You are here

AMD Radeon HD 6870 Graphics Card

AMD Radeon HD 6870 Graphics Card

November 7, 2010 by Lawrence Lee

Product
AMD Radeon HD 6870

PCI-E Graphics Card
Manufacturer
Street Price
~US$240

Cloudy skies, bare trees, scarf-clad pedestrians, and slight dustings of frost on front lawns in the early morning can only mean one of two things: the onset of winter or the release of a new line of ATI graphics cards. Alas this year, there will be no new ATI card eating up your hard earned cash or miraculously appearing in a well-wrapped box underneath the Christmas tree. ATI is dead, done, no more. But relax, the new Radeons you've been waiting for are here and ready to keep you away from your loved ones during those long cold nights, they just now have a slightly different name. Noobs will continue to be fragged, characters will level up, and neglected spouses will hear the familiar placation "one more turn" (translation: see you in the morning).

This year ATI becomes AMD, a symbolic transition perhaps, but an end of an era nevertheless, and the start of a new one celebrated with a major press event for the 6800 series graphics card launch last month, as SPCR Editor Mike Chin documented. With the eagerly anticipated APU (accelerated processing unit, the fusion of the GPU and CPU on a single chip) on the horizon, ATI no longer simply compliments AMD with chipsets and integrated/discrete graphics. They are really one and the same now; it only makes sense for the company to have a single brand. However, they will have to live with the short-term pain of confused lay folk unsure about the compatibility of AMD graphics cards with their Intel-based systems.



The AMD Radeon HD 6870.

The Barts core of the HD 6000 series doesn't boast any revolutionary performance leaps over the HD 5000 line's Cypress core, at least not with the first wave of cards. Shortly after AMD's HD 5000 series supply fiasco, Nvidia came back from their long desktop graphics hiatus with GPUs with the new GF100 core. Of the cards nVidia released, the one that really hit it out of the park was the GTX 460, a US$200 card that easily offered the best value in that price range. With street prices of about US$180 and US$250 respectively, the new HD 6850 and 6870 cards are AMD's weapons to reclaim the midrange market.

\

The tale of the tape: HD 5850 vs. HD 6870.

Despite what the name suggests, the HD 6870 is not meant to be an upgrade for the HD 5870, but rather the HD 5850, with both cards currently retailing close to US$240. While we don't approve of misleading naming schemes, Nvidia has done much worse in the past, rebranding old GPUs with new, higher model numbers several times. Without getting technical, the main difference between Cypress and Barts is that the architecture has been rearranged to make it more efficient. So while the hardware specifications seem to be in the HD 5850's favor, under the hood the HD 6870 is doing more work with what it's got. Barts has a smaller die size, making it more energy efficient, and cheaper to produce, which means AMD can deliver better performance without sacrificing price or power consumption... or margins.



6870 back panel: 2 x Mini DisplayPort, 2 x DVI, 1 x HDMI.

Performance aside, there are a few things Barts can do that Cypress cannot. It has an updated version of UVD (universal video decoder) that supports MVC encoding, allowing the dual video streams required for stereoscopic 3D output. This coupled with an HDMI 1.4a connector gives the HD 6850/6870 the ability to play both games and Blu-ray movies in 3D. Almost as an afterthought, they've thrown in hardware acceleration for the DIVX/XVID codecs which is odd given it has been around for about 10 years and can be decoded via software without issue even on relatively slow machines.

In addition, AMD's Eyefinity multi-display technology has become even stronger. The 6850 and 6870 ship with a pair of Mini DisplayPort 1.2 connectors that can drive up to three monitors each, giving users the option of running six independent displays with a single card. Soon, there will be monitors with DisplayPort input and output, meaning you can daisy-chain them together. Inn the meantime a separate hub is required as a go-between for the extra displays. Unfortunately like the previous generation of cards, you cannot run three monitors through HDMI/DVI alone; at least one must be connected via a DisplayPort, so most users will still need adapters.

