Report: NVIDIA GeForce GTX 550 Ti and Asus DirectCU TOP - BeHardware
>> Graphics cards
Written by Damien Triolet
Published on March 17, 2011
Following the GeForce GTX 460, it’s now the GeForce GTS 450’s turn to be reworked by NVIDIA. As with the other reworked cards in the GeForce 500 range, the resultant GeForce GTX 550 Ti enjoys higher clocks and higher performance. Does this mean NVIDIA now have a more competitive model to face down the Radeon mid-range? We’re going to try and answer this by giving our analysis of the DirectCU TOP model from Asus.
The reference GeForce GTX 550 Ti.
GF116 and heterogeneous memoryThe GF116 is to the GF106 what the GF114 is to the GF104. In other words, it’s not a new GPU but rather a new revision of the chip designed to reduce current leakage and facilitate clock increases. To recap, here are the main points of the internal architecture of the various GPUs in the GeForce 400 and 500 ranges:
- GF110 = 4 GPCs x 4 SMs x (32 processing units + 8 texturing texturing units)
- GF100 = 4 GPCs x 4 SMs x (32 processing units + 4 texturing units)
- GF104/114 = 2 GPCs x 4 SMs x (48 processing units + 8 texturing units)
- GF106/116 = 1 GPC x 4 SM x (48 processin units + 8 texturing units)
The GF106/116 is basically half a GF104/114, except that it has a 192-bit memory bus as against a 256-bit bus. While the GeForce GTS 450 had to make do with a cut down 128-bit bus, the GeForce GTX 550 Ti benefits from a 192-bit bus in total… but with 1 GB of memory. How can this be? In principle a 192-bit memory bus should function with 768 MB or 1.5 GB of video memory.
NVIDIA says that it has slightly reworked its memory controller here to support asymmetric memory density. More or less all the GPUs run with one or more 64-bit memory controllers. In the case of the GF116, three of these controllers share access to the video memory. The innovation comes from the fact that it is now possible to use chips with varying density for each controller. Here’s the typical configuration for the GeForce GTX 550 Ti:
- memory controller 0: 2x 128 MB
- memory controller 1: 2x 128 MB
- memory controller 2: 2x 256 MB
The GeForce GTX 550 Ti therefore has four 128 MB chips and two more 256 MB chips. In practice, its GPU thus has a ‘high speed’ 768 MB memory that’s accessible at 192 bits and a ‘slow’ additional 256 MB accessible at 64 bits. We therefore suppose that NVIDIA have worked mainly on the software side and that the correct functioning of the system is based on the ability of the drivers to put the most performance dependent data in the part of the memory that is accessible at 192 bits.
You’ll note that NVIDIA has decided to revisit its naming structure for this new model. Why they’re using the Ti suffix still isn’t very clear, but we imagine that it will be reserved for the biggest models in each series should other models see the light of day.
Specifications, the test
With a higher GPU clock, the GeForce GTX 550 Ti has 15% more processing power than the GeForce GTS 450 with a 70% higher bandwidth thanks to its larger memory bus!
The testFor this test, we opted for 3 resolutions: 1680x1050 without and with MSAA 4x and 1920x1200 without MSAA, with the exception of Crysis Warhead and Far Cry 2 which were tested in more modes. We set the graphics options high but not at maximum in the most demanding games.
We have also decided to stop showing decimals in game performance results so as to make the graph more readable. We nevertheless note these values and use them when calculating the index. If you’re observant you’ll notice that the size of the bars also reflects this.
The Radeons and the GeForces were tested with texture filtering at the high quality setting. All the Radeons were tested with the Catalyst 11.4 beta (8.84.3 beta 2) driver and the GeForces with the beta 267.59 drivers.
Test configurationIntel Core i7 980X (HT deactivated)
Asus Rampage III Extreme
6 GB DDR3 1333 Corsair
Windows 7 64 bits
Forceware 267.59 beta
Catalyst 11.4 beta (8.84.3 beta2).
Asus GTX 550 Ti DirectCU TOP
The Asus GeForce GTX 550 Ti DirectCU TOPAsus supplied us with its DirectCU TOP:
The Asus GeForce GTX 550 Ti DirectCU TOP is based on a similar design to that of the Radeon HD 6850 in the same series that we tested recently. Its PCB is of course different and shorter, but is still longer than the one used on the reference GeForce GTX 460 and GTS 450: 22.5 cm against 21 cm. The Asus card is therefore similar in size to the reference GeForce GTX 560 Ti. Asus has gone for a power stage with 4 phases for the GPU, but without a radiator.
