Diamond Radeon HD 4870 Review

Author: Kevin Spiess
Editor: Howard Ha
Publish Date: Friday, July 4th, 2008
Originally Published on Neoseeker (http://www.neoseeker.com)
Article Link: http://www.neoseeker.com/Articles/Hardware/Reviews/diamondhd4870/
Copyright Neo Era Media, Inc. - please do not redistribute or use for commercial purposes.

Last week saw the introduction of the HD 4800 series from AMD/ATI, and the reverberations from that release are still resounding across the web. Reviewing the HD 4850 surprised everybody out there, including us  -- everyone had heard it was going to be fast, and offer great value, but the extent of both these claims weren't really properly quantified until we ran through the benchmarks, and saw that it kept up (and passed) even the fastest of video cards in our current testing-lab arsenal.

Today we are going to follow up that review with a look at the HD 4850's big brother: the HD 4870. The Diamond Radeon HD 4870 is certainly going to be fast -- but just how fast, we are eager to find out. Retailing for $299, this reference board design Radeon should be propelled to great speeds, thanks in large part to 512MB's of zipping fast GDDR5 memory. Judging from the HD 4850's great performance, we are guessing the only cards that should be able to keep up are the much more expensive 9800GX2, and the also recently released king-daddy next-gen card from NVIDIA, the monstrous GTX 280.

The HD 4850's release was definitely an important milestone in the history of ATI. Foregoing the strategy of starting a new generation off with a big, bad and loud expensive bang, the HD 4850 was aimed right at the jugular of the market place: most gamers would consider buying a $199 card, but not as many would want to buy something over $400. And for $199, the HD 4850 delivered -- as much as the HD 4870 will today? We'll find out. 

But besides gaming prowess, the HD 4850 and HD 4870 cards also feature a rich slew of features, including support for DirectX10.1 and Shader Model 4.1, the UVD2 high-definition video playback engine, improved power usage, and readiness to be connected with up to three other 3xxx series (or better) ATI cards in a CrossFireX setup.

How fast is fast? Let's find out today in our look at the Diamond HD 4870.

Note: Last week, for the launch of the HD 4800 series, we took a more in-depth look at the changes introduced in this new generation from ATI. If you'd like to read about these changes, go here.


The Diamond Radeon HD 4870 is a big card, with a double-slot cooler, but not as big to the same extent that the recent high-end cards from NVIDIA have been. The HD 4870 is about a half an inch longer than the HD 3870, and a bit heavier in comparison as well. The HD 4870 has very similar dimensions to the HD 2900 XT...actually -- wait a second, it seems the HD 4870 has a bit more in common with the HD 4870 than size alone -- as far as I can tell, they share the exact same cooler, although this cooler is encased in a different plastic shell.

Whether or not is a matter of extreme foresight, or merely saving some money by just sticking with a design that worked, the coolers for the HD 2900 XT and the HD 4870 are basically the same: a copper base sits on the GPU on the center of the card and extends to the bracket. Three medium-sized heat pipes help move the 'GPU's exhaust' out to a six inch bank of aluminum heat fins. A high-speed, many-vaned fan sits on the right side of the card, pushing a stream of air through this bank of fins...and viola: your GPU is cool.

This cooling solution seems sufficient for the regular operation of the HD 4870 in our testing. While perhaps the HD 4850's cooling could have been beefed up just a little bit (the 4850 ran fairly hot), this HD 2900 XT / HD 4870 is effective enough to keep the HD 4870's GPU within reasonable temperature ranges. The fan is also, for the most part, not that loud at all -- when it is manually set to full power, or when the card is powering up, at full force this fan can move a huge amount of air and be very loud in the process; but in standard operation, playing games, at least in our experiences, the fan never needed to spin fast enough make a unreasonable amount of noise.

