NVIDIA GeForce GTX 580 - The Real Fermi Deal

512 Cores Unleashed

When the first Fermi card, the GeForce GTX 480 was released, we were excited but also disappointed. Excited because the card NVIDIA has talked about for so long was finally here in our labs; and disappointed because while the Fermi architecture promised 512 CUDA cores, the GeForce GTX 480 had only 480. There were other undesirable traits like extremely high operating temperatures and power draw to contend to as well.

If you recall from our of the GeForce GTX 480, the full Fermi chip was supposed to have 16 SMs (streaming multiprocessors) with 512 CUDA cores, but the GTX 480, powered by the GF100 chip, only had 15 SMs enabling 480 CUDA cores. Although NVIDIA never explicitly explained why, it is widely believed that, at that point of time at least, yields were too low due to the complexity of the new chip. This wasn’t helped by the fact that TSMC, the foundry tasked to produce the new chips, had numerous issues with their 40nm process, which affected not only NVIDIA but also ATI.

Recently however, in face of competition from AMD and their new , the rumor mill was abuzz with news that NVIDIA is preparing a card that will finally flaunt the full 512 cores that the Fermi architecture is capable of, in a bid to cement their grip on the title of world’s fastest single GPU card. That card is the new GeForce GTX 580 and it has finally arrived and is in our labs.

Architecturally, despite being given the GF110 codename by NVIDIA, the GeForce GTX 580 is very similar to the older GTX 480 with the key exception being an additional SM on the GTX 580. Consequently, this additional SM gives the GeForce GTX 580 an extra 32 CUDA cores and four texture mapping units, bringing the total number of cores and texture mapping units on the GTX 580 to 512 and 64 respectively. And on top of that, NVIDIA has also incorporated two enhancements such as full-speed FP16 texture filtering and support for new tile formats to improve Z-cull efficiency to speed up the GeForce GTX 580.

Performance enhancements aside, NVIDIA was also anxious to improve the power efficiency and thermal characteristics of the new card. This led to significant improvements that were done at the transistor level, where lower leakage transistors were used on less timing sensitive processing paths and higher speed transistors on more critical processing paths. This also allowed NVIDIA to cut back on the overall number of transistors used, and the end result is that the GeForce GTX 580 has a lower rated TDP of 244W compared to the GeForce GTX 480’s 250W.

Impressively ,despite the slightly lower rated TDP, the GeForce GTX 580 runs at significantly higher clock speeds. It’s core is clocked at 772MHz, whereas its shader and memory are clocked at 1544MHz and 4008MHz respectively. Compared to the GeForce GTX 480’s core, shader and memory clock speeds of 701MHz, 1401MHz and 3696MHz DDR, it’s amazing how the GTX 580 can run at higher clock speeds, have more cores, and still have a lower rated TDP, which only goes to show just how hard NVIDIA has worked at optimizing the GF110 chip in the GeForce GTX 580.

Before we continue, here’s how the GeForce GTX 580 stacks up against its closest competitors.

The NVIDIA GeForce GTX 580

As is the case with high-end cards, the GeForce GTX 580 is not what you’d call compact. Measuring at a considerable 10.5 inches and weighing a hefty 884 grams, the GeForce GTX 580 is a massive card, but surprisingly, as the pictures below can attest, it looks slightly more compact than the GeForce GTX 480.

This is due mostly to the new cooler that NVIDIA employs on the GeForce GTX 580. Apart from increasing performance and improving power efficiency, NVIDIA was also concerned about the new card’s acoustic and thermal characteristics. The GeForce GTX 480 was notorious for running really hot and it was also very noisy to boot, as the fan will emit a noticeable whine whenever the card is heavily taxed.

To rectify this, the GeForce GTX 580 has a completely redesigned cooling solution which uses a vapor chamber cooling design at the heart of the heatsink solution. Vapor chamber technology promises faster heat transfer by means of liquid vaporization and condensation and has often been employed by Sapphire on their high-end cards such as . Apart from that, NVIDIA has also made tweaks to the fan so that it produces a lower pitch and is less intrusive; and also the cooler cover, which is slightly tapered so that it can better route air towards the rear bracket when configured for SLI.

Test Setup

To find out how the GeForce GTX 580 performs, we’ll be using our dependable X58 setup with the following specifications:

• Intel Core i7-975 (3.33GHz)
• Gigabyte GA-EX58-UD4P motherboard
• 3 x 1GB DDR3-1333 G.Skill memory in triple channel mode
• Seagate 7200.10 200GB SATA hard drive
• Windows 7 Ultimate

We will be looking closely at how the GeForce GTX 580 performs against the other two reigning speed kings, GTX 480 and ATI’s dual-GPU Radeon HD 5970. And with the GeForce GTX 480 being notorious for running really hot and being power hungry, we are also interested to see if the optimizations carried out by NVIDIA will help give the GeForce GTX 580 a less voracious appetite for power and keep it cooler to run.