As usual, Asus has placed a reinforcement bar on top of the card to strengthen it as its cooler is only attached using four screws, one in each corner of the GPU. As the cooler name indicates, two copper heatpipes are in direct contact with the GPU. They then run across an aluminium radiator that’s cooled by a central fan. The design doesn’t therefore extract hot air from the casing but disperses it without making too much noise.
The Hynix GDDR5 memory is certified at 1.25 GHz (2.5 GHz for data). As we said, there are four 1 Gb modules (128 MB) and two 2 GB modules (256 MB) which make up a memory space of 1 GB interfaced at 192 bits. Note that the back of the PCB provides for a doubling of this memory space to 2 GB.
In terms of connectivity, Asus has gone for a DVI out, an HDMI out and a VGA out. Just one 6-pin power supply connector is required.
Of course, as a TOP model, it has been overclocked: 975 MHz for the GPU against 900 MHz for the reference clocks (a gain of 8%). However the memory is not overclocked. We managed to clock the GPU up to 1050 MHz and here’s the gain we got in Crysis Warhead at 1920x1200:
Asus GeForce GTX 550 Ti DirectCU TOP @ 900 MHz: 31.5 fps
Asus GeForce GTX 550 Ti DirectCU TOP @ 975 MHz: 33.6 fps
Asus GeForce GTX 550 Ti DirectCU TOP @ 1050 MHz: 35.6 fps
The Asus TOP therefore gives a 7% gain on the reference card with another 6% when overclocked to give a total of 13%.
Although on sale at no less than #170, Asus has only provided an entry level style bundle: an installation guide, a CD for the drivers and a a double molex to 6-pin PCI Express power supply cable convertor.
Energy consumption, noise
Energy consumptionWe measured the energy consumption of the graphics card on its own. We took these readings at idle, in 3D Mark 06 and Furmark. Note that we use a version of Furmark that isn’t detected by the stress test energy consumption limitation mechanism put into place by NVIDIA in the GeForce GTX 500 drivers.
At idle, all the GeForces tested draw a similar amount of power and are very well managed.
In load we measured the GTX 550 Ti a good deal higher than the GeForce GTS 450 (between 24 and 29% higher). Note that while the Asus model is overclocked, the Sparkle Calibre that we’ve included in the graph is too. This brings power draw on the GeForce GTX 550 Ti close to that of the GeForce GTX 460, giving a big advantage to the Radeons which use much less energy.
NoiseWe put the cards in an Antec Sonata 3 casing and measured noise at idle and in load. We placed the sonometer 60 cm from the casing.
Asus has calibrated its DirectCU cooling system pretty well, giving a card that’s quiet at idle and not at all noisy in load.
TemperaturesStill in the same casing, we took a temperature reading of the GPU using internal sensors:
The GF116 is well cooled with similar temperatures to those of a GeForce GTX 460.
Here’s what the infrared thermography image shows:
Asus GTX 550 Ti DirectCU TOP at idle
Asus GTX 550 Ti DirectCU TOP in load
At idle, you can see a hot point where the controller is on the power stage, while in load the whole of the power stage gets hotter. It’s not problematic though a little dedicated radiator would probably have brought these temperatures down.
Gains with the 192-bit bus
Gains with the 192-bit busTo get a handle on the gains given by the singular memory architecture introduced with the GeForce GTX 550 Ti (1 GB of memory on a 192-bit bus), we clocked the Sparkle GeForce GTS 450 Calibre at the same clocks as the GeForce GTX 550 Ti. This gave us a comparison between two solutions that differ only in terms of their memory bus, one 128-bit bus and one 192-bit bus, and allowed us to check if the additional bandwidth on the GeForce GTX 550 Ti was making much of a difference.
The gains vary a great deal from one game to another. There’s absolutely no gain at all in BattleForge for example, but in Metro 2033 we measured a 20% gain. The gains with antialiasing were also bigger than without, which makes sense as this mode is more demanding in terms of memory bandwidth and suffers less from the pixel fillrate limitation associated with Fermi architecture.
To recap, a bottleneck limits the amount of data that can be transmitted from the processing units to the ROPs and memory controllers. Only 2 simple (non HDR) pixels can be transmitted per SM and per cycle, which makes it difficult to exploit all the available memory bandwidth fully. Using antialiasing however doesn’t limit this throughput as the additional load comes at ROP level. This is why the major benetits of the GeForce GTX 550 Ti 192-bit memory bus come with antialiasing.
We are nevertheless a long way from the theoretical gain of 50%, partly because of the limitation mentioned above and partly because GPU processing power is a bigger limiting fact in many situations. In other words, the memory bandwith available on the GeForce GTX 550 Ti is somewhat disproportionate to the power of the GPU.
Given the resolutions tested, we do however reckon that the asymmetric organisation of the memory isn’t a disadvantage compared to the standard use of a 192-bit bus.
