What's new
  • Please do not post any links until you have 3 posts as they will automatically be rejected to prevent SPAM. Many words are also blocked due to being used in SPAM Messages. Thanks!

HD 6870 & HD 6850 vs. GTX 460 1GB: An Overclocking Study

Status
Not open for further replies.

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
12,840
Location
Montreal
StarCraft II (DX9)

StarCraft II (DX9)


In order to test StarCraft II performance we recorded a typical multiplayer battle on the Agria Valley map and used it as a replay in order to ensure every run was identical to one another. We used the last 3 minutes of the replay which includes the final assault on the enemy base. MSAA was applied in the NVIDIA and ATI control panels for certain tests.


1680 x 1050

HD6800-OC-76.jpg


HD6800-OC-77.jpg


1920 x 1200

HD6800-OC-78.jpg


HD6800-OC-79.jpg


2560 x 1600

HD6800-OC-80.jpg


HD6800-OC-81.jpg
 

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
12,840
Location
Montreal
Unigine: Heaven v2.0 (DX11)

Unigine: Heaven v2.0 (DX11)


Unigine’s Heaven benchmark is currently the de-facto standard when it comes to simple, straightforward DX11 performance estimates. While it is considered a synthetic benchmark by many, it is important to remember that no less than four games based on this engine will be released within the next year or so. In this test we will be using a standard benchmark run with and without tessellation enabled at three resolutions,


1680 x 1050

HD6800-OC-82.jpg


HD6800-OC-83.jpg


1920 x 1200

HD6800-OC-84.jpg


HD6800-OC-85.jpg


2560 x 1600

HD6800-OC-86.jpg


HD6800-OC-87.jpg
 

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
12,840
Location
Montreal
8xMSAA Testing

8xMSAA Testing


In this section we take a number of games we have tested previously in this review and bring things to the next level by pushing the in-game MSAA up to 8x. All other methodologies remain the same.


BattleField: Bad Company 2 (DX11)

Note that 8x MSAA is enabled via the game’s config file for the NVIDIA cards since it is not a selectable option within the game menu

HD6800-OC-52.jpg


Dirt 2 (DX11)

HD6800-OC-59.jpg


Just Cause 2 (DX10)

HD6800-OC-66.jpg
 

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
12,840
Location
Montreal
Core Temperature & Acoustics / Power Consumption

Core Temperature & Acoustics


For all temperature testing, the cards were placed on an open test bench with a single 120mm 1200RPM fan placed ~8” away from the heatsink. The ambient temperature was kept at a constant 22°C (+/- 0.5°C). If the ambient temperatures rose above 23°C at any time throughout the test, all benchmarking was stopped. For this test we use the 3DMark Batch Size test at it highest triangle count with 4xAA and 16xAF enabled and looped it for one hour to determine the peak load temperature as measured by GPU-Z.

For Idle tests, we let the system idle at the Vista desktop for 15 minutes and recorded the peak temperature.


HD6800-OC-89.jpg

We’ll start with the obvious: overclocking increases heat output and overclocking with higher than reference voltage impacts thermals even more. Unfortunately, without a custom cooler the HD 6870 gains heat at an almost alarming pace and believe it or not, the vapour chamber-totting XFX HD 6850 isn’t that far behind. If you are thinking of going for high clock frequencies with a HD 6800 series product, we highly recommend you look into upgrading the heatsink or at least buying a product with a higher end thermal solution.

The GTX 460 1GB features an excellent heatsink that isn’t anywhere near its limits; even with additional voltage being pumped through its fins.


System Power Consumption


For this test we hooked up our power supply to a UPM power meter that will log the power consumption of the whole system twice every second. In order to stress the GPU as much as possible we once again use the Batch Render test in 3DMark06 and let it run for 30 minutes to determine the peak power consumption while letting the card sit at a stable Windows desktop for 30 minutes to determine the peak idle power consumption. We have also included several other tests as well.

Please note that after extensive testing, we have found that simply plugging in a power meter to a wall outlet or UPS will NOT give you accurate power consumption numbers due to slight changes in the input voltage. Thus we use a Tripp-Lite 1800W line conditioner between the 120V outlet and the power meter.

HD6800-OC-88.jpg

Yup, overclocking has a significant impact upon power consumption as well. We will let the numbers do the talking here but want to add one thing: if you buy a card for its efficiency, don’t expect power consumption to be anywhere near that of the reference design once you overclock it. GDDR5 in particular can draw impressive amounts of power when pushed to its limits.
 