Specifications: AMD Radeon HD 6870
Interface PCI-E 2.1 16x
Fabrication Process 40 nm
Transistor Count 1700 million
Die Size 255 mm2
Core Clock Speed 900 MHz
Memory Clock Speed 1050 MHz
Memory Type GDDR5
Memory Size 1024 MB
Memory Bandwidth 134.4 Gbps
SIMD Engines 14
Stream Processors 1120
Texture Units 56
Color ROP Units 32
Memory Bandwidth 134.4 Gbps
Idle TDP 19W
Max TDP 151W
Additional Features DirectX 11, DirectCompute 11, OpenCL, Eyefinity, Shader Model 5.0, Open GL 3.2, UVD 3, Mini DisplayPort 1.2, HDMI 1.4a

PHYSICAL DETAILS

The Radeon HD 6870 is a dual slot graphics card with a length of 24.8 cm (9.8"), making it slightly longer than the width of a standard ATX motherboard. Like the reference HD 5850/5870, the 6870's reference design includes a box style heatsink with a blower fan pushing air over the heatsink and out the back.



The PCB measures 24.1 cm (9.5") long but the heatsink extends past the edge by an additional 7 mm.



This is the first time we've seen side-mounted PCI-E power ports located at the side of a Radeon graphics card with a single GPU. The connectors can add about an inch to a card's effective length so moving them to the side gives it better case compatibility.



The cooler is secured with 13 screws on the trace side of the board, as well as two screws connecting it to the metal PCI bracket.



The reference cooler features a large copper base with three heatpipes surrounded by a black metal plate that provides cooling for the memory chips and voltage regulators. Taking apart the stock heatsink is a bit of a chore; the folks at Atomic Maximum Power Computing has a disassembly guide for those interested.



Unlike previous board designs, the VRMs (highlighted in red in the photo above) are located on the left side of the card. Four capacitors sit right next to them and may interfere with larger VRM heatsinks provided by third party GPU coolers.



The orientation of the mounting holes around the core is the same as previous generations. This coupled with the core being shifted toward the right to make way for the VRMs, means the 6870 is more compatible with older third party heatsinks than the 5850 and 5870.

TEST METHODOLOGY

Our test procedure is an in-system test, designed to:

1. Determine whether the cooler is adequate for use in a low-noise system.
By adequately cooled, we mean cooled well enough that no misbehavior
related to thermal overload is exhibited. Thermal misbehavior in a graphics
card can show up in a variety of ways, including:

  • Sudden system shutdown, reboot without warning, or loss of display signal
  • Jaggies and other visual artifacts on the screen.
  • Motion slowing and/or screen freezing.

Any of these misbehaviors are annoying at best and dangerous at worst —
dangerous to the health and lifespan of the graphics card, and sometimes to
the system OS.

2. Estimate the card's power consumption. This is a good indicator of how efficient
the card is and will have an effect on how hot the stock cooler becomes due
to power lost in the form of heat. The lower the better.

Test Platform



HD 6870 technical specifications according to GPU-Z.

Measurement and Analysis Tools

Estimating DC Power

The following power efficiency figures were obtained for the
Seasonic S12-600
used in our test system:

Seasonic S12-500 / 600 TEST RESULTS
DC Output (W)
65.3
89.7
148.7
198.5
249.5
300.2
400
AC Input (W)
87.0
115.0
183.1
242.1
305.0
370.2
500
Efficiency
75.1%
78.0%
81.2%
82.0%
81.8%
81.1%
80%

This data is enough to give us a very good estimate of DC demand in our
test system. We extrapolate the DC power output from the measured AC power
input based on this data. We won't go through the math; it's easy enough
to figure out for yourself if you really want to.

H.264/VC-1 Test Clips

H.264 and VC-1 are codecs commonly used in high definition movie videos
on the web (like Quicktime movie trailers and the like) and also in Blu-ray
discs. To play these clips, we use Cyberlink PowerDVD.



1080p | 24fps | ~10mbps
1080p H.264:
Rush Hour 3 Trailer 2c
is a 1080p clip encoded in H.264
inside an Apple Quicktime container.





