What happens when you take a Radeon HD 3870 graphics card and slap a gig of memory on it? That is the question of the hour, since our subject is Diamond’s new card with just such a configuration. The Radeon HD 3870 has established itself as a pretty decent option among mid-range graphics cards, although it’s squeezed by formidable competition from Nvidia. Could doubling up on video memory allow the 3870 to distinguish itself from the likes of the GeForce 9600 GT and GeForce 8800 GT? We decided to find out.
Diamond’s novel facet
Diamond’s new take on the Radeon HD 3870 doesn’t depart too radically from the established formula. The card comes with a default GPU clock of 830MHz, well above the 775MHz baseline created by AMD but not quite as high as some of the cards we tested in our mid-range roundup, which ranged as high as 850MHz. Similarly, the card’s 1GB of GDDR3 memory comes clocked at 870MHz, for an effective data rate of 1740MT/s, on a 256-bit bus. The 512MB versions of the 3870 generally come with GDDR4 memory and consequently feature clocks as high as 1.2GHz. Diamond’s, er, ace is its larger memory size, coupled with the fact that GGDR3 memory tends to have lower latencies and thus perform better than GDDR4, clock for clock. On the flip side, GGDR4 tends to consume less power.
Incidentally, Diamond’s official specs for this card claim an 825MHz GPU clock and 900MHz memory. The card itself tells us it’s clocked at 830/870MHz. If you’ll excuse us, we’re going with the card on this one.
Diamond has outfitted this puppy with a little bit larger cooler than the norm; the oversized blower causes its enclosure to protrude slightly above the top of the card itself, much like a GeForce 8800 Ultra. The cooler doesn’t look to be tall enough to create any sort of clearance problems in your average PC case, but those with small form factor systems may want to proceed with caution.
Beyond that, the 1GB card comes with the usual complement of output ports and accessories, including a single DVI-to-VGA adapter and a DVI-to-HDMI adapter. Analog video junkies will appreciate the assortment of composite, component, and S-Video output cables (while the rest of us will tolerate them, I suppose). Diamond eschews the temptation to bundle a game with the card, and given the usual value proposition involved there, we’re not inclined to complain.
Oh, here’s something interesting. A sticker that refers to the “XHD3870XTG3”. As you may know, AMD did away with the “XT” suffix on its video card names when it adopted its new “numbers only” strategy that begat the 3870 and 3850. This humble little sticker suggests an XT name may have been in the works at some point, before everything changed. Why does this matter? Eh, I dunno.
The bigger question here may be: How does having 1GB of memory onboard help matters? For that, we have our test results on the following pages, of course. If you’ve been around the block a time or two like me, you’ll probably know that video memory size can be a funny thing. Sometimes, having more memory onboard can lead to a small, marginal increase in performance. More commonly, though, running out of video memory is a bit like a dog hitting the end of its leash while running along at top speedperformance drops off precipitously. These cases typically involve the use of lots of video memory, either because of high memory use by the application (with lots of large textures and objects in play) or by high-resolution/high-quality display settings. If you hit that limit, you’ll know it, and a card with more video memory will be dramatically faster in such scenarios. If you don’t hit that limit, well, the additional video RAM may be no help at all.
You can see all of our test configs below, but I’d like to make note of a few things. First, the GeForce 9600 GT card that we tested was “overclocked in the box” a little more fully than most (the core is 700MHz, while most cards are 650-675MHz), so its performance is a little bit higher than is typical. Similarly, we tested the GeForce 8800 GT and Radeon HD 3870 512MB at their stock speeds, which are increasingly rare in this segment. Most shipping products have higher clocks these days.
Beyond that, we’re in pretty good shape. Our examples of the Radeon HD 3850 512MB and GeForce 8800 GTS 512MB are both clocked above the baseline frequencies by typical amounts, and most of the higher end cards tend to run close to their baseline clock speeds.
Our testing methods
As ever, we did our best to deliver clean benchmark numbers. Tests were run at least three times, and the results were averaged.
Our test systems were configured like so:
2 Extreme X6800 2.93GHz
Matrix Storage Manager 7.8
DDR2 SDRAM at 800MHz
to CAS delay (tRCD)
with RealTek 184.108.40.20697 drivers
Diamond Radeon HD
3850 512MB PCIe
with Catalyst 8.2 drivers
Radeon HD 3870 512MB PCIe
with Catalyst 8.2 drivers
Radeon HD 3870 X2 1GB PCIe
with Catalyst 8.2 drivers
Diamond Radeon HD 3870
with Catalyst 8.3 drivers
9600 GT 512MB PCIe
with ForceWare 174.12 drivers
8800 GT 512MB PCIe
with ForceWare 169.28 drivers
GeForce 8800 GTS 512MB PCIe
with ForceWare 169.28 drivers
8800 Ultra 768MB PCIe
with ForceWare 169.28 drivers
9800 GX2 1GB PCIe
with ForceWare 174.53 drivers
9800 GX2 1GB PCIe
with ForceWare 174.53 drivers
Caviar SE16 320GB SATA
Vista Ultimate x86 Edition
|KB936710, KB938194, KB938979,
DirectX November 2007 Update
Please note that we tested the single and dual-GPU Radeon configs with the Catalyst 8.2 drivers, simply because we didn’t have enough time to re-test everything with Cat 8.3. The one exception is Crysis, where we tested single- and dual-GPU Radeons with AMD’s 8.451-2-080123a drivers, which include many of the same application-specific tweaks that the final Catalyst 8.3 drivers do.