The full list of cards tested and driver versions used:

• NVIDIA GeForce GTX 580 1560MB GDDR5 (ForceWare 262.99)
• NVIDIA GeForce GTX 480 1560MB GDDR5 (ForceWare 260.89 )
• NVIDIA GeForce GTX 470 1280MB GDDR5 (ForceWare 258.96)
• ATI Radeon HD 5970 2GB GDDR5 (Catalyst 10.9)
• AMD Radeon HD 6870 1GB GDDR5 (Catalyst 10.10)
• ATI Radeon HD 5870 1GB GDDR5 (Catalyst 10.9)

The list of benchmark used are as follows:

• Futuremark 3DMark Vantage
• Crysis Warhead
• Far Cry 2
• Warhammer: Dawn of War 2
• Battlefield Bad Company 2
• “Heaven" from Unigine v2.1
• S.T.A.L.K.E.R.: Call of Pripyat

3DMark Vantage Results

The GeForce GTX 580 got off to a blazing start on 3DMark Vantage racking up significantly higher 3DMarks compared to the GeForce GTX 480. On the Performance preset it scored 20% higher and this lead increased to as much as as 31% on the Extreme preset. With its improved performance, the GeForce GTX 580 was able to take on the dual-GPU Radeon HD 5970 and was able to match it all the way, and also completely overpower ATI/AMD’s other high-end single GPU cards.

 

Crysis Warhead and Far Cry 2 Results

On Crysis Warhead, the GeForce GTX 580 was able to maintain its impressive lead over the GeForce GTX 480, and was on average about 15% quicker overall. Against the GeForce GTX 470, the lead was even greater at more than 40% overall, so the GeForce GTX 580 is indeed a very powerful card.

Against the offerings from the red camp, the GeForce GTX 580 found itself undone by the Radeon HD 5970, which was on the whole about 10% quicker. However, the Radeon HD 5970 is a dual-GPU card, and against ATI/AMD’s single GPU offerings, the GeForce GTX 580 was obviously quicker. 

On Far Cry 2, the GeForce GTX 580 again recorded substantially faster frame rates compared to the GeForce GTX 480 and GTX 470. It also managed to edge out the Radeon HD 5970, and this was especially apparent on the more intensive runs when 8x anti-aliasing was enabled, where it was about 8% to 17% quicker. And against the Radeon HD 5870 and HD 6870, the GeForce GTX 580 could be as much as 60% quicker.

Dawn of War 2 & Battlefield Bad Company 2 Results

Dawn of War 2 has traditionally been a closely contested benchmark amongst high-end cards and all our cards managed very comparable frame rates of above 70fps across all resolutions. But if we had to nitpick, we would say that the GeForce GTX 580 was marginally slower than the GeForce GTX 480 and the the Radeon cards. However, the difference was only about two frames and should be unnoticeable by most users. 

By virtue of having an additional SM, the GeForce GTX 580 also boasts an additional PolyMorph compared to the GeForce GTX 580. The PolyMorph engine handles tessellation work loads and looking at the graphs, we can see a distinct increase in performance. In all, the GeForce GTX 580 was a good 15% faster than the GeForce GTX 480, and was able to marginally outperform the dual-GPU Radeon HD 5970. And impressively, when compared to ATI/AMD’s fastest single GPU cards, the GeForce GTX 580 was more than 40% quicker. 

Unigine "Heaven" 2.1 Results

Unigine’s “Heaven” is an extremely tessellation heavy benchmark and this should allow the GeForce GTX 580 to flex its muscles. Compared to the GeForce GTX 480, the GeForce GTX 580 was about 15% faster overall, and was a whopping 40% quicker overall compared to the GeForce GTX 470. The GeForce GTX 580 also handily beat the Radeon HD 5970 by about 8% and is miles ahead compared to the Radeon HD 5870 and HD 6870. However for DirectX 10 results, it seems to be that NVIDIA is going full ahead to optimize for DX11 than for DX10, both at the hardware and driver level.

DirectX 11 Results

DirectX 10 Results

S.T.A.L.K.E.R.: Call of Pripyat Results

Fermi’s advantage in applications where tessellation is involved was apparent on S.T.A.L.K.E.R.: Call of Pripyat. On the DirectX 11 runs, the GeForce GTX 580 was more than 10% quicker than the Radeon HD 5970, but when we disabled tessellation and reverted to DirectX10, the Radeon HD 5970 started clawing its way back, especially on the Sunshafts run, where it left the GeForce GTX 580 for dead. In fact, on the DirectX10 Sunshafts run, the GeForce GTX 580 was only a shade quicker than the Radeon HD 6870 and HD 5870, which goes to show just how tessellation-biased the Fermi architecture is. 

