Archive Home arrow Reviews: arrow Video Cards arrow ASUS GeForce GT 430 Fermi GF108 Video Card
ASUS GeForce GT 430 Fermi GF108 Video Card E-mail
Reviews - Featured Reviews: Video Cards
Written by Hank Tolman   
Monday, 11 October 2010
Table of Contents: Page Index
ASUS GeForce GT 430 Fermi GF108 Video Card
ASUS ENGT430 Features
NVIDIA GF108 Specifications
Video Card Testing Methodology
DX10: 3DMark Vantage
DX10: Left 4 Dead 2
DX10: Street Fighter IV
DX10: Far Cry 2
DX10: Resident Evil 5
DX11: Aliens vs Predator
DX11: Battlefield Bad Company 2
DX11: Lost Planet 2
DX11: Unigine Heaven 2.2
ASUS ENGT430 Final Thoughts
ASUS ENGT430 GeForce GT 430 Conclusion

NVIDIA GF108 GPU Fermi Architecture

Based on the Fermi architecture, NVIDIA's latest GPU is codenamed GF108 and is equipped on the GeForce GT 430. In this article, Benchmark Reviews explains the technical architecture behind NVIDIA's GF108 graphics processor and offers an insight into upcoming Fermi-based GeForce video cards. For those who are not familiar, NVIDIA's GF100 GPU was their first graphics processor to support DirectX-11 hardware features such as tessellation and DirectCompute, while also adding heavy particle and turbulence effects. The GF100 GPU is also the successor to the GT200 graphics processor, which launched in the GeForce GTX 280 video card back in June 2008. NVIDIA has since redefined their focus, allowing subsequent GF100, GF104, GF106, and now GF108 GPUs to prove their dedication towards next generation gaming effects such as raytracing, order-independent transparency, and fluid simulations.

While processor cores have grown from 128 (G80) and 240 (GT200), they reach 512 in the GF100 and earn the title of NVIDIA CUDA (Compute Unified Device Architecture) cores. GF100 was not another incremental GPU step-up like we had going from G80 to GT200. GF100 featured 512 CUDA cores, while GF104 was capable of 336 cores and GF106 had 192. Effectively cutting the four SMUs on GF106 in half, NVIDIA's GF108 is good for 96 CUDA cores from just two SMUs. The key here is not only the name, but that the name now implies an emphasis on something more than just graphics. Each Fermi CUDA processor core has a fully pipelined integer arithmetic logic unit (ALU) and floating point unit (FPU). GF108 implements the IEEE 754-2008 floating-point standard, providing the fused multiply-add (FMA) instruction for both single and double precision arithmetic. FMA improves over a multiply-add (MAD) instruction by doing the multiplication and addition with a single final rounding step, with no loss of precision in the addition. FMA minimizes rendering errors in closely overlapping triangles.

GF108 implements 96 CUDA cores, organized as 2 SMs of 48 cores each. Each SM is a highly parallel multiprocessor supporting up to 32 warps at any given time (four Dispatch Units per SM deliver two dispatched instructions per warp for four total instructions per clock per SM). Each CUDA core is a unified processor core that executes vertex, pixel, geometry, and compute kernels. A unified L2 cache architecture (512KB on 1GB cards) services load, store, and texture operations. GF108 is designed to offer a total of 4 ROP units pixel blending, antialiasing, and atomic memory operations. The ROP units are organized in two groups of two. Each group is serviced by a 64-bit memory controller. The memory controller, L2 cache, and ROP group are closely coupled-scaling one unit automatically scales the others.

Asus_ENGT430_Diagram.png

NVIDIA Fermi GF108 Block Diagram

Based on Fermi's third-generation Streaming Multiprocessor (SM) architecture, GF108 could be considered a divided GF106. NVIDIA GeForce GF100-series Fermi GPUs are based on a scalable array of Graphics Processing Clusters (GPCs), Streaming Multiprocessors (SMs), and memory controllers. NVIDIA's GF100 GPU implemented four GPCs, sixteen SMs, and six memory controllers. GF104 implements two GPCs, eight SMs, and four memory controllers. Conversely, GF108 houses one GPC, two SMs, and two memory controllers. Where each SM contained 32 CUDA cores in the GF100, NVIDIA configured GF104 with 48 cores per SM... which has been repeated for GF106 and the GF108. As expected, NVIDIA Fermi-series products are launching with different configurations of GPCs, SMs, and memory controllers to address different price points.

CPU commands are read by the GPU via the Host Interface. The GigaThread Engine fetches the specified data from system memory and copies them to the frame buffer. GF108 implements two 64-bit GDDR3memory controllers (128-bit total) to facilitate high bandwidth access to the frame buffer. The GigaThread Engine then creates and dispatches thread blocks to various SMs. Individual SMs in turn schedules warps (groups of 48 threads) to CUDA cores and other execution units. The GigaThread Engine also redistributes work to the SMs when work expansion occurs in the graphics pipeline, such as after the tessellation and rasterization stages.

