- Mon Jul 19, 2010 5:13 pm
#326864
Hello all,
i was browsing my disc, and found file MM_00.pdf . It is old Maxwell magazine - I think it is the first one edition. If you get it, look at page 10. There is interview with Victor Gonzales. For those who don`t have it, i will quote some sentences (hopefully i am alowed to do that)
PS: As far as engine speed is concerned, what real improvement margin has NL foreseen here and how much do you at NL instead confide in foreseeable technplogical forward-leaps and in distributed rendering procedures? In other words, does the end user need to be prepared to purchase more powerful and/or a larger number of machines or can some substantial optimization be expected? Again, does the real solution lie in cooperative rendering or in individual machine processing power?
"VG: I think both ways are right in some sense. Individual machine power is the main power resource today and the most efficient method to minimize cost and maximize performance. At the moment network rendering is a must for high volume or to reduce computing time, but this is not an option for every user. In the short term we can expect the best benefits coming from the upcoming multicore processors (up to 4 cores and 1 teraflop in a near future) and the new GPU technology.
Without giving away too many details, we can guarantee serious improvements in speed in a single machine based on different research and development areas while we improve the network rendering feature for the upcoming Maxwell Render 1.2."
So this is an old interview, but you can see, Victor and NL were thinking about GPU technology long time ago and nobody know, what we can expect in near future. As some of you stated, this is relatively new technology and can change rapidly with upcoming new GPU hardware...
Patrik
i was browsing my disc, and found file MM_00.pdf . It is old Maxwell magazine - I think it is the first one edition. If you get it, look at page 10. There is interview with Victor Gonzales. For those who don`t have it, i will quote some sentences (hopefully i am alowed to do that)
PS: As far as engine speed is concerned, what real improvement margin has NL foreseen here and how much do you at NL instead confide in foreseeable technplogical forward-leaps and in distributed rendering procedures? In other words, does the end user need to be prepared to purchase more powerful and/or a larger number of machines or can some substantial optimization be expected? Again, does the real solution lie in cooperative rendering or in individual machine processing power?
"VG: I think both ways are right in some sense. Individual machine power is the main power resource today and the most efficient method to minimize cost and maximize performance. At the moment network rendering is a must for high volume or to reduce computing time, but this is not an option for every user. In the short term we can expect the best benefits coming from the upcoming multicore processors (up to 4 cores and 1 teraflop in a near future) and the new GPU technology.
Without giving away too many details, we can guarantee serious improvements in speed in a single machine based on different research and development areas while we improve the network rendering feature for the upcoming Maxwell Render 1.2."
So this is an old interview, but you can see, Victor and NL were thinking about GPU technology long time ago and nobody know, what we can expect in near future. As some of you stated, this is relatively new technology and can change rapidly with upcoming new GPU hardware...
Patrik
- By choo-chee
- By Andreas Hopf