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
12,840
Location
Montreal
Conclusions

Conclusions


Note, all overclocked cards in the charts below are based upon the highest attained overclock before voltage increase.


GTX 460 1GB

HD6800-OC-90.jpg

With the release of the HD 6800 series, many predicted hard times for NVIDIA’s GF104-based solutions but through steep price cuts and a large stable of pre-overclocked cards, board partners seem well placed to take on AMD’s onslaught.

What makes things interesting is the GTX 460’s overclocking capability allowing partners to release unique products that pack a ton if improvements over the reference designs. This is where the EVGA GTX 460 FTW can be written into the equation since it provides a performance increase over the reference card that is simply astounding. Even when a HD 6870 is overclocked to 1GHz, the battle between it and the FTW flip flops from one card to the other depending on the game – there simply is no clear-cut winner until you get to ultra high resolutions. With a price of just $229 USD (and even less when you factor in the $10 - $15 mail in rebates being offered) this EVGA card is well within the financial reach of many gamers. This also makes it simply untouchable from a price / performance viewpoint.

These high clock speeds do act like a double edged sword though. While a reference GTX 460 may overclock like no one’s business, the EVGA FTW has very little headroom for further increasing performance. This particular GTX 460 should also be considered a “best case scenario” since it is the highest clocked GTX 460 currently available in the North American market.

With all of this being said, we think GTX 460 1GB cards –be they overclocked or not- seem to have found a perfect niche between the HD 6850 and HD 6870’s price points. Just be aware these NVIDIA products were never meant for ultra high resolution gaming; that’s the realm of the newly reduced $260 GTX 470 and even more expensive GTX 480.


HD 6870 1GB

Our respect for what AMD accomplished with their Barts XT has been well documented but overclocking the HD 6870’s core will lead to a wall in relatively short order. On the other hand, the memory overclocks extremely well.

Without increasing the voltage, our HD 6870 cards all stopped shy of a 10% core clock increase which resulted in a good performance gain when it was paired with the overclocked GDDR5. Upping the voltage by a reasonable amount led us over the 1Ghz mark but any additional upwards adjustments led to rapidly increasing heat, power consumption and noise with next to no additional gameplay benefits.

The overall capabilities of an overclocked HD 6870 really are impressive but this is one card which is literally begging for a custom heatsink. We’ll likely revisit HD 6870 overclocking once AMD’s board partners start tweaking its design; that should really allow this card to shine.


HD 6850 1GB

HD6800-OC-91.jpg

Considering the HD 6850 uses a slightly cut down version of the Barts core, we were expecting great things and it didn’t disappoint. Even without a voltage bump, it was nearly able to match the HD 6870’s reference clock speeds. Additional tweaking led to clock speeds above the 1Ghz mark but the memory speeds stopped well short of the HD 6870’s achievements.

Performance-wise, our overclocked HD 6850 was able to hang with the HD 6870 in the vast majority of benchmarks but due to its lower shader count, there was sometimes a noticeable gap between the two cards. Upping clock speeds by applying additional voltage brought the HD 6850 in line with its more expensive brother but like all cards, this quickly leads to high heat and diminishing returns in overall framerate increases. Even XFX’s custom heatsink struggled to keep temperatures within check when the core was running in excess of 1Ghz.


Final Thoughts

Buying a new graphics card is usually considered a significant financial investment for many and overclocking is an excellent way to get additional value out of a purchase. Naturally, some architectures have the ability to significantly increase their performance while others are near their limit from day one.

If you are in the market for a sub-$250 graphics card and are comfortable with overclocking while being mindful of the pitfalls, then both the GTX 460 1GB and the HD 6850 should be considered front-runners. They both have some impressive headroom to spare and feature extremely competitive price points. The HD 6870 on the other hand is a purebred racehorse but has tightly constrained overclocking limits at its reference voltage.

For the time being, certain pre-overclocked GTX 460 1GB cards can give NVIDIA’s board partners a slight to significant edge over the HD 6870. You can also buy these NVIDIA cards right now while AMD's board partners aren't quite ready to release products sporting the clock speeds we saw in this article. However, AMD’s AIBs will soon have their own take on pre-overclocking so the market is surely in for some great competition between the $150 and $250 price points in the coming months.



 
Last edited:
Status
Not open for further replies.

Latest posts

Top