Graphics Cards Guide

Compare This

Launch SRP US$499

NVIDIA GeForce GTX 680 (Reference Card)

22 Mar 2012

With its new SMX processing units and sporting 1536 CUDA cores, the NVIDIA GeForce GTX 680 is the current flagship graphics card based on the NVIDIA 28nm Kepler architecture.

Tags: nvidia, kepler, geforce gtx 680

Filter by Manufacturers

HWZ's Overall Rating: 9.0

Performance:

Features:

Value:

8.5

Excellent all-round performance

Extremely quiet for a flagship card

GPU Boost in Kepler architecture

Temperature and power consumption wasn't low but was justified with GPU boost

<div id="product-review" class="article">
			<h2 class="header-green">GeForce GTX 680 - NVIDIA&#039;s 28nm Successor to Fermi</h2>
		<span class="byline">Reviewed by Wong Chung Wee, Kenny Yeo & Vijay Anand on Thursday, 22 Mar 2012</span>
		<div class="share">
	<div
		id="addthis_div_51c1bc918ed14" 
		class="addthis_toolbox addthis_default_style"
		addthis:url="http://www.hardwarezone.com/review-geforce-gtx-680-nvidias-28nm-successor-fermi"		addthis:title="GeForce GTX 680 - NVIDIA&#039;s 28nm Successor to Fermi"	>
		<a class="addthis_button_google_plusone" g:plusone:count="false"></a>
		<a class="addthis_button_facebook"></a>
		<a class="addthis_button_twitter"></a>
		
		<a class="addthis_button_email"></a>
	</div>
</div>
<script type="text/javascript">
//<![CDATA[
	window.addEvent ('domready', function ()
	{
 		//AddThis thing
		addthis_share = window.addthis_share || {};
		addthis_share.url_transforms = {
			add: {
				utm_source: 'addthis',
				utm_medium: 'referral',
				utm_campaign: 'addthis',
				utm_content: 'GeForce%2BGTX%2B680%2B-%2BNVIDIA%27s%2B28nm%2BSuccessor%2Bto%2BFermi'
			}
		}
		
		addthis_config = window.addthis_config || {};
		addthis_config.data_track_clickback = true;
		addthis_config.data_ga_property = 'UA-166763-1';
		//window.addthis && addthis.toolbox (".addthis_toolbox");
		 
