Chaos implemented OpenCL and Cuda.
With Vray-RT Now you can do distributed rendering with Cuda. It's the first software on the planet that can do this.
BTW Octane is more photorealistic. Vray-RT renders are renders, Octane renders are photos.
WE NEED AMD ATI GPU SUPPORT
-
gabrielefx
- Posts: 1701
- Joined: Wed Sep 28, 2011 2:00 pm
quad Titan Kepler 6GB + quad Titan X Pascal 12GB + quad GTX1080 8GB + dual GTX1080Ti 11GB
Seems that NVidia has real problem - Tesla 3000 will be based on two Keplers. This is clear that this architecture is not "locked" in consumer products (like GTX680) but is purely lacking computing power. Period.
Ubuntu 9.10 64 bit, GeForce 275, Core2Quad, 6 GB Ram
Where from You get this information? =) as Keplers aren't officially anounced? You judge this from it's rumored 7mil transistors?..maybe it will not be based on two, maybe gtx was just a half of it (as in this way was more profitable to market the product and later on upgrade it? =)x3studio wrote:Seems that NVidia has real problem - Tesla 3000 will be based on two Keplers. This is clear that this architecture is not "locked" in consumer products (like GTX680) but is purely lacking computing power. Period.
On the other hand, It realy doesn't matter on how much Keplers it will be based as long as it will be ONE card =) I have qx6700 (first quad core cpu, that was actually glued from two core2duo's =) and so? Where's the problem?..after 5years i'm still running it =)
If the computational part will be worthless you can always try looking for older, way cheaaper cards (580/590) but they will have way less of vRam. i do understand why computational power might be not so good, 'cause nVidia is actually concentrating on efficiency in order to make those product appeal more for large data centers - energy consumption is the Topic in that field - & there flows a lot of money too.
Anyway, 580/590 are not produced any more, so longer we talk, harder it will be to find those =p
This is the source: http://videocardz.com/32898/nvidia-prep ... -of-memory
Ubuntu 9.10 64 bit, GeForce 275, Core2Quad, 6 GB Ram
@glimpse
Problem is that always two circuits generate more heat, consume more power, use is prohibited by cooling and power restrictions form both PSU and mainboard etc. It is always more expensive (as you pay for double the size cut from wafer). It is also interesting that - what earlier never happened - they change more efficient and more capable architecture to inferior one! And it was their own decision! Try to tell this to Intel engineer that you want new architecture that will take two times more wafer space to deliver half the performance of old processor?!
Problem is that always two circuits generate more heat, consume more power, use is prohibited by cooling and power restrictions form both PSU and mainboard etc. It is always more expensive (as you pay for double the size cut from wafer). It is also interesting that - what earlier never happened - they change more efficient and more capable architecture to inferior one! And it was their own decision! Try to tell this to Intel engineer that you want new architecture that will take two times more wafer space to deliver half the performance of old processor?!
Ubuntu 9.10 64 bit, GeForce 275, Core2Quad, 6 GB Ram
-
AndreaMannori
- Posts: 153
- Joined: Thu Feb 24, 2011 8:25 am
- Location: Bergamo
I love Nvidia but the latest AMD video cards are very fast and have a lot of ram
I think we have to compare 7990 http://www.tomshw.it/cont/news/amd-rade ... 534/1.html with 690 http://www.hwupgrade.it/articoli/skvide ... index.html
Sorry, theese links are in Italian language but you can focus on the charts only.
I agree with you all when you say you hope to see a great development of Octane render
Cheers!
I think we have to compare 7990 http://www.tomshw.it/cont/news/amd-rade ... 534/1.html with 690 http://www.hwupgrade.it/articoli/skvide ... index.html
Sorry, theese links are in Italian language but you can focus on the charts only.
I agree with you all when you say you hope to see a great development of Octane render
Cheers!
Andrea Mannori
win 8.1 x64; nvidia 770M driver 340.52; octane standalone 2 + c4doctane 2.
win 8.1 x64; nvidia 770M driver 340.52; octane standalone 2 + c4doctane 2.
Well, You might be wrong =) 690 compared to 680 scales way beter than 590 to 580 - there're architectural reasons for that..& actually kepler is more efficient =) that was the key of new architecture..x3studio wrote:@glimpse
Problem is that always two circuits generate more heat, consume more power, use is prohibited by cooling and power restrictions form both PSU and mainboard etc. It is always more expensive..
@x3studio - that were rumors,..but now we get more info about new products..k10 is actually is double card, but just take a look at numbers.. we still have to stuck with older (fermi) 2070 as it seems there's no new product for desktop....
Hopefully the power of 680&690 will be unleashed that we all could happily render with Keplers =) as Karba mentioned - the new release is expected at the end of this summer =)
For now if we have extra 590 is still fastest, if You can live with 1,5gb and 580 is 'bigest' with 3gb - if You need even more..a bit lower speed there is 6gb tesla.. - that's it llive with it =)
Soon we will not need any of gtx, as everything will be done in cloud: Autodesk is moving to cloud, games are moving to clound, adobe is moving to cloud..one thing You're need in the future is good bandwith =p
Even when the cloud becomes more common place it will only be used to professionals. Sadly I do not fall into this category so I can't see myself going over to any cloud rendering service unless it is very competitive in its pricing policy.
(HW) Intel i7 2600k, 16GB DDR3, MSI 560GTX ti (2GB) x 3
(SW) Octane (1.50) Blender (2.70) (exporter 2.02)
(OS) Windows 7(64)
(SW) Octane (1.50) Blender (2.70) (exporter 2.02)
(OS) Windows 7(64)