DirectX 11 Results

  DirectX 10 Results

Temperature

Thanks to the new vapor chamber cooler design and the various other tweaks and optimizations, the GeForce GTX 580 is markedly cooler than the both the GeForce GTX 480 and GTX 470. We recorded a maximum operating temperature of 80 degrees Celsius, which, although not remarkably cool, represents a major improvement over the GTX 480 and GTX 470.

Power Consumption

NVIDIA claims that the GeForce GTX 580 will be faster yet less power hungry than the GeForce GTX 480 and judging from our graph, they have duly delivered. Still,, the fact that even the dual-GPU Radeon HD 5970 recorded “only” 288W shows that NVIDIA has some ways to go before they can truly claim to be power efficient. 

Overclocking

It’s still early days yet but we managed to push our test sample to a respectable 860MHz on the core, giving us 14130 3DMarks on the Extreme preset - a commendable 8% increase in performance. Sadly, we couldn’t manage to run the card on higher memory clock speeds, as artifacts would start showing up, but if we could, no doubt we would see a greater jump in performance. 

GTX 580 - Everything the Original GTX 480 Should Have Been

As far as single GPU cards go, the GeForce GTX 580 is an absolute beast. The older GeForce GTX 480 was already an extremely powerful single GPU card, but the fact that the GeForce GTX 580 is able to perform substantially faster, up to 15% to 20% more depending on benchmark, is just staggering and underlies the sheer brute graphics crunching power of the GF110 chip in the GTX 580.

It was also a good match for the dual-GPU ATI Radeon HD 5970, especially when tessellation was called into play, like Unigine’s “Heaven” benchmark and S.T.A.L.K.E.R.: Call of Pripyat. However, the Radeon card was able to make up for lost ground whenever we reverted to DirectX10, which shows that the Fermi architecture is better geared towards handling tessellation operations and more forward looking games. The problem here is that it will still take a while before a good majority of the popular titles are taking advantage of DX11 and designed to take advantage of the tessellation function. Until then, we still need a card that is a capable DX10 performer, which the GTX 580 still handles reasonably, but not to our expectations. Perhaps it's a card too soon for today, but if games tomorrow are being pumped out fast on the DX11 platform, the GTX 580 has a lot of staying power.

In other aspects, the card wasn't very cool to run, but we were still pleased by NVIDIA’s redesigned vapor chamber cooler, especially when comparing the GeForce GTX 580’s operating temperatures against the GeForce GTX 480 and GTX 470 - 80 degrees Celsius against the 90-plus degrees Celsius of the older cards. Of course, this isn't purely because of the cooler, but also because of the more optimized core at the transistor level.

We were also pleased by the tamer power requirements of the GeForce GTX 580. To be sure, the GeForce GTX 580 is no where near as efficient as the older ATI Evergreen card, but considering it’s faster than the GeForce GTX 480 yet has a lower rated TDP, it’s a laudable improvement in NVIDIA’s books at least.

To match its high levels of performance, the GeForce GTX 580 has an equally high launch price to match - US$499 - which is exactly the same as the GeForce GTX 480 when it was first launched. To be honest, the GeForce GTX 580, like other extreme high-end cards, is not what one would call a value buy. For sheer bang for buck, two Radeon HD 6870 or GeForce GTX 460 cards in a multi-GPU configuration is cheaper and more powerful.

What the GeForce GTX 580 represents, however, is the zenith of single GPU gaming, because as far as single GPU cards go, it is untouchable and is even able to go toe to toe with the Radeon HD 5970, which has two GPU cores. This is an important consideration for the super hardcore users because while the Radeon HD 5970 can only do two-way CrossFire, the GeForce GTX 580 has the ability to go for a 3-way setup, which means, the GTX 580 has more room for future upgrades.

But while the new GeForce GTX 580, is one for hardcore, power-crazed users who seek the very best, it may not be the fastest for long because AMD is set to unveil their high-end Cayman parts from their new Northern Islands line-up. Granted, , we might not be seeing the Cayman cards anytime soon, but still, the fact that AMD has yet to unleash their high-end parts might dissuade some users from jumping in on the GeForce GTX 580 right now as they take a wait-and-see approach. And that, we think, is the biggest problem facing the GeForce GTX 580. That said, AMD will have to pull something very special out of the bag if they want to top the GeForce GTX 580.

At this moment, the GeForce GTX 580 looks like a respectable response in anticipation of the upcoming Radeon HD 6900 series, but how good is it really? Only time will tell.

Our articles may contain affiliate links. If you buy through these links, we may earn a small commission.