To test Starcraft 2, we launched a replay and measured performances following one player’s view.
All graphics settings were pushed to a maximum. The game doesn’t support antialiasing which is therefore activated in the control panels of the AMD and NVIDIA drivers. Patch 1.0.3 has been installed.
The GeForces do a bit better than the Radeons in this first test. At MSAA 4x, the GeForce GTX 550 Ti has a 38% advantage over the GeForce GTS 450, which demonstrates how useful the additional memory bandwidth is under these conditions.
The Mafia II engine passes physics handling over to the NVIDIA PhysX libraries and takes advantage to offer high physics settings which can be partially accellerated by the GeForces.
To measure performances we used the built-in benchmarks and all graphics options were pushed to a maximum, first without activating PhysX effects accelerated by the GPU:
Overall the GeForce GTX 550 Ti is on a par with the Radeon HD 5770.
Next, we set all PhysX options to high:
With PhysX effects pushed to a maximum, performance levels dive. Note that they are in part limited by the CPU, as not all additional PhysX effects are accelerated. Of course the Radeons remain a long way behind.
Crysis Warhead replaces Crysis and has the same resource-heavy graphics engine. We tested it in version 1.1 and 64-bit mode as this is the main innovation. Crytek has renamed the different graphics quality modes, probably so as not to dismay gamers who may be disappointed at not being able to activate the highest quality mode because of excessive demands on system resources. The high quality mode has been renamed as “Gamer” and the very high is called “Enthusiast”. This is the one we tested.
The GeForce GTX 550 Ti is on an equal footing with the Radeon HD 5770 except when it comes to MSAA 8x, where the Radeons have the advantage.
Far Cry 2
Far Cry 2
This version of Far Cry isn’t really a great development as Crytek made the first episode in any case. As the owner of the licence, Ubisoft handled its development, with Crytek working on Crysis. No easy thing to inherit the graphics revolution that accompanied Far Cry, but the Ubisoft teams have done pretty well, even if the graphics don’t go as far as those in Crysis. The game is also less resource heavy which is no bad thing. It has DirectX 10.1 support to improve the performance levels of compatible cards. We installed patch 1.02 and used the “ultra high” quality graphics mode.
The GeForces do particularly well with Far Cry 2.
H.A.W.X. is a flying action game. It uses a graphics motor that supports DirectX 10.1 to optimise results. Among the graphics effects it supports, note the presence of ambient occlusion that’s pushed to a max along with all other options. We use the built-in benchmark and patch 1.2 was installed.
The GeForce GTX 550 Ti, as is often the case with the GeForce 400/500s, does better with activation of 4x antialiasing than higher resolutions.
The first game with DirectX 11, or more precisely Direct3D 11 support, we couldn’t not test BattleForge. An update added in September 2009 gave support for Microsoft’s new API.
Compute Shaders 5.0 are used by the developers to accellerate SSAO processing (ambient occlusion). Compared to standard implementation, via the Pixel Shaders, this technique allows more efficient use of the available processing power by saturating the texturing units less. BattleForge offers two SSAO levels: High and Very High. Only the second, called HDAO (High Definition AO), uses Compute Shaders 5.0.
We used the game’s bench and installed the latest available update (1.2 build 304941).
The Radeon HD 5770 overtakes the GeForce GTX 550 Ti here which only comes back at it with 4x antialiasing.
Pretty successful visually, Civilization V uses DirectX 11 to improve quality and optimise performance in the rendering of terrains thanks to tessellation and in implementing a special compression of textures thanks to the compute shaders. This compression allows it to retain the scenes of all leaders in the memory. This second usage of DirectX 11 doesn’t concern us here however as we used the benchmark integrated on a game card. We zoom in slightly so as to reduce the CPU limitation which has a strong impact in this game.
All settings were pushed to a max and we measured performance with shadows and reflections. Patch 1.2 was installed.
The GeForces do very well here. While the GTS 450 is equal to the Radeon HD 5770, the GeForce GTX 550 Ti has the advantage and is hard on the heels of the Radeon HD 6850.
S.T.A.L.K.E.R. Call of Pripyat
S.T.A.L.K.E.R. Call of Pripyat
This new S.T.A.L.K.E.R. suite is based on a new development of the graphics engine which moves up to version 1.06.02 and supports Direct3D 11 which is used both to improve performance and quality, with the option to have more detailed light and shade as well as tessellation support.
High quality mode was used and we activated tessellation. The game doesn’t support 8x antialiasing. Our test scene is 50% outside and 50% inside and inside you’ll find yourself facing several characters.
The Radeon HD 5770 and GeForce GTX 550 Ti are pretty much on a par in Call of Pripyat, with the Radeon having a small advantage at 1920x1200, and the GeForce with MSAA 4x.