	});
//]]>
</script>		<div class="content resizeable">
						<div id="page-1" class="page">
		 		<h2 class="page-title">Kepler&#039;s Laws - Faster, Smoother and Richer</h2>
		 		<h3 class="page_title">GeForce GTX 680 -&nbsp; Stating Kepler's Laws</h3>
<p>Since its announcement in October 2010 at <a target="_blank" href="http://www.hardwarezone.com.sg/feature-nvidias-fermi-successors-revealed-kepler-and-maxwell">NVIDIA's GPU Technology Conference (GTC)</a>, Kepler, the successor to&nbsp;<a target="_blank" href="http://www.hardwarezone.com.sg/review-nvidia-geforce-gtx-580-real-fermi-deal">Fermi</a>, has kept us waiting in bated breath for its arrival. Today, 22nd March 2012, marks its official launch and marks the first salvo of retaliation by NVIDIA to AMD's Southern Island series of GPUs. Like AMD's latest GPUs, Kepler is based on the 28nm fabrication process and is currently touted to be the fastest and most power efficient GPU. AMD had the first-to-market advantage and the card manufacturers wasted no time in getting their wares based on the <a target="_blank" href="http://www.hardwarezone.com.sg/feature-amd-radeon-hd-7970-hd-7950-heading-southern-islands-tahiti">Radeon HD 7900 GPU series</a>&nbsp;and <a target="_blank" href="http://www.hardwarezone.com.sg/feature-amd-radeon-hd-7770-seizing-day-cape-verde-asus-way">Radeon HD 7700 GPU series</a>&nbsp;to their different target markets. We feel that NVIDIA has some catching up to do in this race as it reveals its first manifestation of Kepler in the form of the&nbsp;<strong>GeForce GTX 680</strong>.</p>
<p class="rtecenter"><img width="450" height="414" title="The die shot of GeForce GTX 680 GPU with an area of 294 square mm. Its die size has shrunk about 43% when compared to GeForce GTX 580; however, GTX 680 packs in 3.54 billion transistors versus GTX 580 transistor count of 3 billion!" alt="" src="/files/img/2012/03/GeForce_GTX_680_Die_Shot_M.jpg" /></p>
<p>The GeForce GTX 680 is positioned to be the most power efficient flagship GPU, claiming this title from its older&nbsp;<a target="_blank" href="/review-nvidia-geforce-gtx-580-real-fermi-deal">GeForce GTX 580</a> when it too touted similar claims in its heyday. According to technical information gleaned during our visit to NVIDIA GeForce Editor's Day 2012, the architecture of Kepler is actually positioned as an improvement from the GF114 core of <a target="_blank" href="/feature-nvidia-geforce-gtx-560-plugging-mainstream-gap">GeForce GTX 560</a> as the GTX 560 is the latest from NVIDIA's GeForce 500 series GPUs. Using that GPU architecture as the base, NVIDIA designed the GK104 core that makes up the GeForce GTX 680.</p>
<p>Now for the techies, some of you might have long guessed from the core code that this the GTX 680 isn't the true flagship and there is some truth to this looking at the rumors on the grapevine and from our close discussions from NVIDIA. However with no official word and the fact that the GeForce GTX 680 can is positioned to compete with the top Radeon HD 7900 series single GPU graphics card, for all intents and purposes, the GeForce GTX 680 is their current flagship.&nbsp;There are three guiding principles in the development of the Kepler architecture and they are summed up neatly in a triplet of adjectives: <strong>Faster</strong>, <strong>Smoother</strong> and <strong>Richer</strong>.</p>
<p>&nbsp;</p>
<h4 class="page_sub_title">Kepler's First Law - Faster</h4>
<p>One of the most important hardware feature of the Kepler GPU&nbsp;is its four new <strong>Graphics Processing Clusters (GPC)</strong>. This term was first used in Fermi's architecture which also featured four GPC units. A GPC is basically an independent block of processing engines that can exist on its own as it contains all the necessary processing stages within its cluster. While the number of GPC units haven't changed in all these years, the hardware and configuration within each GPC has changed drastically. &nbsp;Within each GPC of the GeForce GTX 680 resides the next generation <strong>Streaming Multiprocessor (SMX)</strong>.</p>
<p class="rtecenter"><img width="500" height="540" title="Each pair of GPCs is separately by the L2 cache that has a capacity of 512KB. Take note of the next generation SMX packs in 192 CUDA cores, an increment of 160 CUDA cores over the previous SM. Incidentally, the CUDA cores in both generations of the the Streaming Multiprocessors share the same kind of CUDA core as they've not evolved. Instead, the performance per watt ratio has been drastically improved on the SMX to enable NVIDIA to pack many more cores, thanks to the die shrink and reorganization of the hardware within." alt="" src="/files/img/2012/03/GeForce_GTX_680_Block_Diagram_M.jpg" /></p>
<h4 class="page_sub_title">&nbsp;</h4>
<h4 class="page_sub_title">SMX - X Marks the Spot?</h4>