Thanks to Corsair for providing us with memory for our testing. Their quality, service, and support are easily superior to no-name DIMMs.
Most of our test systems were powered by PC Power & Cooling Silencer 750W power supply units. The Silencer 750W was a runaway Editor’s Choice winner in our epic 11-way power supply roundup, so it seemed like a fitting choice for our test rigs.
Unless otherwise specified, image quality settings for the graphics cards were left at the control panel defaults. Vertical refresh sync (vsync) was disabled for all tests.
We used the following versions of our test applications:
- Bioshock 1.1
- Call of Duty 4: Modern Warfare 1.4
- Crysis 1.1
- Enemy Territory: Quake Wars 1.4
- Half-Life 2 Episode Two
- Unreal Tournament 3 1.1
- 3DMark06 1.1.0
- FRAPS 2.9.4
The tests and methods we employ are generally publicly available and reproducible. If you have questions about our methods, hit our forums to talk with us about them.
Doing the math
|GeForce 9600 GT||10.4||20.8||10.4||57.6||208|
|Palit GeForce 9600 GT||11.2||22.4||11.2||64.0||224|
|GeForce 8800 GT||9.6||33.6||16.8||57.6||336|
|GeForce 8800 GTS||10.0||12.0||12.0||64.0||230|
|GeForce 8800 GTS 512||10.4||41.6||20.8||62.1||416|
GeForce 8800 GTX
|GeForce 8800 Ultra||14.7||19.6||19.6||103.7||384|
|GeForce 9800 GTX||10.8||43.2||21.6||70.4||432|
|GeForce 9800 GX2||19.2||76.8||38.4||128.0||768|
|Radeon HD 2900 XT||11.9||11.9||11.9||105.6||475|
|Radeon HD 3850||10.7||10.7||10.7||53.1||429|
|Radeon HD 3870||12.4||12.4||12.4||72.0||496|
|Diamond Radeon HD 3870 1GB||13.3||13.3||13.3||55.7||531|
|Radeon HD 3870 X2||26.4||26.4||26.4||115.2||1056|
The table above shows how Diamond’s HD 3870 1GB stacks up in theoretical terms. Its higher GPU core clock grants it more fill rate and shader power than the stock HD 3870, although its lower memory clock cuts bandwidth considerably. The 1GB card’s memory bandwidth is still comparable to that of its GeForce 9600 GT and 8800 GT competition, though.
The more interesting question here involves overall performance. Not to give too much away, but the 3870 has somewhat underachieved versus the 9600 GT and 8800 GT given its raw shader FLOPS capacity. Why is that? One possibility is that the RV670 GPU’s five-wide superscalar execution units don’t process data as efficiently as Nvidia’s scalar units. I’m not sold on that explanation, though. AMD has implemented all sorts of voodoo magic in its driver compiler, including serializing a pixel shader program for execution on a that fifth ALU while another executes in vector fashion on the other four ALUs. Also, the performance of the 9600 GT argues against shader power being a primary constraint in today’s games. The more likely explanations involve the RV670’s relatively weak texturing capacity and the fact that R6x0-series GPUseither by design or because of a rumored flaw in the ROP logiccannot perform the resolve step for multisampled antialiasing in their ROP hardware; they must use the shader core for this task.
Another possibility, I suppose, is that the RV670 doesn’t compress and manage memory as efficiently as the GeForces do. If so, Diamond’s 1GB card may be an answer.
3DMark lets us measure performance in some of our theoretical categories. In actuality, sheer pixel throughput tends to be limited by memory bandwidth, which is why Diamond’s 1GB card scores lower in single-texture fill rate than the 512MB GDDR4 version of the HD 3870. Multitextured fill rate hits no such limits; the 1GB card nearly reaches its theoretical peak capacity. However, that capacity is appreciably lower than the GeForce 9600 GT’s, let alone the 8800GT’s.
