GTX470 benchmarks / comparison

[radiance]

great

indeed, as i thought, 48 cores instead of 32, i will add support into 2.3 (it's very simple and quick to add),
so 2.3 should give you the real speed of the GTX460, using all 48 instead of just 32 with 2.2.

Radiance

[waterthinking]

Great to hear that!

Was very disappointed with my 460 performance (1.7 MegaSamples)

[Kitsune]

48 cores instead of 32, i will add support into 2.3 (it's very simple and quick to add)

One line of code to detect if the maximum warp size is 32 or 48?

[radiance]

48 cores instead of 32, i will add support into 2.3 (it's very simple and quick to add)

One line of code to detect if the maximum warp size is 32 or 48?

it's more than a few, but it's already done

Radiance

[Skeletor]

Any word on when 2.3 comes out?

[crazyhorse666]

my god Radiance do you ever sleep dude or do you dream in cuda code..

Respect and have a good sunday...

Crazy

[radiance]

Any word on when 2.3 comes out?

we're aiming for tomorrow, but we've still not fixed the cuda 3.1 issue.
i'm not sure. we're working as hard as we can, and it won't be long now.

Radiance

[Skeletor]

I didn't expect it to be ready so soon! But that's awesome - getting my GF460 2GB on Tuesday

[buda]

What difference in compute model 2.0, 2.1 ??

[max3d]

Version 2.2 on Win7 i7, fast workstation, etc.

My cheap GTX460 reports 0 compute 21. What does that say?

device manager shows it´s aware that the card has 224 cores and it shows a 32 warp size.

Result with pathtracing on 1.59 which seems rather slow.

The card with just 768 mb to test octane reports through GPU-Z 675mhz core clock. Device manager reports 791. How can that be.