Digital Art Zone

 
   
5 of 8
5
OctaneRender for Daz Studio
Posted: 25 September 2012 03:52 PM   [ Ignore ]   [ # 61 ]
Active Member
Avatar
RankRank
Total Posts:  793
Joined  2012-02-12

My thought exactely I’d prefer to buy ATI as their card have more GPU power for rendering and NVIDIA limited their last cards so got to buy an old one. To get it correct you should buy a 580 Gtx with 3 Gigs of ram.

I may do the jump one day but I don’t use unbiased renderers that much for now. I have the time to wait and see, espescially curious about cycles render engine

 

 Signature 

My Deviant Art

Profile
 
 
Posted: 25 September 2012 04:08 PM   [ Ignore ]   [ # 62 ]
Power Member
Avatar
RankRankRank
Total Posts:  2302
Joined  2010-12-18
Takeo.Kensei - 25 September 2012 03:52 PM

My thought exactely I’d prefer to buy ATI as their card have more GPU power for rendering and NVIDIA limited their last cards so got to buy an old one. To get it correct you should buy a 580 Gtx with 3 Gigs of ram.

I may do the jump one day but I don’t use unbiased renderers that much for now. I have the time to wait and see, espescially curious about cycles render engine

Yeah I read the 580 GTX were best for this but I couldn’t find one locally. I prefer computer parts from local stores so I went with a GTX 660 with 3gigs of ram. There were some 500 series cards around but they only had 1.5 or 2 gigs of ram.

unfortunately when I uninstalled my video card drivers it killed my both my network card drivers so now I’m doing a system restore :( Haven’t been able to test out the new card yet. Either way it will be better than that 8800gts. I really like my ati card so I’ll just put it in my new living room rig so that when I play driving games on my racing wheel it will be even better. I don’t really play games too much anymore though.

 Signature 

My deviantArt Pinup Gallery

Profile
 
 
Posted: 25 September 2012 04:30 PM   [ Ignore ]   [ # 63 ]
Addict
Avatar
RankRankRankRank
Total Posts:  4668
Joined  2004-08-11

I’m an NVIDIA fan boy but didn’t hear about the 580 GTX being the best of the best.  If that will help speed up render time I’m there.  I think they are way past the 500 series at this point though, aren’t they? 

Oh, just went to Amazon and looked at that card.  I hate those long ass cards.  They but up against the back of hard drives which produce bad heat.  I don’t like that design at all.  NEXT!

 Signature 

    —- Wolff On The Prowl—-

Pan Project

Profile
 
 
Posted: 25 September 2012 04:36 PM   [ Ignore ]   [ # 64 ]
Power Member
Avatar
RankRankRank
Total Posts:  2302
Joined  2010-12-18
RAMWolff - 25 September 2012 04:30 PM

I’m an NVIDIA fan boy but didn’t hear about the 580 GTX being the best of the best.  If that will help speed up render time I’m there.  I think they are way past the 500 series at this point though, aren’t they? 

Oh, just went to Amazon and looked at that card.  I hate those long ass cards.  They but up against the back of hard drives which produce bad heat.  I don’t like that design at all.  NEXT!

they way my cases are I don’t have that issue fortunately. but i’ve got some big ole cases. the card I just got has the power connectors on the site which helps too.

 Signature 

My deviantArt Pinup Gallery

Profile
 
 
Posted: 25 September 2012 04:39 PM   [ Ignore ]   [ # 65 ]
Power Member
Avatar
RankRankRank
Total Posts:  1870
Joined  2006-02-17

Folks, if yer gonna drool then DROOL!

http://www.nvidia.com/object/tesla-servers.html

Tesla K10—3072 CUDA Cores. 8GB RAM.
Quadro K5000—1344 CUDA Cores, 4GB RAM

EDIT:  BTW, unlike the GTX series, these have the FULL Floating-Point pipeline open, not half of it.  And no, softmodding will not activate the FP Pipelines.

cheese

Kendall

 Signature 

Any opinions expressed in this post are those of Kendall Sears and may, or may not, be more, or less, valid than any other opinion.

The contents of this post are intended for the DAZ forum only, do not re-post any portion to any other forum without his permission.

Profile
 
 
Posted: 25 September 2012 04:40 PM   [ Ignore ]   [ # 66 ]
Power Member
Avatar
RankRankRank
Total Posts:  2302
Joined  2010-12-18

i went from 0, to 96 to ~1300 in about 24 hours. Im ok for now. LOL

 Signature 

My deviantArt Pinup Gallery

Profile
 
 
Posted: 25 September 2012 04:43 PM   [ Ignore ]   [ # 67 ]
Member
Avatar
Rank
Total Posts:  39
Joined  2003-10-09
RAMWolff - 25 September 2012 04:30 PM

I’m an NVIDIA fan boy but didn’t hear about the 580 GTX being the best of the best.  If that will help speed up render time I’m there.  I think they are way past the 500 series at this point though, aren’t they? 