The 3870 1GB shows its shader power, mixing it up with the GeForce cards from test to test. One intriguing result: the stock Radeon HD 3870’s performance suffers in the simple vertex shader test, likely due to GDDR4’s higher access latencies. With its GDDR3 memory, Diamond’s 3870 1GB avoids that fate.
As ever, these results don’t track perfectly with performance in actual games, although they do give us some insight. For gaming performance, we have… actual games.
Call of Duty 4: Modern Warfare
We tested Call of Duty 4 by recording a custom demo of a multiplayer gaming session and playing it back using the game’s timedemo capability. Since these are high-end graphics configs we’re testing, we enabled 4X antialiasing and 16X anisotropic filtering and turned up the game’s texture and image quality settings to their limits.
We’ve chosen to test at 1680×1050, 1920×1200, and 2560×1600resolutions of roughly two, three, and four megapixelsto see how performance scales. I’ve also tested at 1280×1024 with some of the less expensive cards, since some of them struggled to deliver completely fluid rate rates at 1680×1050.
Well, the Radeon HD 3870 1GB is slightly faster than the stock 512MB GDDR4 version, but the margin of improvement is minor. More than likely, the improvement comes from its higher GPU clocks and lower memory latencies, not the presence of additional RAM. At 2560×1600, where memory is most likely to be at a premium, the 1GB is only a smidgen quicker.
Meanwhile, the GeForce 9600 GT outperforms both of them.
Enemy Territory: Quake Wars
We tested this game with 4X antialiasing and 16X anisotropic filtering enabled, along with “high” settings for all of the game’s quality options except “Shader level” which was set to “Ultra.” We left the diffuse, bump, and specular texture quality settings at their default levels, though. Shadows, soft particles, and smooth foliage were enabled. Again, we used a custom timedemo recorded for use in this review.
The story with Quake Wars is similar to CoD4. Even at the highest resolution, the 512MB Radeons don’t appear to be running out of video memory, and the 1GB card shows no benefit.
Half-Life 2: Episode Two
We used a custom-recorded timedemo for this game, as well. We tested Episode Two with the in-game image quality options cranked, with 4X AA and 16 anisotropic filtering. HDR lighting and motion blur were both enabled.
Once more, we see no real benefit from the 1GB memory size, and the Radeon 3870 1GB trails the GeForces.
I was a little dubious about the GPU benchmark Crytek supplies with Crysis after our experiences with it when testing three-way SLI. The scripted benchmark does a flyover that covers a lot of ground quickly and appears to stream in lots of data in a short period, possibly making it I/O boundso I decided to see what I could learn by testing with FRAPS instead. I chose to test in the “Recovery” level, early in the game, using our standard FRAPS testing procedure (five sessions of 60 seconds each). The area where I tested included some forest, a village, a roadside, and some watera good mix of the game’s usual environments.
Due to the fact that FRAPS testing is a time-intensive endeavor, I’ve tested the lower-end graphics cards at 1680×1050 and the higher end cards at 1920×1200, with Radeon HD 3870 1GB included in both groups.
Nothing here suggests memory size is an issue for the Radeons in Crysis, even with its high-quality settings, which produce some fairly low frame rates. The 1GB card is ever-so-slightly faster than the 512MB one, but not by enough to turn any heads. The difference is likely caused by the 1GB’s higher GPU clock.
Unreal Tournament 3
We tested UT3 by playing a deathmatch against some bots and recording frame rates during 60-second gameplay sessions using FRAPS. This method has the advantage of duplicating real gameplay, but it comes at the expense of precise repeatability. We believe five sample sessions are sufficient to get reasonably consistent and trustworthy results. In addition to average frame rates, we’ve included the low frames rates, because those tend to reflect the user experience in performance-critical situations. In order to diminish the effect of outliers, we’ve reported the median of the five low frame rates we encountered.
Because UT3 doesn’t natively support multisampled antialiasing, we tested without AA. Instead, we just cranked up the resolution to 2560×1600 and turned up the game’s quality sliders to the max. I also disabled the game’s frame rate cap before testing.
All this has happened before and will again.
Overtime: another shot at 512MB versus 1GB
Since none of the games in our regular GPU test suite showed any appreciable difference between a RV670 with 512MB and 1GB of memory, I decided to try one more game. I’d read somewhere that Bioshock might be a good candidate to chew up some video memory, and indeed, a DX10 game with high-detail textures seemed like a good candidate. For this eleventh-hour test, I updated to the latest Catalyst 8.4 drivers and compared the Diamond 1GB card to an Asus Radeon HD 3870 with an 850MHz GPU core and 512MB of GDDR4 memory running at 1125MHz. The Asus currently sells for as little as $169.99 at Newegg.
I used FRAPS to test Bioshock, using the graphics options shown in the second screenshot below.
