies follow an S shaped curve. I don't think that workstation performance is doubling every 18 months. In fact looking at CB scores it seems we're getting an increase of around 50% in 18 months. There seems to have been a definite slow down over the last 3-5 years.

Now that the mass market has moved to mobile I don't see the driver to build ever more expensive fabs to produce more powerful CPUs that few people need. Time will tell, but I suspect we may be getting towards the top part of the S. Part of the driver behind Moore's law was the economics of building fabs.

I suspect that Apple are working with a 3D software vendor, the new Mac Pro seems very much designed for graphics given that it will have 2 GPUs and for the first time Pro GPUs at that.
__________________
deviantART Gallery
 
Old 08-26-2013, 09:27 PM   #48
ThePriest
Bad Ass
 
ThePriest's Avatar
portfolio
Stuart Townsend
Environment artist
Berkeley, USA
 
Join Date: Jun 2005
Posts: 2,545
A technology such as the silicon chip can of course reach limits determined by physical constraints. Moore's law has been very accurate in determining this. However the exponential curve of technological advancement, as discussed by the likes of Kurzweil, suggests that newer technologies are certain to take their place, and that they won't be too far off. We already have a handful of candidates technologies to replace the silicon chip.

A current gen 16 core machine still doubles the performance of a last gen 12 core, which says to me that things haven't slowed down quite as much as you've assumed. Those core counts are set to grow larger, though admittedly ghz per core is largely unchanged.

I do agree that cpu's have been pushed aside in favor of consumer electronics, to some degree, but I don't think we've been entirely neglected.
__________________
۩PRIST


Last edited by ThePriest : 08-26-2013 at 10:30 PM.
 
Old 08-27-2013, 12:45 AM   #49
LukeLetellier
Work in Progress
 
LukeLetellier's Avatar
portfolio
Luke Letellier
Portland, Maine, USA
 
Join Date: Aug 2012
Posts: 808
Quote:
Originally Posted by ThePriest

A current gen 16 core machine still doubles the performance of a last gen 12 core


Out of curiousity, which models are you comparing here? Tom's Hardware is showing a 50% increase.

http://www.tomshardware.com/reviews...iew,3149-8.html
 
Old 08-27-2013, 03:33 AM   #50
ThePriest
Bad Ass
 
ThePriest's Avatar
portfolio
Stuart Townsend
Environment artist
Berkeley, USA
 
Join Date: Jun 2005
Posts: 2,545
I maybe exaggerated with double the performance, but keep in mind it's not double the cores either, it's four extra. My old 12 core cinebenches in at about 14, my new machine gets almost 25. Comparing ghz, they're in the same range.
__________________
۩PRIST

 
Old 08-27-2013, 06:44 AM   #51
lllab
VRAYforC4D
 
lllab's Avatar
stefan Laub
VRAYforC4D_architect
vrayforc4d.com
Vienna, Austria
 
Join Date: Sep 2002
Posts: 3,771
yes preist is quite right i think, the cpu speed raised quite extreme last year with the new xeon E5, 16 core.
here from 16-17 to 25 in "one generation",which is a lot in one year, for similar price.

however it is not sure how long this will go on, and it does need a lot new challenges to support those many cpu cores in a good way. render software is pretty good in that, but also here a new tech is needed, only adding cores will not help on the long run.

the 2 gpus on the apple is for supporting the 2d gpu rendering on adobe products i assume, using amd chips indicate they are not targeted at 3d so much. i doubt apple does cooperate with a 3d software here. its more adobe and the like.
__________________
see vray on facebook: facebook/vrayforc4d
vray user gallery: http://tiny.cc/v6j5o

VRAYforC4D.com
3dtools.eu
rhino.io
 
Old 09-03-2013, 04:27 PM   #52
stoiQa
PRO
portfolio
stoiqa.com
LA, USA
 
Join Date: May 2008
Posts: 131
Send a message via AIM to stoiQa Send a message via Yahoo to stoiQa
getting a little better on my 2011 mac pro, I was hoping for better, we`ll see

[/URL]
__________________
My link
 
Old 09-03-2013, 04:35 PM   #53
Spaceland
Veteran
 
Spaceland's Avatar
portfolio
Denis Guay
Saint-hubert, Canada
 
Join Date: Feb 2007
Posts: 75
Quote:
Originally Posted by lllab
yes preist is quite right i think, the cpu speed raised quite extreme last year with the new xeon E5, 16 core.
here from 16-17 to 25 in "one generation",which is a lot in one year, for similar price.

however it is not sure how long this will go on, and it does need a lot new challenges to support those many cpu cores in a good way. render software is pretty good in that, but also here a new tech is needed, only adding cores will not help on the long run.

the 2 gpus on the apple is for supporting the 2d gpu rendering on adobe products i assume, using amd chips indicate they are not targeted at 3d so much. i doubt apple does cooperate with a 3d software here. its more adobe and the like.


Thru, they need a new technology in CPU. Adding more core to CPU is like extending the life on the current technology. In the end it won't be enough.

