![\"\"](\"http:\/\/i2.wp.com\/alienbabeltech.com\/main\/wp-content\/uploads\/2011\/03\/GeForce_GTX_590_3qtr.jpg\" "\\"GeForce_GTX_590_3qtr\\"")
<\/a>Nvidia released its long awaited GeForce GTX based on its brand new Fermi DX11 GF100 architecture back in April of last year, six months later than AMD’s own DX11 Cypress video cards. This new Fermi GPU \u2013 Graphics Processing Unit – a term originally originated by Nvidia is a continuation of their strategy since their G80 which launched over three year ago to create a General Purpose Processor \u2013 co-equal with the CPU \u2013 that also renders amazing graphics. The culmination of Nvidia’s efforts with their GF100 DX11 Fermi architecture was the GTX 480 with the caveat that it runs rather hot and the cooling solutions based on the reference design were rather noisy.<\/p>\nThings changed very rapidly as Nvidia’s introduced a new refined GTX 400 series ‘Tank’, the Galaxy GTX 480 SuperOverclock on a mature process which we covered in this review<\/a>. Shortly thereafter, enter the completely redesigned Nvidia Tank – at $499 suggested etail pricing and designed to be faster and more efficient than even the super-overclocked GTX 480s.<\/p>\n We saw AMD introduce their new line up, HD 68×0 series to replace HD 58×0 series in our review here<\/a>. We found out that the “Barts” GPU it is based on is only a mid-range launch so far with the HD 6870 only slightly faster than the HD 5850; the best part is that it replaces it for less money. And at the end of last year we saw AMD’s Cayman release in the form of HD 6970 and HD 6950. Nvidia took aim at the HD 6970 with the GTX 570, and the GTX 560 Ti takes on the HD 6870 and the HD 6950.<\/p>\n In turn, AMD released their flagship dual-GPU, $700 video card a few days ago on March 8 and we reviewed it here<\/a>.\u00a0 It absolutely blew away the GTX 580 and AMD’s own HD 6970 with the caveat that it is rather noisy under full load. Well, now Nvidia has just now releases their own dual-GPU graphics card, their flagship GTX 590.\u00a0 It boasts 512 CUDA cores each.\u00a0 That means two independent NVIDIA GF110 GPUs are internally connected to deliver 1024 total cores of processing power. The GTX 590 has six 64-bit memory controllers which make up the 384-bit bandwidth per each GPU. \u00a0 This is matched with\u00a0 3GB of total GDDR5 video frame buffer – 1.5 GB per GPU just like two down-clocked GTX 580.<\/p>\n Here are the specifications for the GTX 590:<\/p>\n Since, Nvidia’s new GTX 590 now also comes with a MSRP of $699, which one is worth your hard earned dollars and are they worth the $700 that one would currently spend for a single video card?<\/em><\/strong><\/p>\n <\/em><\/strong><\/p>\n <\/em><\/strong><\/p>\n To properly bring you this review, we are using our reference HD 6990 and HD 6970 which we put through their paces last week with the release 11.4 (beta) Catalyst drivers in the launch article.\u00a0 We are putting HD 6990 head to head in 28 modern games and in 3 synthetic benchmarks to see which card may be for you using 1680×1050, 1920×1200 and\/or 2560×1600 resolutions. Since we are using fast single-GPU video cards, it makes sense to test them at the highest resolutions and with the most demanding playable settings that they can handle. Since we are matching the two top dual-GPU video cards to each other in a performance showdown, we do also include HD 6970 and HD CrossFire as well as GTX 580 and GTX 560 Ti SLI configurations to get an idea of value.<\/p>\n We also examined the performance of Galaxy’s GTX 480 SuperOverclock<\/a> and we also reran GTX 480 against stock and overclocked versions of HD 5870, HD 6870 and HD 6850 here<\/a> just a few weeks ago. A recent review<\/a> covered the GTX 580 a few months ago. We also covered GTX 570<\/a> and the launch of the HD 69×0 series<\/a> against GTX 460. And Senior Editor Leon Hyman covered GTX 460-768M vs. HD 5830 here<\/a> last week. Now you are up to date.