1080p | 24fps | ~8mbps
WMV-HD:
Coral Reef Adventure Trailer
is encoded in VC-1 using
the WMV3 codec commonly recognized by the "WMV-HD" moniker.

x264/MKV Video Test Clip

MKV (Matroska) is a very popular online multimedia container
used for high definition content, usually using x264 (a free, open source
H.264 encoder) for video. The clip was taken from a full length movie;
the most demanding one minute portion was used. We use Media Player Classic
Home - Cinema to play it as its default settings allow it to use DXVA
(DirectX Video Acceleration) by default.



1080p | 24fps | ~14mbps

x264 1080p: Spaceship is a 1080p x264 clip encoded from
the Blu-ray version of an animated short film. It features a
hapless robot trying to repair a lamp on a spaceship.

Testing Procedures

Our first test involves recording the system power consumption using a Seasonic
Power Angel as well as CPU and GPU temperatures using SpeedFan and GPU-Z during
different states: Idle, under load with CPUBurn running to stress the processor,
and CPUBurn plus FurMark running to stress both the CPU and GPU simultaneously.
This last state is an extremely stressful, worst case scenario test which generates
more heat and higher power consumption than can be produced by a modern video
game. If it can survive this torture in our low airflow system, it should be
able to function nominally in the majority of PCs.

If the heatsink has a fan, the load state tests are repeated at various fan
speeds (if applicable) while the system case fan is left at its lowest setting
of 7V. If the card utilizes a passive cooler, the system fan is varied instead
to study the effect of system airflow on the heatsink's performance. System
noise measurements are made at each fan speed.

Our second test procedure is to run the system through a video test suite featuring
a variety of high definition clips. During playback, a CPU usage graph is created
by the Windows Task Manger for analysis to determine the average CPU usage.
High CPU usage is indicative of poor video decoding ability. If the video (and/or
audio) skips or freezes, we conclude the GPU (in conjunction with the processor)
is inadequate to decompress the clip properly. Power consumption during playback
of high definition video is also recorded.

TEST RESULTS

Our test system has been upgraded since our last video card review, now featuring an Athlon II X3 processor and a Scythe Ninja 3 heatsink. Our previous Pentium D based system seemed out of date and the CPU temperature sensor was never very accurate (it would go insane after being placed on load for an extended period of time). The 785G board we chose also gives us southbridge temperature readings. These changes give us a more realistic testing environment and allows us to more accurately monitor the effect a graphics card/cooler has on CPU/system temperatures.

BASELINE, with Integrated Graphics: First, here are the results of
our baseline results of the system with just its integrated graphics, without
a discrete video card. We'll also need the power consumption reading during
CPUBurn to estimate the actual power draw of discrete card later.

System Measurements
Measurement
Idle
CPUBurn
CPU Temp
33°C
51°C
SB Temp
33°C
41°C
13 dBA
System Power (AC)
52W
140W
System Power (DC, est.)
unknown
111W
Ambient temperature: 22°C

Ambient noise level: 11 dBA

AMD Radeon HD 6870 1GB:

System Measurements
Measurement
Idle
CPUBurn
CPUBurn + FurMark
CPU Temp
32°C
54°C
60°C
SB Temp
37°C
43°C
49°C
GPU Core Temp
44°C
46°C
90°C
GPU VRM Temp
40°C
45°C
81°C
GPU Fan Speed
1060 RPM
1210 RPM
2950 RPM
15 dBA
15~16 dBA
34 dBA
System Power (AC)
71W
161W
336W
System Power (DC, est.)
unknown
129W
274W
Ambient temperature: 23°C

Ambient noise level: 11 dBA

System noise level: 13 dBA

At idle, the reference HD 6870 is impressive. Our test system measured only 15 [email protected] (2 dB higher than the same machine running on integrated graphics) and sounded quite smooth with no hint of whine, though we could hear a slight rattle at close proximity with the side panel removed. Temperatures were fairly low all around, and the power draw at the wall was just 71W. With CPUBurn stressing all the CPU cores, the southbridge temperature was 2°C hotter than when it was running on integrated graphics, and GPU temperatures increased only slightly compared to idle.