Three really powerful cards: the XFX 9800GX2, the Diamond HD 4870, and the GTX 280

One thing that the HD 2900 XT certainly doesn't have in common with the HD 4870 is power needs: the 4800 series, as we saw in our review of the HD 4850 last week, put out a lot of power while maintaining a reasonable power consumption rates. While the HD 2900 XT required both a 8-pin and 6-pin PCIe power connection, the HD 4870 takes two 6-pin's. Diamond recommends only a 450W power supply (or greater) for this HD 4870 -- although, if you have a less expensive 500W or 450W PSU, you might be pushing it somewhat. Still, after shelling out $299 for a video card, it is nice to think that you might get by without having to buy a new PSU, if your still using a 500W PSU.


The most notable -- and one of the most impressive -- features of the Diamond HD 4870 is memory performance. The HD 4870 cards are the first ones to run with GDDR5. The vast majority of cards out use GDDR3 -- which has been the standard for sometime now. Even though ATI was first to pioneer using GDDR4, its adoption was fairly limited. But it seems likely that GDDR5 will probably be the de-facto standard for the high-end cards over the next fairly long (in tech terms) stretch of time. This memory is tied to a 256-bit memory interface, which is a bit more limited -- but much cheaper to make -- than the 512-bit and 448-bit memory bus found on NVIDIA's just released GTX 280 and GTX 260. However, as the benchmarks will soon show, seems like 256-bit memory bus offers enough bandwidth to keep the high-speed GDDR5 very happy indeed.

The 4800 series also introduced a number of improvements in AA performance. With the big match-up of the last generation, the HD 3870 versus the 8800 GT, the HD 3870 did lag behind somewhat the 8800 GT when anti-aliasing was used. The Diamond Radeon HD 4870 offers up to 24x custom filter anti-aliasing, and marketing materials suggest that this card will not break hardly a sweat at all at 2x and 4x AA levels.

It seems like in the older days, if you bought a mid-range card, you would play the newest games without AA to reach those higher framerates at first, while perhaps cranking the AA on the older titles in your gaming rotation. Perhaps this is not the case now -- but the benchmarks will help us figure this out.

  9800 GTX


8800 GTS 512MB

8800GT 512MB


HD 3850

GTX 260

GTX 280
HD4850 HD4870

Processing Cores








240 800 800

Core Clock








602 625 750

Shader Clock








1296 625 750

Memory Clock








1107 993 900

Memory Interface

256 bit

 256 bit

256 bit

256 bit

256 bit

256 bit

448 bit

512 bit   
256 bit 256 bit

Memory Type








1024MB GDDR3 512MB GDDR3 512MB

Fabrication Process








65nm 55nm 55nm


Box and bundle

Not deviating from the Diamond usual, this box is all black, but this time, it features an eye-catching, silver colored Ruby standing on the right. The color contrasts of the box probably help it stand out from some of the often garish boxes you find other cards in. 

On the back of the box is an image from S.T.A.L.K.E.R: Clear Sky, and some standard marketing exposition on the qualities of the HD 4870.

This video card comes with the following: a printed quick-start guide, a driver/manual CD, a credit for 50 free song downloads from emusic.com, a Crossfire bridge, a VGA to DVI adapter, a DVI to HDMI adapter, and a S-video out cable with adapter. All in all, a well put together, complete bundle that should have everything that most people are looking for.


The HD 4870 seems to run a bit cooler than the temperatures we were hitting with the HD 4850 last week, so we were eager to see how far this card could be pushed.

However this early in the game, there are limited overclocking tools available, so we were 'stuck' using the Catalyst Control Centre. In the CCC, the top core clock you can select is 790 MHz -- only 40 MHz above the stock 750 MHz speed. From what I can tell, it looks like this is unnecessarily set a bit lower than you would expect -- perhaps ATI wanted to leave room for further expansion of the product line. Already partners, including Diamond, have announced Black Editions and overclocked editions that have a higher possible top speed for the core unclocked. However, there was much more leeway when it came to the memory.

