Worth the Wait: NVIDIA's Kepler GTX Geforce 680 is New Graphics Market King
March 23, 2012 12:00 AM
comment(s) - last by
New GPU is more powerful, but also quieter, cooler; beats AMD's similar offering in price
In January, Advanced Micro Devices Inc. (
) shipped the world's first 28 nm graphics processing unit,
. Leveraging AMD's
long-awaited new architecture
, Graphics Core Next, the ensuing
Radeon HD 7950/70 card
snatched the performance crown away from rival NVIDIA Corp. (
In the months that follow AMD fleshed out its lineup with four more cards, the
Radeon HD 7750/70
Radeon HD 7850/70
While pricing was a bit high, in all but the Radeon HD 7700 series the AMD card was the best buy because NVIDIA's
28 nm counterpunch
was missing in action.
I. The New Gaming King
Missing in action, that is, until now. A week after a
popped up, NVIDIA has pulled the wraps off of its flagship desktop
graphics card, the GeForce GTX 680.
Almost everything in the GK104 architecture chip has been improved. The die is a petite 294 mm2, with 3.5b transistors onboard, versus AMD's 365 mm2 4.3b transistor
. Likewise, NVIDIA not only one-ups AMD in core clock speed (1008 MHz on the GTX 680 vs. 925 MHz on the Radeon HD 7970), but it also installs a promising new dynamic clocking system, which allows smartphone-esque throttling up or down, based on performance demands.
GK104 [Image Source: NVIDIA]
In "unlocked" card models, NVIDIA expects the card to dip as a low as 325 MHz at idle allowing massive power savings. On the opposite end of the spectrum, in times of extremely demanding performance, unlocked cards can dynamically clock up over the 1.1 GHz barrier, all automatically.
(Click to enlarge) [Image Source: NVIDIA]
NVIDIA's frame buffer (memory) is a bit smaller -- 2 GB of GDDR5 vs. 3 GB of GDDR5 in the Radeon HD 7970, and the bus is narrower -- 256-bit vs. 384-bit. Despite NVIDIA holding a slight edge in memory clock (6.008 GHz v. 5.5 GHz), memory throughput will like favor AMD.
(Click to enlarge) [Image Source: NVIDIA]
shows it to be faster in almost all games, though the AMD flagship manages to eke out a win in some tests. In power and heat NVIDIA has dramatically improved over the 500 series, but it only earns a tie with AMD. However, it is much quieter than AMD's cards.
II. GPU Computing -- Some Steps Forward, Some Spinning of the Wheels
The new card mostly impresses when it comes to GPU computing.
The card streamlines the Fermi architecture, eliminating the high performance, but divergent higher shader clock. In its place it uses the core clock ubiquitously in all its computing functional units. As a result, most of the components of its functional units doubled -- such as the number of CUDA cores, load/store units, and special function units. For example, the CUDA core count in a block within a functional unit doubles from
32 to 64
16 to 32. As a result, NVIDIA is able to keep pace on a functional unit level even while eliminating its higher performance shader clock.
To move things forward, NVIDIA then doubles the number of "blocks" of cores from 3 to 6 per functional unit, effectively doubling performance. In total 192 CUDA cores (6 blocks of 32) now lurk inside a GK104 streaming multiprocessor (SM), vs
48 per SM (3 blocks of 16 cores) in the previous generation architecture.
[Image Source: NVIDIA]
SMs are grouped in blocks called GPCs. There's twice as many GPCs (4) as Fermi (2), but they each half half the number of SMs (2 vs 4 in Fermi), so the SM count stays the same.
A couple remaining oddities are that it declines to boost the shared memory space from 64 kB (a disappointment considering 192 cores are now sharing the resources previously shared by 96 cores). Also it offers 8 special CUDA cores per function unit that offer full 1/1 64-bit floating point (FP64) performance, versus 32-bit floating point. This is the first GPU computing chip to ever offer 1/1 FP64 vs. FP32, however that achievement is dulled by the fact that there are only 8 of these cores per functional unit, meaning an effective speed of 1/4 FP64 per functional unit or 1/24 FP64 per SM.