The latest Codemaster title, F1 2010 uses the same engine as DiRT 2 and supports DirectX 11 via patch 1.1 that we installed. As this patch was developped in collaboration with AMD, NVIDIA told us that they had only received it late in the day and haven’t yet had the opportunity to optimise its drivers for the game in the DirectX 11 version.
We pushed all the graphics options to a max and we used the game’s own test tool on the Spa-Rancorchamps circuit with a single F1.
In F1 2010, the Radeons are particularly at ease and the Radeon HD 5770 ends up at the same level as the GeForce GTX 460 1 GB. The GeForce GTX 550 Ti is slightly behind the Radeon HD 5750 here.
Probably the most demanding title right now, Metro 2033 forces all recent graphics cards to their knees. It supports GPU PhysX but only for the generation of particles during impacts, a rather discreet effect that we therefore didn’t activate during the tests. In DirectX 11 mode, performance is identical to DirectX 10 mode but with two additional options: tessellation for characters and a very advanced, very demanding depth of field feature.
We tested it in DirectX 11 mode, at a very high quality level and with tessellation activated, both with and without 4x MSAA.
The GeForce GTX 550 Ti is on a level here with the Radeon HD 5770.
Performance recapAlthough individual game results are worth looking at, we have calculated a performance index based on all tests with the same weight for each game. Mafia II is included with the scores obtained without GPU PhysX effects.
We attributed an index of 100 to the GeForce GTX 460 1 GB at 1680x1050:
Hold the mouse over the graph to view the cards by performance at 1680x1050.
It’s interesting to see that AMD and NVIDIA are on an equal footing in several market segments: GeForce GTX 460 1 GB and Radeon HD 6850, GeForce GTS 450 and Radeon HD 5750… and GeForce GTX 550 Ti and Radeon HD 5770.
Another thing that comes out in this analysis is a trend we’ve noted with the GeForce 400s and 500s: they suffer a little less when antialiasing is activated, but a little more at higher resolutions, which is explained by the reduced fillrate given by their internal architecture.
Although the GeForce GTX 550 Ti gives a nice gain over the GeForce GTS 450, of 19% without MSAA and 23% with, this isn’t enough to bring the GeForce GTX 460 into range – the GeForce GTX 460 1 GB still has an advantage of almost 30%! The GeForce GTX 560 Ti gives 70% more than the newcomer in terms of performance, which creates quite a gap in the new NVIDIA range.
The little GeForce GTX 460 SE also gives higher performance than the GeForce GTX 550 Ti. Moreover, it equals the GeForce GTX 460 768 MB with antialiasing which is penalised by a lack of memory in a few cases.
Note finally that the Asus GeForce GTX 550 Ti DirectCU TOP gives a gain of almost 7% on the reference card with a GPU clock that is 8% higher. As the memory clock hasn’t been increased, the Asus has just a 5% gain with MSAA 4x.
ConclusionOnce again, NVIDIA is having trouble getting to grips with the Radeon 5770 and 5750, which are likely to be renamed as the Radeon HD 6770 and 6750 soon. A heavier architecture and advantages which diminish as you go down the range mean the latest GeForce GPUs are rather toothless when faced with the performance / price ratio you get on certain Radeons. These mid and lower end NVIDIA cards struggle to support 3D Vision and PhysX and it is therefore harder to justify their higher pricing.
With this GeForce GTX 550 Ti, which looked pretty good at first, it’s difficult not to feel that NVIDIA has tried to retain too high a profit margin at the cost of competitiveness, no doubt hoping that the high end aura of a GTX Ti card will help customers swallow the pill. Even if you ignore the ferocious competition from AMD in this segment, the new GeForce gives a lower price/performance ratio than the GeForce GTS 450 and GTX 460, without introducing any other advantages.
Announced at €140, with some versions that could go as low as €130, the GeForce GTX 550 Ti is a good deal more expensive than the Radeon HD 5770, which gives a similar level of performance. It also suffers at the hands of the cheaper GeForce GTX 460s, which are a notch above them in terms of performance. If NVIDIA wants to make this GeForce GTX 550 Ti competitive, they’ll need to bring its price down by €20.
As things stand, we can only advise you to look at the GeForce GTX 460 1 GB or a well-priced Radeon HD 6850, both gamer references, or the Radeon HD 5700 if you have a more limited budget. The Radeon HD 5700 will also give you a much better energy yield in load.
Looking at the customised and overclocked cards from NVIDIA’s partners, models such as the Asus GTX 550 Ti DirectCU TOP, announced at €170, unfortunately suffer even more in comparison with the GeForce GTX 460, which is similarly priced but gives much higher performance.
Copyright © 1997-2014 BeHardware. All rights reserved.