After a series of crash and burns, it seems liked the GPU overclocking wasn't working at nearly as well as the memory overclocking was. In the end, we were very happy running without problems at 780 MHz for the core, and a very impressive 1090 MHz for the memory. Seeing these results really makes us look forward to see what some extensive cooling and better overclocking tools will be able to accomplish. With the GDDR5 already very fast at 900 MHz, running at 1090 induced some big, fat grins in the testing lab.


With the performance of this last generation of video cards reaching new heights, we felt it was time to upgrade our benchmarking rig. Compared to our last setup, this new testing bench offers a significant increase in processing power, allowing us to better stay clear of any bottlenecks:


To compare against the Diamond HD4870, we selected the following cards: NVIDIA's GTX 280, an Asus EN8800 GTX, a VisionTek HD3870, a XFX 9800GX2, a Palit 9600GT Sonic, an Asus EN3850X2, and a VisionTek HD4850.


Driver-wise, I used NVIDIA's 175.19 Forceware drivers with everything but the GTX 280 which used Forceware 177.35's, while for the ATI cards, we used Catalyst 8.6 drivers, except for the HD4870, which used an unreleased driver supplied by ATI ('sample 5').

On the software side of things, here is the list of programs used:

Bioshock: For this benchmark, all of the Detail settings were set to 'High'. All of the graphic option switches were set to 'On', with the exception of the following three settings: Vsync, Windowed mode, and Force Global Lighting. We used FRAPS to measure frame rate performance. The FRAPS run was 138 seconds, triggered from pulling the switch in the sub at game's beginning. The sub's dive involves many big models moving around, which should strain the GPU's and be a good measure of the game's engine.

Call of Juarez: We used the stand-alone Call of Juarez DX10 benchmarking program for these results. For our AA testing, we used a setting of 2x.

Crysis: These benchmarks were performed using 'fly-by' GPU test found within the single-player pre-release demo version of the game. All graphic settings were on High.' For AA, we used a setting of 4x. DX10 mode was used. The game has also been fully patched.

Enemy Territory: Quake Wars: We use this id FPS benchmark to test out higher resolutions. We used the highest possible detail settings. We tested the resolutions at 4x AA as well as at 8x AA. 16x AF was also used.  

3DMark06 and 3DMark Vantage: These popular synthetic benchmarking programs were used at a resolution of 1280x1024. Vantage was run in 'Performance' mode, and only the two GPU tests were used.

World In Conflict: We used the built-in benchmark of the demo version of this game. We ran the benchmark in DX9 rendering mode, with a 'High' level of quality. For the AA testing, we used a setting of 4x, and a setting of 16x for AF.

Media Error demo: A runner-up winner in the scene.org best demo category for 2007, this demo made by Fairlight , CNCD & Orange. If you have no idea what a demoscene demo is, you might want to go here for an explication, and it is recommended you follow the link above to see for yourself. Media Error in particular presents a serious work out for a video card. Although to my knowledge, no one in the history of hardware reviewing has used a demoscene demo for benchmarking purposes before now, demos such as these are naturally well-suited to the task. This demo was run at a 1600x1200 resolution, with no AA and with 4xAA,  and FRAPS was used to measure framerate averages.   

If you would like any further information about our benchmark settings, feel free to ask us in the forums.

The three fastest cards in today's benchmarking -- the Diamond HD 4870, the XFX 9800 GX2, and the GTX 280 -- all danced around the top three spots through the tests, with some cards fairing better on some tests than others.

As far as these synthetic benchmarks go, perhaps the score offered by Vantage is easiest way to visualize the raw horsepower offered from these cards -- although whether or not this same pecking order will be maintained in the games is another matter. 

Sure, Call of Juarez always runs a bit better on ATI cards as it does on NVIDIA cards -- but there's not getting around this simple fact: straight out of the gate, the HD 4870 puts in better performance than the GTX 280, which costs more than double the price.

Nothing more needs to be said!

Let's check out some more benchmarks.


Besides falling a bit behind with AA on at 1600x1200 (perhaps due to the 256-bit memory bus?), the HD 4870 puts in another stellar performance, keeping up and beating the GTX 280, which is so impressive that it hardly needs further explanation.

