The reviews about the GTX 650 ti say that nvidia released that card to replace the 560 ti with less power consumption more cuda cores and more clock speed…
thats for gaming!!
What about cycles?? does the 150 dollars 650 ti really replaces the performance of 560 ti in cycles.
the 650 ti being the kepler architecture might be 20% slower but the 560 ti has 384 cuda cores while the 650 ti has 768 cuda cores which is 20% more then the 560 ti?
So does the cuda cores in 650 ti cover the performance lag of 6 series card through the number of cores then the 560 ti?
The double to cores doesn’t mean anything. The 6 series is designed for faster gaming - not faster rendering. They are completely different.
I found this response from a cude core programming site (in a blender thread none the less…)
The reasons why the 5 series is quicker at rendering than the 6 series…
5xx shaders run at 2x speed, but there are fewer of them…this is a good for some applications, bad for others, but the 6xx chips use less power
the 6xx chips have reduced double-precision arithmetic units. Games do not (for the most part) use doubles, so cutting this for gaming cards again reduces the power consumption and allows other parts to run faster
the 6xx has new compute capabilities (look up “cuda compute capability level”) mostly around scheduling. This is definitely a good thing, but most software doesn’t take advantage of it yet.
Don’t get fixated on the Cuda cores numbers. Just buy as much VRAM as you can, and either get a 5series (580>570) if you really need that extra few minutes, or buy a 6 series. I personally will be buying a 6 series next month - the 5 series are almost the same in price on Amazon for an equavilent VRAM amount. The extra few minutes rendering won’t matter to me.
EDIT : a Benchmark done with Blender, cards, OS, systems last year. Note that the 580 was fastest, but I don’t think the 6 series was released for comparison.
I know what you mean, I’m not technical at all…well, not at computers The 6 series was kind of a bust for Rendering - but it was designed as a gamers card, so if you are doing both, get one. The more VRAM means the more you can fit in your scene (models / textures / etc). Just keep an eye out on some TI cards, some of those seem to be lesser cards - just compare the ones in your price range, and get the best of everything you can afford!
To be fair I will correct my last post, I have now managed to do some renders a bit faster with the GPU now, the difference is not at all “stunning” and does seem to depend on settings and the amount of texturing involved. My card has only 1G and I have already managed to crash it. I still would not recommend it.
I’m using a 460 with less than 1GB of ram and it’s still much faster than my CPU, more than double actually.
HOWEVER, I have to always make sure that my card is running at full speed. I’ve set blender to have a high performance profile in my video drivers, but after some use and when I close blender, my card falls back to 2D clock speeds and gets stuck there until I reboot.
Use a performance monitoring application and make sure your video card is running at it’s full clock speed when blender is running.