<\/p>\n The GeForce GTX 590 is basically two GTX 580s on a single PCB.\u00a0 However, it was exceptional engineering feat to get these two very powerful GPUs into a single enclosure in a two slot video card that can handle the extreme thermals and wattage without sounding like a hairdryer.<\/p>\n One of the reasons the GTX 590 is barely any louder than a GTX 580 – about on the same level as a HD 6970 – is because the thermals are tamed by Nvidia’s vapor chambers.<\/p>\n Take a look at the coolers.<\/p>\n The GTX 590 was designed from the ground up to deliver exceptional tessellation performance, which is a key component of Microsoft’s DirectX 11 development platform for PC games. Tessellation allows game developers to take advantage of the GeForce GTX 590 dual-GPU’s tessellation ability to increase the geometric complexity of models and characters to deliver far more realistic and visually rich gaming environments. You will soon see that although the clocks of Nvidia’s GTX 590 are clocked far lower at 605\/1707MHz\u00a0 than the reference GTS 580 version at 772\/2001.<\/p>\n In our testing, we lowered the clocks of a pair of GTX 580s in SLI down to match the reference GTX 590 and we discovered that it generally used 150W less than the stock-clocked pair!\u00a0 And this was without even lowering the voltage.\u00a0 Of course, Nvidia has had quite a bit of time to tweak these cards and we find that the GTX 590 generally outperforms the underclocked GTX 580 SLI’s pair.\u00a0 We will draw attention to this in our regular testing of the individual games that we tested with underclocked SLI’s GTX 580s vs. GTX 590.<\/p>\n Of course, we know that GTX 580 SLI beats HD 6970 CrossFire and HD 6990.\u00a0 We do not know if the GTX 590 can manage the same feat as it rather underclocked.\u00a0 We also want to know if there is any more headroom to overclock it further; we suspect there is.<\/p>\n We see the new GPU supports the new HDMI 1.4a connector standard as there is one mini HDMI output and 3 DVI outputs.\u00a0 We can finally enjoy Nvidia’s Surround – their answer to AMD’s Eyefinity – on a single card.\u00a0 Unfortunately, we did not realize that all three displays must have the same native resolution, unlike Eyefinity which only requires that all three displays support a common resolution.\u00a0 Next time we will bring you Eyefinity vs. Surround.<\/p>\n The GTX 590 is much shorter than the AMD card and is physically able to fit in a lot more cases.\u00a0 No doubt the OEMs will really appreciate this.<\/p>\n Needless to say, the new Fermi Dual-GF110\u00a0 GTX 590 brings a lot of features to the table that current Nvidia customers will appreciate, including improved CUDA’s PhysX, 2D and 3D Surround to drive up to 3 LCDs with a single card, superb tessellation capabilities and two really fast GPUs in comparison to their GT200 series and even their hot running GF100 series cards.<\/p>\n Nvidia themselves believes that the GTX 590 is about 1.5 times faster than a single GTX 580.\u00a0 Since the GTX 580 retails for $500, the performance to value ration remains the same.\u00a0 Now you can even have ultimate performance with Quad-SLI – 2 times GTX 590 as long as you have the right motherboard.<\/p>\n Another reason to choose a GTX 590 would be where you only have one 16x PCIe slot available and you want the single fastest video card.\u00a0 Either the GTX 590 or the Radeon HD 6990 might fit the bill. But which one to choose?\u00a0 Read on.<\/p>\n If you have a top PC and $1500 to spend on graphics, you might consider Quad-SLI.\u00a0 It is supported by GTX 590 and there is improved scaling. There are also recently more compelling reasons besides increased performance to consider GTX 590 SLI which includes being able to experience Nvidia’s multi-display 3D Vision Surround. The only other way to experience similar but lesser performance is with 3 x GTX 580 Tri-SLI for about the same costs.