GF108 Specifications

  • 96 CUDA Cores
  • 16 Texture Units
  • 4 ROP Units
  • 128-bit GDDR3
  • DirectX-11 API Support

GeForce 400-Series Specifications

Graphics Card

GeForce GT 430

GeForce GTS 450

GeForce GTX 460

GeForce GTX 465

GeForce GTX 470

GeForce GTX 480

GPU Transistors

585 Million

1.17 Billion

1.95 Billion

3.2 Billion

3.2 Billion

3.2 Billion

Graphics Processing Clusters

1

1

2

4

4

4

Streaming Multiprocessors

2

4

7

11

14

15

CUDA Cores

96

192

336

352

448

480

Texture Units

16

32

56

44

56

60

ROP Units

4

16

768MB=24 / 1GB=32

32

40

48

Graphics Clock
(Fixed Function Units)

700 MHz

783 MHz

675 MHz

607 MHz

607 MHz

700 MHz

Processor Clock
(CUDA Cores)

1400 MHz

1566 MHz

1350 MHz

1215 MHz

1215 MHz

1401 MHz

Memory Clock
(Clock Rate/Data Rate)

900/1800 MHz

902/3608 MHz

900/3600 MHz

837/3348 MHz

837/3348 MHz

924/3696 MHz

Total Video Memory

1024MB GDDR3

1024MB GDDR5

768MB / 1024MB GDDR5

1024MB GDDR5

1280MB GDDR5

1536MB GDDR5

Memory Interface

128-Bit

128-Bit

768MB=192 / 1GB=256-Bit

256-Bit

320-Bit

384-Bit

Total Memory Bandwidth

28.8 GB/s

57.7 GB/s

86.4 / 115.2 GB/s

102.6 GB/s

133.9 GB/s

177.4 GB/s

Texture Filtering Rate
(Bilinear)

11.2 GigaTexels/s

25.1 GigaTexels/s

37.8 GigaTexels/s

26.7 GigaTexels/s

34.0 GigaTexels/s

42.0 GigaTexels/s

GPU Fabrication Process

40 nm

40 nm

40 nm

40 nm

40 nm

40 nm

Output Connections

1x Dual-Link DVI-I 1x HDMI 1x VGA

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

Dual-Slot

Power Input

None

6-Pin

2x 6-Pin

2x 6-Pin

2x 6-Pin

6-Pin + 8-Pin

Thermal Design Power (TDP)

49 Watts

106 Watts

768MB=150W / 1GB=160W

200 Watts

215 Watts

250 Watts

Recommended PSU

300 Watts

400 Watts

450 Watts

550 Watts

550 Watts

600 Watts

GPU Thermal Threshold

95°C

104°C

105°C

105°C

105°C

GeForce Fermi Chart Courtesy of Benchmark Reviews



 

Comments 

 
# huh?Dave 2010-10-11 14:02
"NVIDIA released the GTX480 video card, they reclaimed the top performance position"

How do you figure? The 5970 can't be touched in benchmark tests. Price aside, the 5970 is the beast to beat.
Report Comment
 
 
# RE: huh?Olin Coles 2010-10-11 14:08
The author is probably speaking figuratively, as in NVIDIA has reclaimed the title of producing the best GPU... which they did when GF100 launched. It wasn't by much, but it was certainly enough to see a difference.
Report Comment
 
 
# re: huh?warthorn 2010-10-11 14:11
GTX 480 has a more powerful GPU core and is the most powerful single-GPU on the market. Additionally, if talking about "ultimate" current performance, GTX 480 Quad-SLI usually beats HD 5970 CFX (4GPU vs 4GPU).
Report Comment
 
 
# RE: ASUS GeForce GT 430 Fermi GF108 Video CardXeann 2010-10-11 17:37
i apprecciate all the bennchmarks expose here. Definitly, gives me a clear example in where the GT 430 can do a good job. but I think it would be a fairer comparison with a GT 240, Radeon HD 5600 for example.

Anyway, thanks for this benchmark
Report Comment
 
 
# RE: ASUS GeForce GT 430 Fermi GF108 Video CardRealNeil 2010-10-11 18:02
5 minutes after I pulled the trigger on this card at newegg, I got the email about this review.
Good to see that it will work for what I want it to.
Report Comment
 
 
# GTX480 vs 5970Hank 2010-10-11 19:24
As was stated by the other posts, I was referring to the performance of the GPU, I should have been more specific.
Report Comment
 
 
# Image quality?Sweatshopking 2010-10-12 03:57
after reading the anandtech review of this card, I have a few concerns. One is the fact that a 5600 series costs less, and performs better. It will also be silent, which this card isnt. As well, according to anand, this card has inferior image quality. Knowing that a more affordable card performs better, and looks better, how can you recommend this card?
Report Comment
 