Oh, just went to Amazon and looked at that card.  I hate those long ass cards.  They but up against the back of hard drives which produce bad heat.  I don’t like that design at all.  NEXT!

The cards aren’t too long, your tower is too small wink

The 600 series is out. But unfortunately, these are cards for gamers. CUDA processing power is lower than the previous generation ( which is why Octane users are still using 5xx cards ).

I would love to get a really good Octane export plugin. It is true that you are limited by your graphic card meomory but Octane wasn’t designed for Studio users. It is perfect for photorealistic renderign and architecture. Nonetheless, I’ve seen amazing results made by Studio users…

Profile
 
 
Posted: 25 September 2012 04:47 PM   [ Ignore ]   [ # 68 ]
Addict
Avatar
RankRankRankRank
Total Posts:  4545
Joined  2007-09-13

In order to actually use one of those big video cards, a FULL tower case is essential.

That’s all there is to it.  Those cards were not designed to fit into mid or smaller cases.

Several reasons, with the most obvious one being that with that ‘level’ video card, the rest of the system is usually geared toward performance…and in such systems, it is a lot easier to keep them cool if the whole thing is built around a full tower case.

Kendall Sears - 25 September 2012 04:39 PM

Folks, if yer gonna drool then DROOL!

http://www.nvidia.com/object/tesla-servers.html

Tesla K10—3072 CUDA Cores. 8GB RAM.
Quadro K5000—1344 CUDA Cores, 4GB RAM

EDIT:  BTW, unlike the GTX series, these have the FULL Floating-Point pipeline open, not half of it.  And no, softmodding will not activate the FP Pipelines.

cheese

Kendall

And at the low price of $3400…I’ll take two.

 Signature 

1432 old posts

My ShareCG gallery.

Just because something costs a lot, doesn’t mean it’s the best…

It just means it’s expensive.

Profile
 
 
Posted: 25 September 2012 04:49 PM   [ Ignore ]   [ # 69 ]
Power Member
Avatar
RankRankRank
Total Posts:  2302
Joined  2010-12-18
mjc1016 - 25 September 2012 04:47 PM

In order to actually use one of those big video cards, a FULL tower case is essential.

That’s all there is to it.  Those cards were not designed to fit into mid or smaller cases.

Several reasons, with the most obvious one being that with that ‘level’ video card, the rest of the system is usually geared toward performance…and in such systems, it is a lot easier to keep them cool if the whole thing is built around a full tower case.

Kendall Sears - 25 September 2012 04:39 PM

Folks, if yer gonna drool then DROOL!

http://www.nvidia.com/object/tesla-servers.html

Tesla K10—3072 CUDA Cores. 8GB RAM.
Quadro K5000—1344 CUDA Cores, 4GB RAM

EDIT:  BTW, unlike the GTX series, these have the FULL Floating-Point pipeline open, not half of it.  And no, softmodding will not activate the FP Pipelines.

cheese

Kendall

And at the low price of $3400…I’ll take two.

a good tower will outlast the hardware in it a few generations over as well. I’ve had my main for 8 or so years now so 3-4 generations of guts.

 Signature 

My deviantArt Pinup Gallery

Profile
 
 
Posted: 25 September 2012 05:10 PM   [ Ignore ]   [ # 70 ]
Addict
Avatar
RankRankRankRank
Total Posts:  4545
Joined  2007-09-13

Yeah…that too.

 Signature 

1432 old posts

My ShareCG gallery.

Just because something costs a lot, doesn’t mean it’s the best…

It just means it’s expensive.

Profile
 
 
Posted: 25 September 2012 05:27 PM   [ Ignore ]   [ # 71 ]
Member
Rank
Total Posts:  52
Joined  2011-09-11

the 670 is the best price / cuda value
the 680 3gb is the best for scene memory

the 690 3gb has 2x the speed of the 680 but shares the memory so it’s 1.5

power is something to consider if you want more than one graphic card.

A 15 min render, 3000 samples, pmc - no post work
for the best light use volumetric but it can triple the render time

I would look at http://raytracey.blogspot.com for information on the other OTOY product, for me I’m more interested in motion so i my standards my be a bit lower when it comes to the perfect render.

Image Attachments
09.png
Profile
 
 
Posted: 25 September 2012 06:39 PM   [ Ignore ]   [ # 72 ]
Power Member
Avatar
RankRankRank
Total Posts:  2302
Joined  2010-12-18

i’m finally up and running, ATI CCC gave me hell for 2 hours. Used CCleaner and Driver fusion and got the mess cleaned up. Anyway found the kepler version of the software after I realized the 600series isn’t supported yet. Things seem ok so far, now I just need to learn a ton of stuff smile like how to add a light lol

I’m watching the training videos now.

