Archive Home arrow Reviews: arrow Video Cards arrow NVIDIA GeForce GTX 570 Video Card Performance
NVIDIA GeForce GTX 570 Video Card Performance E-mail
Reviews - Featured Reviews: Video Cards
Written by Olin Coles   
Tuesday, 07 December 2010
Table of Contents: Page Index
NVIDIA GeForce GTX 570 Video Card Performance
GeForce GTX 570 Closer Look
GeForce GTX 570 Detailed
Features and Specifications
Video Card Testing Methodology
DX10: 3DMark Vantage
DX10: Crysis Warhead
DX11: Aliens vs Predator
DX11: Battlefield Bad Company 2
DX11: BattleForge
DX11: Lost Planet 2
DX9 SSAO: Mafia II
DX11: Metro 2033
DX11: Tom Clancy's HAWX2
DX11: Unigine Heaven 2.1
GeForce GTX 570 Temperatures
VGA Power Consumption
NVIDIA GeForce GTX 570 Conclusion

NVIDIA Fermi Features

In today's complex graphics, tessellation offers the means to store massive amounts of coarse geometry, with expand-on-demand functionality. In the NVIDIA GF100-series GPU, tessellation also enables more complex animations. In terms of model scalability, dynamic Level of Detail (LOD) allows for quality and performance trade-offs whenever it can deliver better picture quality over performance without penalty. Comprised of three layers (original geometry, tessellation geometry, and displacement map), the final product is far more detailed in shade and data-expansion than if it were constructed with bump-map technology. In plain terms, tessellation gives the peaks and valleys with shadow detail in-between, while previous-generation technology (bump-mapping) would give the illusion of detail.

id-imp-tessellated-character.jpg

Using GPU-based tessellation, a game developer can send a compact geometric representation of an object or character and the tessellation unit can produce the correct geometric complexity for the specific scene. Consider the "Imp" character illustrated above. On the far left we see the initial quad mesh used to model the general outline of the figure; this representation is quite compact even when compared to typical game assets. The two middle images of the character are created by finely tessellating the description at the left. The result is a very smooth appearance, free of any of the faceting that resulted from limited geometry. Unfortunately this character, while smooth, is no more detailed than the coarse mesh. The final image on the right was created by applying a displacement map to the smoothly tessellated third character to the left.

Benchmark Reviews also more detail in our full-length NVIDIA GF100 GPU Fermi Graphics Architecture guide.

Tessellation in DirectX-11

Control hull shaders run DX11 pre-expansion routines, and operates explicitly in parallel across all points. Domain shaders run post-expansion operations on maps (u/v or x/y/z/w) and is also implicitly parallel. Fixed function tessellation is configured by Level of Detail (LOD) based on output from the control hull shader, and can also produce triangles and lines if requested. Tessellation is something that is new to NVIDIA GPUs, and was not part of GT200 because of geometry bandwidth bottlenecks from sequential rendering/execution semantics.

In regard to the GF110 graphics processor, NVIDIA has added a new PolyMorph and Raster engines to handle world-space processing (PolyMorph) and screen-space processing (Raster). There are sixteen PolyMorph engines and four Raster engines on the GF110, which depend on an improved L2 cache to keep buffered geometric data produced by the pipeline on-die.

GF100 Compute for Gaming

As developers continue to search for novel ways to improve their graphics engines, the GPU will need to excel at a diverse and growing set of graphics algorithms. Since these algorithms are executed via general compute APIs, a robust compute architecture is fundamental to a GPU's graphical capabilities. In essence, one can think of compute as the new programmable shader. GF110's compute architecture is designed to address a wider range of algorithms and to facilitate more pervasive use of the GPU for solving parallel problems. Many algorithms, such as ray tracing, physics, and AI, cannot exploit shared memory-program memory locality is only revealed at runtime. GF110's cache architecture was designed with these problems in mind. With up to 48 KB of L1 cache per Streaming Multiprocessor (SM) and a global L2 cache, threads that access the same memory locations at runtime automatically run faster, irrespective of the choice of algorithm.

NVIDIA Codename NEXUS brings CPU and GPU code development together in Microsoft Visual Studio 2008 for a shared process timeline. NEXUS also introduces the first hardware-based shader debugger. NVIDIA's GF100-series is the first GPU to ever offer full C++ support, the programming language of choice among game developers. To ease the transition to GPU programming, NVIDIA developed Nexus, a Microsoft Visual Studio programming environment for the GPU. Together with new hardware features that provide better debugging support, developers will be able enjoy CPU-class application development on the GPU. The end results is C++ and Visual Studio integration that brings HPC users into the same platform of development. NVIDIA offers several paths to deliver compute functionality on the GF110 GPU, such as CUDA C++ for video games.