The raw horsepower of the GTX 280 and 9800 GX2 help propel these NVIDIA cards to the top here, but the HD 4870 sill puts in some great numbers, not even considering the price difference between these products.

Perhaps most impressive here is how far the AA performance has come between the HD 3870 and the HD 4870 -- looks like some engineers were doing their homework.

The HD 4870 virtually ties with the GTX 280 here without AA, and with 4xAA, the HD 4870 maintains a healthy lead in this GPU-intensive benchmark.

Well -- can't win them all!

For our Unreal Tournament 3 benchmark, we use some very high resolutions. Perhaps if you have an excellent and capable display, the GTX 280 might offer better performance at the most extreme of resolutions. Then again, for whatever reason, the dual-GPU HD3850x2 does not do so poorly here -- so perhaps some driver work will bring the HD 4870's scores up here.

The Diamond HD4870 out-muscles the 9800GX2 here, but can't keep up with the GTX 280. However, when you consider that the HD 4870 is half the price of the GTX 280 currently, there are very few people that would complain about its level of performance.

The GTX 280 and HD 4870 virtually tie here, once again. For whatever reason, Bioshock seems to really benefit from the dual-GPU nature of the 9800GX2, as it races well ahead of the competition.

Power Consumption

To measure power usage, we used a Kill A Watt P4400 power meter. Note that the above numbers represent the power drain for the entire benchmarking system, not just the video cards themselves. For the 'idle' readings we measured the power drain from the desktop, with no applications running; for the 'load' situation, we ran a demanding part of 3DMark06.

These numbers are a bit surprising: while the HD4850 seems quite efficient, it seems like the HD 4870 is relatively less so. That idle is particularly high for a GPU that is about half the size of the GTX 280's GPU. This card seems to like the juice! Not to an alarmingly extent; but a bit more than expected. Still less than the GTX 280 and 9800 GX2 monster cards however.


Usually, when each team releases a new generation of graphics cards, fans from either side can argue that one or the other is the better bet to go with. This time around though, this might be a much harder proposition. 

While the GTX 280 is an engineering marvel, and offers substantial performance, in almost every respect the HD 4870 comes out ahead. Just judging from a value perspective alone, there is no contest: the HD 4870 by offers more for your money than does the GTX 280.

Perhaps the earlier switch to 55nm paid off -- this time around, ATI just delivered a thunderous one-two punch combination with the HD 4850 and HD 4870 cards. 

I fully expect that NVIDIA will adjust and react accordingly, and will offer high-end cards (very soon) at more competitive prices, but at these early stages of  the game, it will be very difficult to top the performance offered by the HD 4870, for around the $300 mark. The HD 4870 is cheaper to produce, and seemingly more efficient than the GTX 280 currently, and really, after looking at these benchmarks results, there seems to be very few reasons to recommend a GTX 280 over a HD 4870. ATI just bust down the door and stole the show this time around -- and it has been a long time indeed since you could say that.

Beside the actual 4800 series cards themselves, perhaps the best thing about this new generation's launch is that the level of competition between the red team and the green team just shot way up -- probably almost off the charts. Who knows... perhaps NVIDIA even decided to 'throw' this round because they have something explosive coming up in the pipes -- you never can tell. But for now, there really isn't much beating the value and gaming goodness offered by the HD 4870. The HD 4850 represents a better deal at $199, when it comes to down to the bang-for-buck, but if you are looking for something a little bit more, at the next step up  the HD 4870 stands pretty much alone. Sure, the GTX 280 can beat it sometimes -- but not as much as you'd expect for something that costs more than twice as much.

With the complete bundle, a fair price, and a helpful tech-support line offered by Diamond, if you are looking for a super-charged upgrade to your gaming rig, it would be hard to do much better this summer than picking up a Diamond Radeon HD 4870. 


Copyright Neo Era Media, Inc., 1999-2014.
All Rights Reserved.

Please do not redistribute or use this article in whole, or in part, for commercial purposes.