NVIDIA Kepler speculation thread

Seems normal:
perfwatt_1920.gif
http://www.techpowerup.com/reviews/NVIDIA/GeForce_GTX_680/1.html (yup, that's the tpu review :)
 
The Adaptive Vsync is a true bliss here -- zero input lag, no tearing and no stuttering. No need for triple buffering (wasted memory), too. :D
 
Now, that's weird (from TR's review):
The organization of the SMX's execution units isn't truly apparent in the diagram above. Although Nvidia likes to talk about them as individual "cores," the ALUs are actually grouped into execution units of varying widths. In the SMX, there are four 16-ALU-wide vector execution units and four 32-wide units. Each of the four schedulers in the diagram above is associated with one vec16 unit and one vec32 unit. There are eight special function units per scheduler to handle, well, special math functions like transcendentals and interpolation. (Incidentally, the partial use of vec32 units is apparently how the GF114 got to have 48 ALUs in its SM, a detail Alben let slip that we hadn't realized before.)

Although each of the SMX's execution units works on multiple data simultaneously according to its width—and we've called them vector units as a result—work is scheduled on them according to Nvidia's customary scheme, in which the elements of a pixel or thread are processed sequentially on a single ALU. (AMD has recently adopted a similar scheduling format in its GCN architecture.) As in the past, Nvidia schedules its work in groups of 32 pixels or threads known as "warps." Those vec32 units should be able to output a completed warp in each clock cycle, while the vec16 units and SFUs will require multiple clocks to output a warp.
Click to expand...
Asymmetric SIMDs? :???:
 
fellix said:
The Adaptive Vsync is a true bliss here -- zero input lag, no tearing and no stuttering. No need for triple buffering (wasted memory), too. :D
Click to expand...

I was pleasantly surprised to see that nvidia did not cap this feature for Kepler owners only! It's actually a driver feature!

+1 on my Nvidia rating book! Good job and thanks Nv!
 
ECH said:
It does draw a :?: as to why tweaktown's benchmarks and power consumption numbers are different from the other reviews. If you read the comments he does say that they didn't receive a card from them and he's not under NDA. Tweaktown's review shows higher power consumption for the GTX 680. So will the retail card's performance and power consumption actually be different then what's reviewed?
Click to expand...

I've just measured our first partner sample besides the GTX 680 and the differences are negligible, very well within normal variance of 1-2 watts at different load states. TT might not have a driver that correctly enabled the power management stuff.
 
silent_guy said:
With the limited amount of info available, I don't see how GPU Boost differs from PowerTune?
Click to expand...
Big difference in implementation. PowerTune being a calculated "infered" power at the chip level we can choose to implement it in a deterministic or non-deterministic manner; initial implementations we have chosen deterministic in order to ensure all cards deliver the same performance regardless of the ASIC characteristics or the environment it is in. With the Boost implementation here it appears to be using the GTX 580 input power draw circuitry, which meas it is entirely non-determinisic and you can see different performances with different chip characteristics and even by the conditions the end user is running.
 
You must log in or register to reply here.

Similar threads

A
RDNA3 Efficiency [Spinoff from RDNA4]
3 4 5
Replies
85
Views
10K
Bondrewd
Bondrewd
iroboto
CES 2025 Thread (AMD, Intel, Nvidia, and others!)
9 10 11
Replies
209
Views
11K
DavidGraham
D
D
The Silent Hand of Driver "Optimization" (spinoff)
2
Replies
31
Views
3K
DavidGraham
D
F
AMD RDNA5 Architecture Speculation
2
Replies
38
Views
6K
Seanspeed
S
Shortbread
Nvidia is Now The Second-Largest Company in The US Market
Replies
4
Views
3K
MfA
M
Back
Top