<\/p>\n At this time, Nvidia does not support GTX 580 plus GTX 590 Tri-SLI although they could easily enable it in the drivers if they choose.\u00a0 It would give enthusiasts a cheaper way to experience Trr-SLI without a very expensive motherboard as the two card can be run together without a PCIe slot in-between as shown below.\u00a0 Notice the GeForce logo lights up on the GTX 590; a nice touch!<\/p>\n The GTX 580 was used with the GTX 590 as a dedicated PhysX card; the SLI bridge is only shown for illustration.\u00a0 Here is what the control panel looks like.<\/p>\n New Power Monitoring Hardware – or no more Furmark!<\/strong><\/p>\n <\/strong>Nvidia has added a power draw limitation system to their card beginning with the GTX 580 and also with the GTX 570 and GTX 560 Ti. When either Furmark or OCCT are detected, sensors measure the incoming current and voltage to calculate the total power draw. If the power draw exceeds a certain predetermined limit, the GTX 590 will automatically downclock to avoid damage to hardware components. After the power draw drops back to safe limits, the GPU returns to normal clocks much the same as with thermal management.<\/p>\n Because of this, we will no longer use Furmark for showing power draw and will return to using games to illustrate real world situations. Currently, this power management only switches on when Furmark or OCCT are detected and it should not limit overclocking unless Nvidia extends this management to regular PC games. Evidently this works by having the GeForce driver detect the program and treat it as a virus. In the case of the GTX 580 and the GTX 590, this power limiting circuitry implementation is mandatory by Nvidia’s parners.<\/p>\n As a total package, the new GTX 590 looks (and sounds) great! It looks and feels solid. Let’s show you the results of our one day (yes, we got it yesterday!) hand’s on test drive, shall we? We will put it to the test in 29 PC games and in three synthetic tests. But first, head to the next page to check out our test bed configuration. <\/span><\/p>\n Test Configuration<\/strong><\/span><\/span><\/span><\/p>\n <\/strong><\/span><\/p>\n Test Configuration – Hardware <\/strong><\/span><\/span><\/p>\n Test Configuration – Software<\/span><\/strong> <\/span><\/p>\n The Benchmarks<\/span><\/strong><\/span><\/strong> <\/span><\/strong><\/p>\n We have got an interesting project going. Let’s check our results and see if we can determine which card can wear “the world’s fastest video card” crown. <\/p>\n Vantage is Futuremark\u2019s DX10 test. It is really useful for tracking changes in a single <\/em>system \u2013 especially driver changes. There are two mini-game tests, Jane Nash<\/em> and Calico<\/em> and also two CPU tests, but we are still focusing on the graphics performance. Here is a scene from Vantage\u2019s second mini-game.<\/p>\n Let\u2019s go right to the graphs and first check the basic tests with the default benchmark scores:<\/p>\n We also underclocked our GTX 580 SLI pair to the same clocks as the GTX 590, 605\/1707MHz.\u00a0 We got a score of 33069, considerably lower than our GTX 590.\u00a0 We tested the vRAM clocks separately from the core clocks – it is a delicate balance to get the right performance.\u00a0 Overclocking only the core from 605 to 690MHz got us 41311, while leaving the core stock and overclocking the memory from 1707 to 1850MHz got a much lower score of 38701.\u00a0 Overclocking both to as far as they would work together optimally – 690\/1825MHz – got us 41592.\u00a0 The memory clocks need to go up as the core is raised – but not too much or there is instability.<\/p>\n For our purposes here, Vantage is a meaningless test although they do attempt to compare one video card’s performance to another by using two short timedemos. Let’s move on to the latest Futuremark benchmark, 3DMark11 which is DX11 only.