When FurMark was added the equation, GPU temperatures skyrocketed and the GPU fan speed followed suit, eventually settling at just under 3000 RPM resulting in a measured noise level of 34 [email protected] This is quite loud by out standards. On the positive side, the character of the noise was generally broadband, sounding similar to white noise or static. The VRM temperature on the card measured 81°C which is very low compared to previous cards we've tested, and the GPU core stabilized at 90°C — acceptable, but a bit on the high side. The extra heat generated by the card also put more thermal stress on the rest of the system, heating up both the southbridge and CPU by an additional 6°C.

Noise & Cooling Comparison
Card
Idle
Load
SPL @1m
GPU Temp
SPL @1m
GPU Temp
ATI HD 4870
13 dBA
66°C
22 dBA
87°C
HIS HD 4890 Turbo
16 dBA
61°C
30~31 dBA
83°C
HIS HD 5870 Turbo
18 dBA
40°C
34 dBA
88°C
AMD HD 6870
15 dBA
45°C
34 dBA
90°C
Ambient temperature: 23°C

Ambient noise level: 11 dBA

System noise level: 13 dBA

The 6870's reference cooler was noticeably quieter compared to previous cards we've tested. On load however, the fan had to work like a dog, bringing the overall noise level well past 30 [email protected] whilst keeping the GPU temperature at a balmy 90°C. Like the HIS 4890 and 5870, the 6870 is a prime candidate for improved cooling with an aftermarket hetainsk/fan.

BIOS Fan Settings



BIOS fan control settings.



Fan speed/temperature curve.

BIOS level fan control keeps the fan at the minimum allowed speed (22%) until the GPU core temperature hits 55°C at which point the fan speed increases linearly topping out at 102°C. For more customized control, we found that MSI's Afterburner utility works well with the 6870, though we don't think running the fan any slower at load is advisable.

POWER

The power consumption of an add-on video card can be estimated by comparing
the total system power draw with and without the card installed. Our results
were derived thus:

1. Power consumption of the graphics card at idle - When CPUBurn is
run on a system, the video card is not stressed at all and stays idle.
This is true whether the video card is integrated or an add-on PCIe 16X device.
Hence, when the power consumption of the base system under CPUBurn is subtracted
from the power consumption of the same test with the graphics card installed,
we obtain the increase in idle power of the add-on card over the
integrated graphics chip.

2. Power consumption of the graphics card under load - The power draw
of the system is measured with the add-on video card, with CPUBurn and FurMark
running simultaneously. Then the power of the baseline system (with integrated
graphics) running just CPUBurn is subtracted. The difference is the load power
of the add-on card. Any load on the CPU from FurMark
should not skew the results, since the CPU was running at full load in both
systems.

Both results are scaled by the efficiency of the power supply (tested
here
) to obtain a final estimate of the DC power consumption.

Note: the actual power
of the add-on card cannot be derived using this method because the integrated graphics may draw
some power even when not in use. If we assume our old test system with GMA 950 graphics used 1~2W, we estimate that the integrated HD 4200 graphics chip in our new setup uses 2W more than that, based on tests conducted using a low-end reference card. However, the relative difference between the cards should be accurate.

Estimated Power Consumption (DC)
Card
Idle
Load
ATI HD 4870 1GB
55W
147W
HIS HD 4890 Turbo 1GB
57W
153W
HIS HD 5870 Turbo 1GB
22W
216W
AMD HD 6870 1GB
18W
163W

By our estimates, the HD 6870 uses 18W when idle, a touch better than the 19W claimed, and a savings of 4W compared to the factory-overclocked HIS HD 5870 Turbo. This may be because of the core/memory speeds, which according to GPU-Z, are only 100/80 MHz when idle. On load the 6870 uses 53W less than the HIS HD 5870 Turbo, a substantial improvement that should make cooling the 6870 a much easier task.

VIDEO PLAYBACK

Video Playback Comparison
Test State
AMD Radeon HD 6870
HIS Radeon HD 5870 Turbo
Avg.

CPU
Avg. Power*
Avg.

