OTOY Forums

hmmmmm... please make an update for the 3dsmax plug-in :/ its not moving !!

Hi wrapthereal66,
have you already tried with this version?
viewtopic.php?f=80&t=55377
ciao beppe

WOW ! thanks you very much !

Why people is talking about optimization? How realistic is to hope in magic optimizations?
If history have some value just take a look at the official OctaneBench scores, you can cleary see that after 4 years a gtx 680 kepler architecture (which bring 3.0 Tflops) has a -20% of performance compared to the good old gtx 580 (fermi with 1,6 Tflops)...so where's the optimization?

When 9xx series entered in the scene we saw the exact same performance that we have now, they just fixed the memory performance problem related to the 970s (which initially performed very bad).

I don't think it is a good idea to hope on future optimizations...just look at the bad reality: the miracle that nvidia were claim with 1xxx gpus is just a marketing illusion.

erbombo wrote:Why people is talking about optimization? How realistic is to hope in magic optimizations?
If history have some value just take a look at the official OctaneBench scores, you can cleary see that after 4 years a gtx 680 kepler architecture (which bring 3.0 Tflops) has a -20% of performance compared to the good old gtx 580 (fermi with 1,6 Tflops)...so where's the optimization?

When 9xx series entered in the scene we saw the exact same performance that we have now, they just fixed the memory performance problem related to the 970s (which initially performed very bad).

I don't think it is a good idea to hope on future optimizations...just look at the bad reality: the miracle that nvidia were claim with 1xxx gpus is just a marketing illusion.

There is no magic bullet when it comes to render times/GPU performance. If you're looking for a massive catalyst, you won't find it in optimizations. You're better off building a time machine or waiting a handful of years for the technology to drastically improve. Personally, I went from a GTX 780 Ti to a GTX 1080. The difference is day and night.

The performance we are seeing with the Pascal cards now, without optimization, is tenable.
Bear in mind that this is on a hybrid CUDA 7.5 & 8.0 RC build for Octane that we asked the Octane team to build in the interim.

Once the full launch of CUDA 8.0 hits, we will see a new stable release of Octane 3.0 with full Pascal support (optimized).
There is no reason to believe that optimizations won't bring some noticeable performance boost. Unless, of course, you're some kind of pessimist. Time will tell. I have faith in the Octane Team.

zeigg wrote:

erbombo wrote:Why people is talking about optimization? How realistic is to hope in magic optimizations?
If history have some value just take a look at the official OctaneBench scores, you can cleary see that after 4 years a gtx 680 kepler architecture (which bring 3.0 Tflops) has a -20% of performance compared to the good old gtx 580 (fermi with 1,6 Tflops)...so where's the optimization?

When 9xx series entered in the scene we saw the exact same performance that we have now, they just fixed the memory performance problem related to the 970s (which initially performed very bad).

I don't think it is a good idea to hope on future optimizations...just look at the bad reality: the miracle that nvidia were claim with 1xxx gpus is just a marketing illusion.

There is no magic bullet when it comes to render times/GPU performance. If you're looking for a massive catalyst, you won't find it in optimizations. You're better off building a time machine or waiting a handful of years for the technology to drastically improve. Personally, I went from a GTX 780 Ti to a GTX 1080. The difference is day and night.

The performance we are seeing with the Pascal cards now, without optimization, is tenable.
Bear in mind that this is on a hybrid CUDA 7.5 & 8.0 RC build for Octane that we asked the Octane team to build in the interim.

Once the full launch of CUDA 8.0 hits, we will see a new stable release of Octane 3.0 with full Pascal support (optimized).
There is no reason to believe that optimizations won't bring some noticeable performance boost. Unless, of course, you're some kind of pessimist. Time will tell. I have faith in the Octane Team.

Again, the reason to don't be too positive is just in the history of last 5 years of new cards/new cuda. Anyway if you are happy with your new buy it is ok, I'm not saying people that buy 1080s are stupid just because they spent 700-800$ on a videocard that goes less than a 980ti, the reason of my post is because people may read and hope on a miracle optimization (that in fact it never happens in 5 years when it comes on old cards). I think, if people don't have the hurry to buy, it is better to wait and see how these Pascal cards go on gpu rendering. Maybe this time is different (I hope), but as you say "time will tell"

I´m actuallty considering between a couple of 1070 or 1080 - somebody may "know" how much the performance difference is between those two?! Of course at rendering with octane not gaming ^^
Not for sure by now if the price difference (1080) is worth the more in performance =/ Especially both have 8G VRAM

Do Pascal GPUs work now ? What about the (C4D) plugins ?

FlowFireGames wrote:Do Pascal GPUs work now ? What about the (C4D) plugins ?

Yes they work from v3.03.2, here is the latest c4dOctane version:
viewtopic.php?f=85&t=54672
ciao beppe

zeigg wrote:

erbombo wrote:Why people is talking about optimization? How realistic is to hope in magic optimizations?
If history have some value just take a look at the official OctaneBench scores, you can cleary see that after 4 years a gtx 680 kepler architecture (which bring 3.0 Tflops) has a -20% of performance compared to the good old gtx 580 (fermi with 1,6 Tflops)...so where's the optimization?

When 9xx series entered in the scene we saw the exact same performance that we have now, they just fixed the memory performance problem related to the 970s (which initially performed very bad).

I don't think it is a good idea to hope on future optimizations...just look at the bad reality: the miracle that nvidia were claim with 1xxx gpus is just a marketing illusion.

There is no magic bullet when it comes to render times/GPU performance. If you're looking for a massive catalyst, you won't find it in optimizations. You're better off building a time machine or waiting a handful of years for the technology to drastically improve. Personally, I went from a GTX 780 Ti to a GTX 1080. The difference is day and night.

The performance we are seeing with the Pascal cards now, without optimization, is tenable.
Bear in mind that this is on a hybrid CUDA 7.5 & 8.0 RC build for Octane that we asked the Octane team to build in the interim.

Once the full launch of CUDA 8.0 hits, we will see a new stable release of Octane 3.0 with full Pascal support (optimized).
There is no reason to believe that optimizations won't bring some noticeable performance boost. Unless, of course, you're some kind of pessimist. Time will tell. I have faith in the Octane Team.

Care to post some test results of 1080 (ideally at 2GHz clock) vs 780Ti? I have 780Ti at work and was looking at 1080, even possibly 2, so i am really curious how they compare.

This whole Titan X thing gives me headache. Nvidia really likes to complicate everything. If they kept Titan for next year, i would just get 1080(s) now and be happy with them, as they would be the best things one can currently get. No matter what happens next year, its obvious something faster is going to come out eventually. But 1080 was basically 2 months straight not on stock, bar those FE versions i did not want. So i waited and waited and now they released TitanX and i dont know what to do. I am on limited budget, so 2 Titans are out of question, but somehow, i dont want 1080s anymore... Seems the best thing to do is to wait for 1080Ti, but god, i was waiting to get new GPU since last summer, when my 590 broke, and even my patience has limits...