 
# WellBez 2010-10-21 01:31
Anand was testing the Nvidia reference card, so when it comes to noise at least, you can't compare with the ASUS card.
Report Comment
 
 
# RecommendationHank 2010-10-13 15:27
Because this card supports Physx and, more importantly, 3D media playback. I recommended it for a media environment. Show me a Radeon that can do that. I think the Radeon is much better if you want to play games, and I said so.
Report Comment
 
 
# Any???Sweatshopking 2010-10-14 05:17
Please see
##xbitlabs.com/news/video/display/20100324235120_ATI_Radeon_Finall y_Supports_Stereosco pic_3D_Output.html

any dx 11 card will support 3d on a 120hz monitor, since 10.3 drivers. yes, it's less functional, but it works.
Report Comment
 
 
# OkHank 2010-10-14 06:26
And PhysX?

So basically, what you are telling me is that, since you can get 3D with less functionality on any other card, and who cares about PhysX, I shouldn't recommend this card for HTPC use?

Hank
Report Comment
 
 
# well the second parts true...Sweatshopking 2010-10-14 06:36
I would say, that functional isn't the best word. It IS functional, and it's not proprietary, so you can use a variety of software. Check out bit cauldron for some sweet products. I'm not a fan boy either way. I like both companies, my concern is to recommend a card that underperforms a card 1 year old, delivers poor image quality, at a higher cost. Also, You're not going to be running any physx on this card, so it's moot.

That being said, it's your site, do as you want. If you think it's worth it, fine, give'r. But to me, it seems like a lackluster product. Slower, hotter, louder, poorer image quality, and requires proprietary 3d, and more expensive. Recommend away.
Report Comment
 
 
# Bottom LineHank 2010-10-14 06:40
The bottom line here is, I recommend this card for HTPC or media PC use. I clearly stated so in the article. That's my opinion. You obviously have a different opinion. You can go out an buy a proprietary HDTV, hope it works with the proprietary set of glasses and IR receiver you get somewhere else, and buy some proprietary software that might work with your TV and glasses, all so you can buck the recommendation of a GT430 card that plays 3D media.

One last thing, the ENGT430 that I tested was silent. It also doesn't require a PSU input as the 5670 does. Sounds good for a media PC to me.

Hank
Report Comment
 
 
# Whoa Brother,Sweatshopking 2010-10-14 06:50
First of all, The one you have pictured Is NOT silent, it clearly has a fan, however quite. it is not passively cooled.
Second, I'm not disputing that it might be ok for HTPC, just that it's not the best.
third, NVIDIA HAS PROPRIETARY GLASSES AND REQUIRES A KIT. IT'S NOT OPEN, and the drivers are not reviewable by other manufacturers. Amd does not, and is not. They use the VESA standard, and will work with ALMOST ALL glasses and tv's, vs nvidia, who only works with nvidia glasses, and requires the nvidia kit.
I'm not sure why you got so worked up, i realize this site has been getting a lot of flack recently over some of the nvidia reviews, and that might be part of the reason. I do appreciate the fact that you reply to my concerns. I realize as well, that this isn't my site, and come to enjoy it, at no cost to myself.

i would like to thank you for the effort to review, benchmark, and discuss this card and this article.

thanks

Josh
Report Comment
 
 
# No problems hereHank 2010-10-14 09:31
I'm sorry if I seemed worked up. I don't think that this card is better than a 5670, but I do think it would be better for a media PC. I could only find one 5670 that doesn't have a fan, I assumed you meant just silent to the reviewer's ears. You are right, it is not passively cooled.
The difference that I have found between the NVIDIA and ATI 3D is this: You will have to find a TV that is supported by the NVIDIA hardware, but it's just that. With ATI, you have to make sure the glasses are supported, you have to get 3rd party drivers for the glasses, and you have to make sure the TV is supported. The article that you linked specifies that as well. In fact, the link it had for DDD showing what TVs it works with is down and iZ3d only works with its own monitor. So it seems easier to me to buy the NVIDIA kit, then just worry about the TV. Anyway, like I said, you are entitled to your opinion and we certainly appreciate you commenting here. It's good to have contrasting viewpoints.
Report Comment
 
 
# It's all good!Sweatshopking 2010-10-14 09:49
Thanks, I appreciate the candid response.

I've heard that ATI's implementation was becoming significantly streamlined in compared to nvidia's. It's interesting to hear that might not be the case. Let me know how it's hanging Hank, i'll continue to follow your articles.
Report Comment
 
 
# AMD and 3DOlin Coles 2010-10-14 09:56
Just for reference, AMD/ATI does not have a dedicated 3D solution. There are 3rd party kits that work with Radeon video cards, but they work equally well with GeForce products, too. Alternatively, 3D Vision offers expanded special effects through coordinated development.
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