But the possibility of using CPU and GPU for rendering at the same time would be great. For every software.
__________________
[Denis]

| Coreldraw X6 | Cinema 4D R15 Prime | Lightwave 3D 11.6 | Moi 3D |


Google +
 
Old 09-10-2013, 09:22 PM   #54
ThePriest
Bad Ass
 
ThePriest's Avatar
portfolio
Stuart Townsend
Environment artist
Berkeley, USA
 
Join Date: Jun 2005
Posts: 2,545
Just a quick heads up. The new 12 core xeons are available today on newegg.


I was a little concerned about this information though; from this article

"it's worth noting that not all applications will support this many threads. I attempted to benchmark a Z820 (above) with two 12-core 2.7GHz Xeon E5-2697 v2 chips at an embargoed HP event in New York last week, and found that our Cinebench test application which is based on Cinema 4D R11.5 was unable to render with more than 32 threads, so showed no improvement over the kind of score I'd expect from the previous generation of chips."
__________________
۩PRIST

 
Old 09-10-2013, 09:37 PM   #55
wbj
Lord of the posts
 
Join Date: Oct 2003
Posts: 604
Quote:
Originally Posted by ThePriest
Just a quick heads up. The new 12 core xeons are available today on newegg.


I was a little concerned about this information though; from this article

"it's worth noting that not all applications will support this many threads. I attempted to benchmark a Z820 (above) with two 12-core 2.7GHz Xeon E5-2697 v2 chips at an embargoed HP event in New York last week, and found that our Cinebench test application which is based on Cinema 4D R11.5 was unable to render with more than 32 threads, so showed no improvement over the kind of score I'd expect from the previous generation of chips."


No reason to be concerned. Due to a bug in Windows, R11.5 couldn't use more than 32 threads by default, unless you would manually set more threads in the custom render settings (up 64) - this was an exclusive Windows problem (not existing on Linux or OS X).

R13 and newer use newer Windows API calls (where MS fixed that bug) and support up to 256 threads (which is also the limit of the Windows server editions).

Best regards,

wbj
 
Old 09-10-2013, 09:55 PM   #56
ThePriest
Bad Ass
 
ThePriest's Avatar
portfolio
Stuart Townsend
Environment artist
Berkeley, USA
 
Join Date: Jun 2005
Posts: 2,545
Is 256 threads a limit with R15's Team Render?
I have between 240-480 threads in mind for a studio based farm.
__________________
۩PRIST

 
Old 09-11-2013, 05:06 AM   #57
wbj
Lord of the posts
 
Join Date: Oct 2003
Posts: 604
Quote:
Originally Posted by ThePriest
Is 256 threads a limit with R15's Team Render?
I have between 240-480 threads in mind for a studio based farm.


Of course it's not. It's the limit of hardware cpu threads supported by Windows on a local machine (and again this is a limitation of Windows given by the limited structures of the Win32 api).

Best regards,

wbj
 
Old 09-11-2013, 02:36 PM   #58
rob rhodes
Arch Vizer
Rob Rhodes
United Kingdom
 
Join Date: Oct 2003
Posts: 583
The new 2650 v2 (8core 2.6ghz) looks like a good sweetspot, its about 200 per chip cheaper than the v1 of the same spec. I was just going to get or build another overclocked hex core machine soon but this is starting to tempt me more. Reckon I could build one for about 3K with 64GB RAM but without GPU as still weighing up if its worth Titaning it or not. Will have to wait for some benchmarks...
 
Old 09-16-2013, 01:10 PM   #59
ooo
actually my name is odo
 
ooo's Avatar
 
Join Date: May 2003
Posts: 2,724
Finally OSX Mavericks gets OpenGL 4.1 support and seems to get equal performance as on Windows 8

http://architosh.com/2013/09/maveri...r-new-mac-pros/

odo
 
Old 09-16-2013, 02:14 PM   #60
ThePriest
Bad Ass
 
ThePriest's Avatar
portfolio
Stuart Townsend
Environment artist
Berkeley, USA
 
Join Date: Jun 2005
Posts: 2,545
Quote:
Originally Posted by ooo
Finally OSX Mavericks gets OpenGL 4.1 support and seems to get equal performance as on Windows 8

http://architosh.com/2013/09/maveri...r-new-mac-pros/

odo


1300 was the recorded speed, comparable to a GTX 480-470.
A GTX 670 scores almost 3x as much. A GTX Titan boasts almost 6x that power.

An improvement, if you consider 2010 tech to still be viable hardware for 3D. But still an improvement. After all, I considered the GTX 285 a great card for 3D at the time.

You're also still limited to 12 low powered cores, which is probably okay if you have multiple machines and Team Render.

If Mac is truly your first choice, one Darth machine is where I would stop. If expansion is key, it doesn't matter what's on your network, fill it up with PC's. A 24 core workstation probably won't cost a great deal more than a the new Mac.
__________________
۩PRIST


Last edited by ThePriest : 09-16-2013 at 02:23 PM.
 
Thread Closed share thread


Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off
CGSociety
Society of Digital Artists
www.cgsociety.org

Powered by vBulletin
Copyright 2000 - 2006,
Jelsoft Enterprises Ltd.
Minimize Ads
Forum Jump
Miscellaneous

All times are GMT. The time now is 11:11 AM.


Powered by vBulletin
Copyright ©2000 - 2016, Jelsoft Enterprises Ltd.