<p>According to NIVIDA, SMX&nbsp;promises to deliver more performance, due to its higher number of CUDA cores, than the Fermi architecture and it will deliver the stipulated level of performance with less power; in fact, it is touted to be twice as efficient as the <a target="_blank" href="/review-fermi-arrives-nvidia-geforce-gtx-480/introducing-fermi">previous generation <strong>SM</strong></a>. Upon scrutinizing the block diagrams of both SM&nbsp;and SMX, we feel that the Kepler architecture is evolutionary when compared against the Fermi architecture. Both architecture still sport the same CUDA core, the basic processing block; albeit SMX has 6 times more shader cores. This progression is only natural considering that fact that Kepler's manufacturing process has improved from Fermi's 40nm to the current 28nm.</p>
<p class="rtecenter"><img width="400" height="682" title="The logical representation of the hardware blocks that make up the SMX. It features 192 CUDA cores and 16 texture units. Compared to the previous generation SM of Fermi architecture, the SMX has 6 times more CUDA cores and 4 times as many texture units. Its Polymorph Engine has also been updated to version 2.0." alt="" src="/files/img/2012/03/GeForce_GTX_680_SM_Diagram_FINAL_M.jpg" /></p>
<p>Other improvements we have noticed are the new PCIe 3.0 bus interface as well as a new version of its Polymorph Engine which promises double the tesselation performance. NVIDIA also states that the Kepler architecture features a new improved memory subsystem. While it's total memory bandwidth is still identical to that of the GeForce GTX 580 at 192.3GB/s, NVIDIA has paired the GPU with the fastest graphics memory ever that operate at <strong>6GHz</strong>.&nbsp;With much speedier memory, its bandwidth is limited only by its reduced <strong>256-bit memory interface</strong> (down from 384 bits wide on the GTX 580). &nbsp;According to NVIDIA GPU&nbsp;managers we spoke to, any potential increment to Kepler's memory interface will entail in a corresponding die size increment. At that juncture, their die size target limited Kepler's memory interface to 256 bits wide, while more importance was shifted to the GPU's overall processing prowess with a massive increase in CUDA processing cores - the GTX 680 has <strong>1536 stream processors</strong> as compared to 512 on its predecessor. While NVIDIA maintains that its memory bandwidth is sufficient for the GPU, we'll soon find out from our testing if it holds true.</p>
<p>In any case, NVIDIA has also engineered new anti-aliasing techniques to massively reduce memory footprint and processing requirements with FXAA and the newer TXAA sampling techniques (we'll touch on these more later). This is also another reason why NVIDIA is confident introducing a new flagship with the same memory bandwidth, but has partially offset its limitation with much speedier memory.</p>			</div>
		</div>
			<div class="pagination" >
		<span class="desc">Page 1 of 12</span>
		<ul>
			<li class="prevnext prev"><a href="/review-geforce-gtx-680-nvidias-28nm-successor-fermi" class="disabled">&lsaquo; Prev</a></li>
			<li class="skip">
				<form id="PageAddForm" method="post" action="/go"><fieldset style="display:none;"><input type="hidden" name="_method" value="POST" /></fieldset>					<select name="data[Page][url]" class="page-selector" id="PageUrl">
<option value="/review-geforce-gtx-680-nvidias-28nm-successor-fermi" selected="selected">Kepler&#039;s Laws - Faster, Smoother and Richer</option>
<option value="/review-geforce-gtx-680-nvidias-28nm-successor-fermi/faster-lesser-gpu-boost">Faster with Lesser - GPU Boost</option>
<option value="/review-geforce-gtx-680-nvidias-28nm-successor-fermi/txaa-adaptive-vsync-3d-vision-surround-improvements">TXAA, Adaptive VSync &amp; 3D Vision Surround Improvements</option>
<option value="/review-geforce-gtx-680-nvidias-28nm-successor-fermi/nvidia-geforce-gtx-680">The NVIDIA GeForce GTX 680</option>
<option value="/review-geforce-gtx-680-nvidias-28nm-successor-fermi/test-setup-evga-precisionx-tweaker">Test Setup &amp; The EVGA PrecisionX Tweaker</option>
<option value="/review-geforce-gtx-680-nvidias-28nm-successor-fermi/3dmark-11-battlefield-bad-company-2-results-0">3DMark 11 &amp; Battlefield Bad Company 2 Results</option>
<option value="/review-geforce-gtx-680-nvidias-28nm-successor-fermi/crysis-warhead-far-cry-2-results-11">Crysis Warhead &amp; Far Cry 2 Results</option>
<option value="/review-geforce-gtx-680-nvidias-28nm-successor-fermi/crysis-2-batman-arkham-city-dirt3-results-0">Crysis 2, Batman: Arkham City &amp; Dirt3 Results</option>
<option value="/review-geforce-gtx-680-nvidias-28nm-successor-fermi/unigine-heaven-21-results-4">Unigine &quot;Heaven&quot; 2.1 Results </option>
<option value="/review-geforce-gtx-680-nvidias-28nm-successor-fermi/stalker-call-pripyat-results-9">S.T.A.L.K.E.R.: Call of Pripyat Results</option>
<option value="/review-geforce-gtx-680-nvidias-28nm-successor-fermi/temperature-power-consumption-overclocking-19">Temperature, Power Consumption &amp; Overclocking</option>
<option value="/review-geforce-gtx-680-nvidias-28nm-successor-fermi/conclusion-keplers-laws-or-keplers-flaws-0">Conclusion - Kepler&#039;s Laws or Kepler&#039;s Flaws?</option>
</select>					<input type="submit" class="submit" value="Go" />				</form>			</li>
			<li class="prevnext next"><a href="/review-geforce-gtx-680-nvidias-28nm-successor-fermi/faster-lesser-gpu-boost" class="">Next &rsaquo;</a></li>
		</ul>
	</div>
<script type="text/javascript">
//<![CDATA[
	window.addEvent ('domready', function ()
	{
 