I started testing at 1920×1200, since that’s a more common resolution, and the Asus card came out on top. So I cranked up the resolution to a full four megapixels, raising the memory footprint. Again, the Asus card came out on top. All I can say is that I tried, folks. We perhaps could have contrived a situation where the 512MB card ran out of video memory, possibly one involving 16X CFAA or something like that, but we didn’t come by it organically with our common settings, which are the ones we’d use to play most games.
We measured total system power consumption at the wall socket using an Extech power analyzer model 380803. The monitor was plugged into a separate outlet, so its power draw was not part of our measurement. The cards were plugged into a motherboard on an open test bench.
The idle measurements were taken at the Windows Vista desktop with the Aero theme enabled. The cards were tested under load running UT3 at 2560×1600 resolution, using the same settings we did for performance testing.
The 1GB Radeon draws slightly more power at idle than the 512MB GDDR4 version, but they’re dead even when running at gameand, characteristically for a Radeon HD 3870, quite power-efficient.
We measured noise levels on our test systems, sitting on an open test bench, using an Extech model 407727 digital sound level meter. The meter was mounted on a tripod approximately 12″ from the test system at a height even with the top of the video card. We used the OSHA-standard weighting and speed for these measurements.
You can think of these noise level measurements much like our system power consumption tests, because the entire systems’ noise levels were measured, including the stock Intel cooler we used to cool the CPU. Of course, noise levels will vary greatly in the real world along with the acoustic properties of the PC enclosure used, whether the enclosure provides adequate cooling to avoid a card’s highest fan speeds, placement of the enclosure in the room, and a whole range of other variables. These results should give a reasonably good picture of comparative fan noise, though.
Unfortunatelyor, rather, quite fortunatelyI wasn’t able to reliably measure noise levels for most of these systems at idle. Our test systems keep getting quieter with the addition of new power supply units and new motherboards with passive cooling and the like, as do the video cards themselves. I decided this time around that our test rigs at idle are too close to the sensitivity floor for our sound level meter, so I only measured noise levels under load.
My ears both agreed with this assessmentDiamond’s cooler is a little louder than the stock oneand disagreed, because the Diamond’s blower produces a relatively unobtrusive, lower-pitched hiss than some coolers. Despite the relatively high reading on the decibel meter, I didn’t mind the sound much, and the blower spins down so it’s nice and quiet at idle.
Per your requests, I’ve added GPU temperature readings to our results. I captured these using AMD’s Catalyst Control Center and Nvidia’s nTune Monitor, so we’re basically relying on the cards to report their temperatures properly. In the case of multi-GPU configs, I only got one number out of CCC. I used the highest of the numbers from the Nvidia monitoring app. These temperatures were recorded while running UT3 in a window.
Here’s why that larger Diamond cooler makes more noise than the stock one: it’s keeping GPU temperatures considerably lower. This seems to be a trend among newer cards; the GeForce 9800 GTX is similar. I suppose that’s good for GPU longevity, if nothing else.
Raising the Radeon HD 3870’s memory size from 512MB to 1GB doesn’t seem to improve performance measurably in any of the games we tested, even at a monster 2560×1600 resolution. As a result, the Radeon HD 3870 1GB delivers 3D gaming performance that’s usually lower than the GeForce 9600 GT’s.
This outcome isn’t really surprising to us. Radeon HD 3870 cards seem to get along fine with 512MB of RAM, even in extreme circumstances. We saw strong performance scaling from the CrossFire X multi-GPU scheme, which imposes some memory overhead, in Crysis with three and four Radeon HD 3870 GPUs, each with 512MB. In the same situation, multiple G92 GPUs with 512MB ran into an obvious performance wall, while their 768MB cousins did not. AMD clearly has its GPU memory management mojo working better than Nvidia’s, so going beyond 512MB isn’t yet necessary.
This reality shouldn’t be any sort of great calamity for Diamond’s 1GB wonder. After all, video card makers slap outsized memory amounts on their products all of the time. Memory is cheap, and it does add a bit of future-proofing. No harm, no foul.
But this is the part where you can cue the scary music. You see, Diamond is selling its Radeon HD 3870 1GB for $299.99 at multiple online vendors. The card’s selling for $289.99 at Newegg with a $30 mail-in rebate, if you like to exchange money and inconvenience for presumably better-than-even odds of receiving a check in 90 days. Any way you slice it, though, that’s expensiveon par with the GeForce 9800 GTX, a product that vastly outperforms the Radeon HD 3870. 300 bucks will also buy you a pair of GeForce 9600 GT cards; one would be faster than the 3870 1GB, and two would roughly double its performance in many cases. All of which adds up to one inescapable conclusion: Diamond’s Radeon HD 3870 1GB is a singularly poor value. I’m all for pushing the boundaries with new memory sizes when the time comes, but nothing we’ve seen from this product justifies its price.