 Signature 

My deviantArt Pinup Gallery

Profile
 
 
Posted: 25 September 2012 07:41 PM   [ Ignore ]   [ # 73 ]
Addict
Avatar
RankRankRankRank
Total Posts:  4668
Joined  2004-08-11
joelegecko - 25 September 2012 04:43 PM
RAMWolff - 25 September 2012 04:30 PM

I’m an NVIDIA fan boy but didn’t hear about the 580 GTX being the best of the best.  If that will help speed up render time I’m there.  I think they are way past the 500 series at this point though, aren’t they? 

Oh, just went to Amazon and looked at that card.  I hate those long ass cards.  They but up against the back of hard drives which produce bad heat.  I don’t like that design at all.  NEXT!

The cards aren’t too long, your tower is too small wink

The 600 series is out. But unfortunately, these are cards for gamers. CUDA processing power is lower than the previous generation ( which is why Octane users are still using 5xx cards ).

I would love to get a really good Octane export plugin. It is true that you are limited by your graphic card meomory but Octane wasn’t designed for Studio users. It is perfect for photorealistic renderign and architecture. Nonetheless, I’ve seen amazing results made by Studio users…

I have a medium sized case, I don’t need a huge tower, nor do I want something monstrous in my little computer room.  The rig uses a full sized Mb… so not sure why my case is too small, it’s industry standard gaming rig that I use for graphics…..

 Signature 

    —- Wolff On The Prowl—-

Pan Project

Profile
 
 
Posted: 25 September 2012 08:01 PM   [ Ignore ]   [ # 74 ]
Addict
Avatar
RankRankRankRank
Total Posts:  4545
Joined  2007-09-13
RAMWolff - 25 September 2012 07:41 PM

I have a medium sized case, I don’t need a huge tower, nor do I want something monstrous in my little computer room.  The rig uses a full sized Mb… so not sure why my case is too small, it’s industry standard gaming rig that I use for graphics…..

Back in ancient times (Window 3.x/DOS)  there was basically 1 sized case…the ‘full tower’.  It was the ‘industry standard’...when Win 95 came around cases started getting smaller and other categories were invented.  But as far as it goes, the ‘full tower’ is still the ‘standard’ and everything else is adjusted down from there.  And really, a full tower isn’t that much bigger (yes mid-towers come in small, medium and large flavors) than the ‘real’ mid-tower spec, but some times, that little extra does matter.  Like with a real ‘full sized’ video card…

 Signature 

1432 old posts

My ShareCG gallery.

Just because something costs a lot, doesn’t mean it’s the best…

It just means it’s expensive.

Profile
 
 
Posted: 25 September 2012 08:17 PM   [ Ignore ]   [ # 75 ]
Power Member
Avatar
RankRankRank
Total Posts:  1870
Joined  2006-02-17
mjc1016 - 25 September 2012 04:47 PM

In order to actually use one of those big video cards, a FULL tower case is essential.

That’s all there is to it.  Those cards were not designed to fit into mid or smaller cases.

Several reasons, with the most obvious one being that with that ‘level’ video card, the rest of the system is usually geared toward performance…and in such systems, it is a lot easier to keep them cool if the whole thing is built around a full tower case.

Kendall Sears - 25 September 2012 04:39 PM

Folks, if yer gonna drool then DROOL!

http://www.nvidia.com/object/tesla-servers.html

Tesla K10—3072 CUDA Cores. 8GB RAM.
Quadro K5000—1344 CUDA Cores, 4GB RAM

EDIT:  BTW, unlike the GTX series, these have the FULL Floating-Point pipeline open, not half of it.  And no, softmodding will not activate the FP Pipelines.

cheese

Kendall

And at the low price of $3400…I’ll take two.

I didn’t say they were cheap.  But it is something to dream about.  No, I don’t have any… yet.  I have seen them in action though.  VERY very fast!  And I would like to have a few, but there’s no justification yet. :-(

Since these both have the full floating-point pipelines available the CUDA cores run up to 300% faster each than the CUDA cores on the GTX series.

Here are some numbers of a softmodded Quadro-FX 4600 compared to a real Quadro FX 1700.  As most will know the FX-1700 is a LOW END Quadro.  Imagine the numbers from the K-10 (and the new K-20…)

Software           Softmod 4600           Real 1700
                    (128 CUDA)              (32 CUDA)

3ds Max               16.13                     37.12
CATIA                 15.38                     38.85
EnSight               19.46                     26.32
Maya                   30.13                     60.00
Pro/ENGINEER       13.16                     36.43
SolidWorks             23.42                   48.20

Kendall

 Signature 

Any opinions expressed in this post are those of Kendall Sears and may, or may not, be more, or less, valid than any other opinion.

The contents of this post are intended for the DAZ forum only, do not re-post any portion to any other forum without his permission.

Profile
 
 
   
5 of 8
5