Still for all its gains in GPU computing,
's benchmarking shows it to only be roughly on par with AMD's flagship card, winning in some GPUCompute benchmarks, losing in others. Of course a tie still works in NVIDIA's favor as it has arguably the best supported GPU programming API -- CUDA -- which is slightly easier to learn and master than OpenGL, thanks in part to the large amount of resources and support NVIDIA throws at developers.
III. Buy One if You Can
NVIDIA's card is available today for $500 USD. NVIDIA is going to tell you that it's the fast card on the market and toss out terms like "revolutionary". The good news, is that when it comes to gaming it is a solid card, though its less of a revolution and more of a nice iterative bump.
Still, that bump is enough to make it the new king of the graphics market on the high end.
The choice is now easy for customers -- buy a GTX 680. That's the good news.
The bad news is that the choice may not be that easy.
writes that NVIDIA indicated that launch supplies may be slightly scarce. Thus it's very possible that GTX 680s could be sold out, taking this option off the plate temporarily.
This all gets back to the yield difficulties reportedly experienced by Taiwan Semiconductor Manufacturing Comp., Ltd. (
) on their
new 28 nm node
. Like AMD, NVIDIA is likely aggressively binning the good chips coming off the line for use in its flagship cards, but the problem is that higher quality 28 nm silicon appears to be having very low yields. As a result, expect supply of NVIDIA's unannounced lower-end
derivatives to be a bit more liberal, but that they'll have lower clock speeds similar to AMD's chips.
So get your hands on the GTX 680 if you can find one -- it's the best thing you can find -- for now -- until the rumored "Big Kepler" comes along.
This article is over a month old, voting and posting comments is disabled
3/26/2012 10:50:29 AM
see Taft12 post. LOL
"So, I think the same thing of the music industry. They can't say that they're losing money, you know what I'm saying. They just probably don't have the same surplus that they had." -- Wu-Tang Clan founder RZA
Quick Note: Apple Sells 3 Million iPads During Launch Weekend
March 19, 2012, 5:55 PM
AMD Completes Its GCN Lineup With Impressive Mid-Range 7850/7870 Cards
March 5, 2012, 3:10 PM
AT&T Throttling Unlimited Data Users After Only 1-2 GB
February 15, 2012, 9:35 AM
EU Nails Samsung With Formal Investigation Over 3G Patent Abuse
January 31, 2012, 9:32 AM
AMD Regains Single-GPU Performance Crown From NVIDIA, For Now
December 22, 2011, 2:07 PM
4/16/2014 Hardware Reviews
April 16, 2014, 9:01 AM
Quick Note: Kingston's 1 TB USB Stick Hits $899 on Lightning Deal
April 15, 2014, 3:35 PM
4/15/2014 Hardware Reviews
April 15, 2014, 11:30 AM
4/11/2014 Hardware Reviews
April 11, 2014, 11:03 AM
Global PC Shipments Declined 1.7 Percent in Q1 2014
April 10, 2014, 9:58 AM
Intel Previews Devil's Canyon Chip, "Black Book", and Broadwell
March 21, 2014, 8:15 AM
Most Popular Articles
Cities to Carpoolers: Sharing Your Car is Illegal, We Will Seize Your Cars
April 4, 2014, 9:17 PM
iPad Exploiter is Freed by Federal Appeals Court
April 11, 2014, 7:40 PM
A-10 Warthog May Live to Fight Another Day with Support from Lawmakers
April 14, 2014, 9:41 AM
Taiwan's AOU Claims to Have World's Highest-Res. OLED Smartphone Display
April 11, 2014, 1:44 PM
EFF: NSA May Have Used IRC Botnets to Exploit Heartbleed for Last Two Years
April 14, 2014, 4:43 PM
Latest Blog Posts
Facebook Aims to Provide Internet to "Every Person in the World" with Drones, Satellites
Apr 1, 2014, 10:20 AM
Retail Mobile Sites Experience Outages in Light of Simplexity's Bankruptcy
Mar 14, 2014, 8:48 AM
Tesla vs. BMW: Who Has the Safer EV?
Feb 1, 2014, 2:56 PM
Justice Leaks Details of Next HTC One Two Flagship Phone
Dec 5, 2013, 4:04 PM
Global Cyber Espionage Concerns Reveal Growing Cyber Armies
Nov 29, 2013, 11:04 AM
More Blog Posts
Copyright 2014 DailyTech LLC. -
Terms, Conditions & Privacy Information