CPU
Avg. Power*
Rush Hour

(1080p H.264)
10%
+40W
12%
+46W
Coral Reef

(WMV-HD)
12%
+41W
11%
+46W
Spaceship

(1080p x264)
16%
+45W
15%
+51W
*compared to idle (AC)

Though UVD 3 adds a few features to the 6870 in terms of video playback, the H.264 and VC-1 acceleration remains the same as it was in UVD 2. As a result, the average CPU usage of our test system was similar during H.264/VC-1 playback whether we used the 6870 or the older 5870. The 6870 gave us a small power savings but this is likely due to its lower idle power.

MP3 SOUND RECORDINGS

These recordings were made with a high
resolution, lab quality, digital recording system
inside SPCR's
own 11 dBA ambient anechoic chamber
, then converted to LAME 128kbps
encoded MP3s. We've listened long and hard to ensure there is no audible degradation
from the original WAV files to these MP3s. They represent a quick snapshot of
what we heard during the review.

These recordings are intended to give you an idea of how the product sounds
in actual use — one meter is a reasonable typical distance between a computer
or computer component and your ear. The recording contains stretches of ambient
noise that you can use to judge the relative loudness of the subject. Be aware
that very quiet subjects may not be audible — if we couldn't hear it from
one meter, chances are we couldn't record it either!

The recording starts with 5~10 seconds of room ambiance, followed by 5~10 seconds
of the VGA test system without a video card installed, and then the actual product's
noise at various levels. For the most realistic results, set the volume
so that the starting ambient level is just barely audible, then don't change
the volume setting again.

FINAL THOUGHTS

Gaming: Please check out gaming-oriented reviews of the HD 6870 at sites
like like HardwareCanucks,
The Tech Report, and
AnandTech. The general
consensus is that the HD 6870 performs somewhere in-between the HD 5850 and HD 5870 and is a much more energy efficient alternative to Nvidia's GTX 470 It is optimal for gaming at resolutions of 1920x1080/1200 with extra image quality features enabled.

Power Consumption: By our estimates, the Radeon HD 6870 uses about 18W
when idle, a 4W improvement over the overclocked HIS HD 5870 we reviewed earlier this year, and an excellent result for a graphics card of its caliber. When stressed to the theoretical limit with FurMark, the card consumed approximately 163W, a significant 50W+ savings over the aforementioned 5870.

Cooling: The reference cooler attached to our HD 6870 sample was very quiet when idle, measuring 15 [email protected] in our test system, only a few dB higher than our anechoic chamber's noise floor. Like most high performance cards, it runs hot and loud on full load. If you only care about the noise level of the card when it's sitting idle, you can stick with the stock cooler. Otherwise it is a perfect candidate for an aftermarket heatsink like the GELID Icy Vision, now in its second revision with HD 6850 and 6870 compatibility. The GPU core has been shifted away from the back panel so some older coolers that don't fit on the HD 5850 will work, though they may not be able to handle the heat output.

The HD 6870 is an upgrade over the HD 5850, bringing to the table a nice performance bump but without using nearly as much power as the 5870. No, it isn't quite as fast the 5870, but it uses approximately 50W less on load, and is about US$100 cheaper. Solid performance, well priced, and energy efficient, we feel it is the best overall choice at the US$240~250 mark. The 6000 series also offers a few extra niceties like stereoscopic 3D output and the potential for six displays running off a single card with DisplayPort. DP has not been fully embraced by display manufacturers yet, but HDMI adoption did not happen overnight, either; it will come. For now, this means extra costs in adapters past on to end-users, but it is still preferable to what Nvidia requires — a pair of cards in SLI or a dual GPU card to drive just three displays. Usable and practical benefits, excellent performance for money, and improved energy efficiency spell out the new AMD Radeon 6870's formidable strengths.

Our thanks to AMD for the Radeon HD 6870 sample.

* * *

Articles of Related Interest

Arctic Cooling Accelero Xtreme Plus GPU Cooler


GELID Icy Vision Dual Fan VGA Cooler


HIS HD 5550 & 5570 "Silence" Graphics Cards

HIS HD Radeon 5870 iCooler V Turbo

Asus EAH5750 Formula Graphics Card

ATI Radeon HD 5450 & HD 5570 Graphics Cards

* * *

Discuss this article in the
SPCR forums.

Sections: 

Google

www SPCR