<\/p>\n <\/p>\n 3DMark11 is Futuremark’s brand new DX11-only benchmark. We are keeping track of the overall (for these tests, meaningless) scores and the framerates of the 4 graphics tests (which are more meaningful). First the basic tests results.<\/p>\n (Above the HD 6990 does not score zero; correct is 9647<\/strong>)<\/em><\/p>\n We note the rankings. Unfortunately, scores are completely meaningless when they are presented in this way. However, the next set of tests actually measures framerates of four short timedemos that focus on DX11 graphics performance.<\/p>\n Let’s move on to PC games and to real world situations and we will create our “snapshot” of current performance of our two dual-GPU video cards to see how they scale in CrossFire or SLI and we will compare them to the single-GPU fastest video cards as well as to two each CrossFired and SLI’d configurations to attempt to determine price to performance.<\/p>\n <\/p>\n <\/em> <\/em> <\/em>F.E.A.R. \u2013 First Encounter Armed Assault<\/em> is a DX9c game by Monolith Productions that was originally released in October 2005 by Vivendi Universal Production. Later, there were two expansions with the latest, Perseus Mandate<\/em>, released in 2007. Although the game engine is aging, it still has some of the most spectacular effects of any game. F.E.A.R.<\/em> showcases a powerful particle system, complete with sparks and smoke for collisions as well as featuring bullet marks and other effects including \u201csoft shadows\u201d. This is highlighted by the built-in performance test, although it was never updated.<\/p>\n This performance test will tell you how F.E.A.R.<\/em> will run, but both of its expansions are progressively more demanding on your PC graphics and will run slower than the demo. We always run at least two sets of tests with all in-game features at \u2018maximum\u2019. F.E.A.R.<\/em> uses the Jupiter Extended Technology engine from Touchdown Entertainment. We test this game with the most demanding settings. We use fully maxed details with 4xAA\/16xAF; soft shadows \u2018off\u2019, as they do not play well with AA. Let\u2019s start first at 2560×1600:<\/p>\n We see an interesting situation. Here the HD 6870 as a single GPU drops down in the minimums. Adding a second HD 6870 in CrossFire allows it to easily play this game fluidly<\/em>. We also see the minimums nearly triple when a second card is added for CrossFire. However, generally we see scaling of the average framerates is not great with SLI nor CrossFire in F.E.A.R.<\/em><\/p>\n Let’s look at 1920×100:<\/p>\n<\/a><\/p>\n<\/a><\/p>\nBefore we do performance testing, let’s take a look at the original Fermi GF100 GTX 480 and quickly recap its new DX11 architecture and features of the original Fermi GF100 which we covered in our reviews of the GTX 480, published here,<\/span><\/a> here<\/a><\/span> and here<\/a><\/span>. Senior Editor BFG10K reviewed GTX 470 here<\/a> and here <\/a>and Senior editor MrK covered GTX 465 here<\/a>.<\/p>\nSpecifications<\/strong><\/h3>\n
<\/a><\/p>\nVapor Chamber Cooling<\/h3>\n
<\/a>Here is a good look behind the fan.<\/p>\n<\/a><\/p>\n<\/a><\/p>\n<\/a>Now lets compare the brackets and the connectors of the GTX 590 (above) with the HD 6990 (below).<\/p>\n<\/a><\/p>\nWHY choose a GTX 590?<\/h3>\n
<\/a><\/p>\nShould you SLI your GTX 590?<\/h3>\n
<\/a><\/p>\n<\/a><\/p>\n<\/a><\/p>\n\n
\n
\n
Vantage<\/strong><\/h2>\n
<\/a><\/p>\n<\/a>We note the rankings. Unfortunately, scores are completely meaningless when they are presented this way.<\/p>\n<\/a><\/p>\n3DMark11<\/h2>\n
<\/a><\/p>\n<\/a>We see an interesting lineup. Unfortunately for our purposes, 3DMark11 scores are just as meaningless as Vantage in an attempt to compare one video card’s performance to another – even in the same system.<\/p>\nF.E.A.R.<\/strong><\/em><\/h2>\n
<\/a> <\/em><\/p>\n<\/a><\/p>\n