Image processing, simulation, and hybrid rendering are three primary functions of GPU compute for gaming. Using NVIDIA's GF100-series GPU, interactive ray tracing becomes possible for the first time on a standard PC. Ray tracing performance on the NVIDIA GF100 is roughly 4x faster than it was on the GT200 GPU, according to NVIDIA tests. AI/path finding is a compute intensive process well suited for GPUs. The NVIDIA GF110 can handle AI obstacles approximately 3x better than on the GT200. Benefits from this improvement are faster collision avoidance and shortest path searches for higher-performance path finding.

GF110 Specifications

  • 512 CUDA Cores
  • 16 Geometry Units
  • 4 Raster Units
  • 64 Texture Units
  • 48 ROP Units
  • 384-bit GDDR5
  • DirectX-11 API Support

GeForce GTX-Series Products

Graphics Card

GeForce GTX 460

GeForce GTX 470

GeForce GTX 480

GeForce GTX 570 GeForce GTX 580
GPU Transistors 1.95 Billion 3.2 Billion 3.2 Billion 3.0 Billion 3.0 Billion

Graphics Processing Clusters

2

4

4

4

4

Streaming Multiprocessors

7

14

15

15 16

CUDA Cores

336

448

480

480 512

Texture Units

56

56

60

60 64

ROP Units

768MB=24 / 1GB=32

40

48

40 48

Graphics Clock
(Fixed Function Units)

675 MHz

607 MHz

700 MHz

732 MHz 772 MHz

Processor Clock
(CUDA Cores)

1350 MHz

1215 MHz

1401 MHz

1464 MHz 1544 MHz

Memory Clock
(Clock Rate/Data Rate)

900/3600 MHz

837/3348 MHz

924/3696 MHz

950/3800 MHz 1002/4016 MHz

Total Video Memory

768MB / 1024MB GDDR5

1280MB GDDR5

1536MB GDDR5

1280MB GDDR5

1536MB GDDR5

Memory Interface

768MB=192 / 1GB=256-Bit

320-Bit

384-Bit

320-Bit

384-Bit

Total Memory Bandwidth

86.4 / 115.2 GB/s

133.9 GB/s

177.4 GB/s

152.0 GB/s 192.4 GB/s

Texture Filtering Rate
(Bilinear)

37.8 GigaTexels/s

34.0 GigaTexels/s

42.0 GigaTexels/s

43.9 GigaTexels/s

49.4 GigaTexels/s

GPU Fabrication Process

40 nm

40 nm

40 nm

40 nm

40 nm

Output Connections

2x Dual-Link DVI-I
1x Mini HDMI

2x Dual-Link DVI-I
1x Mini HDMI

2x Dual-Link DVI-I
1x Mini HDMI

2x Dual-Link DVI-I
1x Mini HDMI

2x Dual-Link DVI-I
1x Mini HDMI

Form Factor

Dual-Slot

Dual-Slot

Dual-Slot

Dual-Slot

Dual-Slot

Power Input

2x 6-Pin

2x 6-Pin

6-Pin + 8-Pin

2x 6-Pin

6-Pin + 8-Pin

Thermal Design Power (TDP)

768MB=150W / 1GB=160W

215 Watts

250 Watts

219 Watts 244 Watts

Recommended PSU

450 Watts

550 Watts

600 Watts

550 Watts

600 Watts

GPU Thermal Threshold

104°C

105°C

105°C

97°C

97°C

GeForce Fermi Chart Courtesy of Benchmark Reviews



 

Comments 

 
# No award ?Succellus 2010-12-07 08:46
No award and no final rates ?
Report Comment
 
 
# RE: No award ?Olin Coles 2010-12-07 08:53
I'm experimenting. I occasionally get complaints about our ratings, so with this review you guys can draw your own conclusion.
Report Comment
 
 
# RE: NVIDIA GeForce GTX 570 Video Card Performanceaberkae 2010-12-07 09:08
High end gpu for $350, now we are starting to see the affects of gpu wars!
Report Comment
 
 
# RE: NVIDIA GeForce GTX 570 Video Card PerformanceRealNeil 2010-12-07 09:49
Very compelling video card here. This looks to be a good replacement for the pair of GTX460's I was going to get. Maybe I'll get a pair of them,...(one for each box)...Ha-Ha! For the performance rendered, it really IS a good price point.
Report Comment
 
 
# sweet deal !LASER 2010-12-07 13:39
thnks guys for fast review.. its 3Am but i dont really sleep so readind reviews..

i didnt thought 152hp machine can beat 177hp.... but yea it matches it and so much price difference...
totally stunning card compared to older price/perfor ratios by Nvidia...
i was going for GTX 580 but now i gotta think 170$ extra for 6fps in heavy games...