		var container = document.getElement ('#product-review'),
			content = container && container.getElement ('.content'),
			pagination = container && container.getElement ('.pagination'),
			form = pagination && pagination.getElement ('form'),
			menu = form && form.getElement ('.page-selector'),
			links = {
				prev: pagination && pagination.getElement ('.prev a'),
				next: pagination && pagination.getElement ('.next a')
			},
			page = -1,
			desc = 'Page {curPage} of {lastPage}'
		;

// Tag as Javascript-enabled
		pagination.addClass ('js');

menu.addEvent ('change', function (e)
		{
			var curPage = this.selectedIndex + 1,
				lastPage = this.length,
				url = this.get ('value'),
				pageContent = container && container.getElement ('.page')
			;

// Disable controls
			menu.set ('disabled', 'disabled').blur ();
			if (pageContent) pageContent.setStyle ('opacity', .5);

new Request.HTML ({
				'url': url,
				'method': 'get',
				'link': 'cancel',
				'update': content,
				'onSuccess': function ()
				{
					// Update controls
					menu.erase ('disabled');
					links.prev
						.set ('href', menu.getChildren ('option')[(curPage > 1)? curPage - 2: curPage - 1].get ('value'))
						[(curPage > 1)? 'removeClass': 'addClass'] ('disabled')
					;
					links.next
						.set ('href', menu.getChildren ('option')[(curPage < lastPage)? curPage: curPage - 1].get ('value'))
						[(curPage < lastPage)? 'removeClass': 'addClass'] ('disabled')
					;
					pagination.getElement ('.desc').set ('text', desc.substitute ({ curPage: curPage, lastPage: lastPage}));

captureIntraLinks ();
					new Fx.Scroll (window).toElement (container);
				},
				'onFailure': function ()
				{
					// Restore controls
					menu.erase ('disabled');
					if (pageContent) pageContent.setStyle ('opacity', 1);
				}
			}).send ();
		});
		$$([links.prev, links.next]).addEvent ('click', function (e)
		{
			e.stop (); this.blur ();
			if (!this.hasClass ('disabled'))
				menu.set ('value', this.get ('href')).fireEvent ('change');
		});
		form.addEvent ('submit', function (e)
		{
			e.stop (); this.blur ();
			menu.fireEvent ('change');
		})