Finally i gotta again wait for price downs on GTX 580 ... or i will go for GTX 570 which is sweet deal !!
Report Comment
 
 
# RE: NVIDIA GeForce GTX 570 Video Card PerformanceServando Silva 2010-12-07 19:01
Is it possible to update the article with GTX 460 SLI results? Many people are still thinking about getting a pair of 460s.
Thanks for the article!
Report Comment
 
 
# more than agreeresere 2010-12-08 06:51
call us lazy or whatever, but it would be nice to see in the same article the results of both 460 SLI/ 570.
Anyway good job once again.
Report Comment
 
 
# BF: BC2 resolutionBunzing 2010-12-08 05:32
The Bad Company benchmark is probably done a 1920x1080 resolution, but I just want to no for sure. Thanks

Lookin like a great card btw. First time in years that i'm looking to buy a nvidia card instead of an ati. I sent thermaltake a mail to see wether the awesome coolingapparatice ISGC-V320 will fit (believe me it's a great cooler, but has to be supported some way because of the enormous weight).
Report Comment
 
 
# Old School Gaming MonitorBruceBruce 2010-12-08 07:29
The second set of results (on the right) are for 1920 x 1200. The world has gone 1080P crazy, but we still like the "extra" 20% of vertical view on the old-school 1920 x 1200 monitors.
Report Comment
 
 
# RE: BF: BC2 resolutionOlin Coles 2010-12-08 08:12
I've updated that chart to show the test resolutions. Sorry for the confusion!
Report Comment
 
 
# BFBC 2 BenchmarkManuKey 2010-12-08 07:08
The result on BFBC2 are WAY to high... especially with 8X AA and HBAO...
Report Comment
 
 
# You haven't been paying attentionBruceBruce 2010-12-08 07:23
Take a look at some other reviews, and you'll see that these results are typical. In my recent testing, I'm getting 1-2 FPS higher than this, with a P55-based system. See my MSI Radeon HD 6870 article on this site for examples:

GTX 460 - 47.5 FPS
HD 6870 - 55.7 FPS
HD 5870 - 65.9 FPS
Report Comment
 
 
# Resolution for hi-res testingBruceBruce 2010-12-08 07:31
The results above are for 1920 x 1200 screen resolution. Sorry, forgot to mention that.
Report Comment
 
 
# BFBC 2 Benchmark --> SPManuKey 2010-12-08 11:43
Anyway I juste realised that this is a Singleplayer benchmark (for a multiplayer oriented game... lol)... Multiplayer w/ 32+ players are way more demanding!
Report Comment
 
 
# RE: BFBC 2 Benchmark --> SPOlin Coles 2010-12-08 12:00
Actually, the frame rates are about the same. I've tested it both ways, and the multiplayer maps generally produce the same FPS scores.
Report Comment
 
 
# RE: RE: BFBC 2 Benchmark --> SPManuKey 2010-12-08 12:22
I think you're the first I hear saying that! just check the forums... in general, people seem's to get better performances in Singleplayer than multi...
Report Comment
 
 
# Higher ping=Lower FPSNator 2010-12-09 12:13
Usually this computes. Ergo, playing online gives lower FPS.
Report Comment
 
 
# RE: Higher ping=Lower FPSOlin Coles 2010-12-09 18:35
I think you're confusing network latency (lag) for video frame rate (FPS). Lag might appear to display choppy frames, but it's because of a network issue and not the video card. Most people avoid playing first person shooter games with a ping over 60ms.
Report Comment
 
 
# temps____ 2010-12-08 18:52
The temps are higher than the 580!? Are you sure the 520 is not throttled? I can't think of anything else.
Report Comment
 
 
# RE: tempsOlin Coles 2010-12-08 19:20
I'm going to presume you meant the 570, and not 520.

The temperature difference still has me a little baffled. The cooling solution is the same, and the test methodology was identical. Yet, somehow, the GeForce GTX 570 sample I received heats up more than both the GTX 580 I have.

If I hadn't retested three different times, I would still think there was something amiss. It is what it is.
Report Comment
 
 
# RE: RE: tempsTim 2010-12-09 12:00
My GTX 580 heats up above 80 degrees under load so in that way its similar to 570
Report Comment
 
 
# RE: RE: RE: tempsOlin Coles 2010-12-09 12:10
No, it doesn't. Our reference GTX 580 did not heat up to 80 degrees, it produced only 70C under full load. This was explained in the 'GeForce GTX 570 Temperatures' section of this article.
Report Comment
 
 
# Could just be a very well produced 580Nator 2010-12-09 12:16
Temperatures can vary a bit between cards, even though they appear the same, its all down to the measuring stick, and a little bit to how good the contact between cooler and gpu/ram is. The temp. sensors can be off by a ways sometimes.
Report Comment
 
 
# RE: NVIDIA GeForce GTX 570 Video Card PerformanceTim 2010-12-10 00:42
I use a Gigabyte gtx 580. First I tried it in a 2008 Mac Pro (running Windows 7) For some reason the cooling didn't work properly and the card went too hot and fail (above 100 degrees). Putting the card in a i7 system and it does work ok. Idle temps are around 48 degrees. Playing BFBC2 temp goes up to about 82.
Running Adobe apps / premiere pro MPE temps are about 55 degrees.
Report Comment
 
 
# RE: RE: NVIDIA GeForce GTX 570 Video Card PerformanceOlin Coles 2010-12-10 08:19
Is your Gigabyte GTX 580 overclocked at all? I'm just curious, because our overclocked ASUS GTX 580 could reach 80C but the reference card only produced 70C.
Report Comment
 
 
# RE: RE: RE: NVIDIA GeForce GTX 570 Video Card PerformanceTim 2010-12-11 00:51
No it is a stock Gigabyte GTX 580. I am talking degrees in Celcius no Fahrenheit :-)
Report Comment
 
 
# GTX480 Max Tempaaron686 2010-12-10 22:08
NVIDIA GeForce GTX 480 - 82°C (max)

How can the 480 have an absolute max temp of 82C? I've seen this card go to 95 - 100 easily, having both the 480 and the 570 at the same max temp is baffling.
Report Comment
 
 
# RE: NVIDIA GeForce GTX 570 Video Card PerformanceTim 2010-12-11 00:53
By the way, running Vray RT v2 GPU mode, the temp goes up to 86 degrees Celsius.
Report Comment
 
 
# RE: NVIDIA GeForce GTX 570 Video Card PerformanceLaszlo 2010-12-13 00:46
I would like to see the 570's preformance against 470 with same GPU, shader and mem clock speed (and memory timings too). Then - I think - that the differences would be almost nothing, only that little which comes from 448 vs 480 CUDA cores.
I searched for overclocking maximum results and not surprisingly the fastest 470 results clocked about 900+ (air, stock cooling) MHz - just the 570 results I found so far. With my lame mind (as I'm not an insider in the hw industry) I think that the ~900 is a limit that is nvidia made (with their 'protection'), not the limit that comes from the 40nm manufacturing process. And this things clearly show that the 570's higher clocks doesn't come from the more refined manufacturing (as older gen CPUs got more overclockable when their newer revisions came out) just from marketing strategy. Like you design a Porsche GT3, but first you sell a version to the public which is limited just a bit higher than the Boxter's (and call it Boxter2). Then playing that you worked hard (but did nothing), you limit the product less, say 911 Turbo. But literally the same product.
I will wait for 1: when the prices starting to drop; 2: the next gen.
Report Comment
 
 
# RE: RE: NVIDIA GeForce GTX 570 Video Card PerformanceLaszlo 2010-12-13 01:59
Sorry, 800+, not 900+ (the GPU overclock). Mem-wise: if the gddr5 can be clocked as high as 1200 on the 5870, then maybe NV have some surplus GPU power on their shelf to unleash (as the 512 bit wide membus too - remember GTX285). Not too surprising that mem can be pushed ~1150. My ex8800 GTS512's mem - which is GDDR3 - was stable on that clk, and if it's 65nm GPU was stable on 750 (750/1950/1150 @1,15V), then a two (55nm then the current 40nm) gen later 40nm chip's 600-800 MHz clk is somewhat funny.
Report Comment
 

Comments have been disabled by the administrator.

Search Benchmark Reviews
QNAP Network Storage Servers

Follow Benchmark Reviews on FacebookReceive Tweets from Benchmark Reviews on Twitter