// Capture all intra-page links and hook to js
		function captureIntraLinks ()
		{
			content.getElements ('a[href^="#"]').flatten().each (function (link)
			{
				var target = link.get ('href').substr (6).trim ();
				if (target && target.match (/^\d+$/) && target <= menu.length)
					link.addEvent ('click', function (e)
					{
						e.stop (); this.blur ();
						menu.selectedIndex = target - 1;
						menu.fireEvent ('change');
					});
			});
		}
		captureIntraLinks ();

// Trigger initial pagination state
		var target = location.hash && location.hash.substr (6).trim ();
		if (target && target.match (/^\d+$/) && target <= menu.length && target != page)
		{
			menu.selectedIndex = target - 1;
			menu.fireEvent ('change');
		}
 
	});
//]]>
</script>
				<div class="footer">
		
<div class="tags" style="">
	<span class="label" style="">Tags:</span> <a href="/product-guide/all/reviews/tagged/130-nvidia">nvidia</a>, <a href="/product-guide/all/reviews/tagged/1341-kepler">kepler</a>, <a href="/product-guide/all/reviews/tagged/7897-28nm-architecture">28nm architecture</a>, <a href="/product-guide/all/reviews/tagged/9357-geforce-gtx-680">geforce gtx 680</a></div>
		</div>
				</div>

<script type="text/javascript">
//<![CDATA[
	window.addEvent ('domready', function ()
	{
 
		var section = $('tab-product-content-review'),
			dropdown = section.getElement ('#reviews-list'),
			tId = 0
		;

if (section && dropdown)
		{
			var list = dropdown.getElement ('ul'),
				links = list.getElements ('li a')
			;

dropdown.addClass ('js');
			list.set ('morph', { 'link': 'cancel' });

$$(dropdown, list).addEvents ({
				'mousemove': function ()
				{
					$clear (tId);
				},
				'mouseenter': function ()
				{
					list.set ('morph', {'duration': 'short'}).morph ({ 'opacity': 1 });
				},
				'mouseleave': function ()
				{
					tId = function () { list.set ('morph', {'duration': 'short'}).morph ({ 'opacity': 0}); }.delay (1000);
				}
			}).fireEvent ('mouseleave');

links.addEvent ('click', function (e)
			{
				e.stop ();
				//unescape the 'alt' prooperty
				var o = new Element('i', { 'html': this.get ('alt')});
				new Request.HTML ({
					'url': o.get('text') + '&as=ajax',
					'method': 'get',
					'link': 'cancel',
					'update': section.empty().grab(new Element('div')
						.addClass ('waiter')
						.grab (new Element ('span', {
							text: 'Please wait while we load your article ...'
						}))
					)
				}).send ();
			})

section.getElements ('.footer .comments a').addEvent ('click', function (e)
			{
				e.stop ();
				new Fx.Scroll (window).toElement ($('tab-product-details'));

var href = this.get ('href').match (/#.+$/), target = href && href[0];
				if (target)
					(function () { new Fx.Scroll (window).toElement ($(target.substr (1))); })
						.delay (3000);

$('tab-product-comments').getChildren ('a')[0].fireEvent ('click');
			})
		}
		if( $$('.image-caption-container').length < 1 )
		{	
			$$('.article .content img[title]').each (function (img)
			{
				// skip if title is empty
				if (!img.get ('title')) return;
			
				new Element ('span', {'class': 'image-caption-container'})
					.addClass (['left', 'right'].contains (img.get ('align'))? img.get ('align'): '')
					.set ('width', img.get ('width'))
					.replaces (img)
					.grab (img)
					.grab (new Element ('div', {
						'class': 'image-caption',
						'html': img.get ('title').trim ()
					}))
				;
				img.erase ('align');
			
				var link = img.getParents ('a');
				if (link) link.addClass ('captioned');
			});
		}
 
	});
//]]>
</script>

Graphics Cards Guide

NVIDIA GeForce GTX 680 (Reference Card)

Filter by Manufacturers

You may also be interested in...