EVGA 8800GT 512MB DDR3 - "2007" system

[harrynowl]

So welcome to my review of the EVGA GeForce 8800GT.

 

At launch this card was a great price : performer and kind of steamrolled the market. It came in at $200-250 USD and put AMDs 2900XT to shame. 

 

 

We have a GF92 core, as used in the 8800GTS and GTX. It has 112 shaders and 16 ROPs. It's a 8800GTX without the price tag. A GDDR3 memory archtechture is used to provide insane bandwidth to the GPU. A Whopping 60.8GB/s is achieved with this memory and the massive bitwidth of the bus.

 

I overclocked the GPU to 702MHz on the Core, 950MHz on the GDDR3 memory. This means a 16.6% overclock on the core and a 5% increase on the RAM. Also increased is the shader core for 1500 to 1728 MHz. A 15.2% increase.

 

And now we move to the games. In 2007 at launch this thing was baller. I included an older title (Modern Warfare 2) to showcase how good it was.

 

I had to run resolutions of 1280x1024 as the 1080p monitor I'd usually use requires more than 512MB of RAM. It managed around 40 FPS on some games but the textures were buggy with heavy micro stuttering. So the results are completely worthless.

 

For a CPU I'm using a gimped FX8320, I disabled the 2nd core in each module to make it a 4c/4t CPU and clocked it at 3.3GHz. With AMDs amazing penryn IPC this should be reasonably accurate for a Q6600.

 

 

 

First up: Battlefield 3.

 

I run through the first mission in full.

 

I ran the game at these settings: 

 

And this is the FPS graph:

I achieved a min / max / average of 27, 70 and 39 FPS respectively.

 

Next up: World of Warcraft.

 

For this benchmark I go through the vale of sorrows, starting at moshu-gun palace and ending near the entrance to the dread wastes.

 

The settings:

 

And the results:

 

 

And the min / max / average: 23 / 61 / 33

 

And lastly a game from the cards era:

 

Call of Duty: Modern Warfare 2

 

For this one I run through the 1st spec op "the pit". The run averages around 35 seconds.

 

The settings:

 

And the results:

 

 

And the min / max / average: 117 / 285 / 169

 

So clearly for games that were kind of designed for these cards it's formidable for the price. And even in modern(ish) games the card does respectfully well.

 

So the hardware I've used:

FX8320 4c/4t - 3.3GHz == Q6600 3.3GHz

8GB DDR3 @ 1066mhz

EVGA 8800GT

 

Tried to make it close to what you'd actually have in the era (2007).

 

[helping]

only obvious mistake is assuming 8320 matches a Q6600 clock per clock. 

[nuclear465]

I had one of those cards, they were pretty awesome !! i have used it between 2008 and 2011 until i sold it and bought a GTX 550Ti.

[harrynowl]

only obvious mistake is assuming 8320 matches a Q6600 clock per clock. 

Well I just went of Cinebench performance. Sure not the best way of doing things but the score was very close. 11 hours sleep last 3 days is enough to warrant not benchmarking a processor 30 times

 

http://www.c4dcafe.com/ipb/Benchmarks/cinebenchold.html?sort_col=record_updated&sort_order=desc&per_page=20&st=100

[flibberdipper]

Cool.

[ericlee30]

Awe bringing back the good old days sweet

[TheRetiredSlave]

11 hours sleep last 3 days is enough to warrant not benchmarking a processor 30 times 

 

What, are you training to replace Slick?

 

jk.. Good work with the benchmarks! It's always interesting to see how older hardware performs in some newer games.

[harrynowl]

What, are you training to replace Slick?

 

jk.. Good work with the benchmarks! It's always interesting to see how older hardware performs in some newer games.

Well I might use it as a point of reference for people that say "You need to spend £500 upgrading a PC every 6 months to play games"!

[The_Strict_Nein]

only obvious mistake is assuming 8320 matches a Q6600 clock per clock. 

 

Based on the cinebench score, it does

 

My Q6700 (which is effectively a Q6600 with a higher base clock) managed 4.12 at 3.80GHz, and from what I can remember around 350 at 3.50GHz, so 3.45 for 3.30GHz sound like a reasonable approximation

 

EDIT:

 

Just did a run of R11.5 with my Q6700 at 3.20GHz. Scored 3.53. If anything, the FX scores worse than a Q6600

[Sauron]

Ehm sorry to be the one who tells you this, but cod isn't exactly known for its superior graphics and power hunger... it runs on a modified version of id tech 3, from 1999... to really showcase a 2007 card's performance you should try bioshock. That's where the real graphics were.

[harrynowl]

Ehm sorry to be the one who tells you this, but cod isn't exactly known for its superior graphics and power hunger... it runs on a modified version of id tech 3, from 1999... to really showcase a 2007 card's performance you should try bioshock. That's where the real graphics were.

I didn't say it was. But it was a relevant game for the era of this card

It's the only game I had around that age other than total war but that's very difficult to benchmark accurately.

[jmaster299]

I had this card under the rebranded 9800GT name. Was a good card and only replaced it because I accidently got water on it while it was running. My old system made Frankenstein's monster look pretty so I kept the side panel off as it improved temps. I normally made it a habit of keeping things like my water glass away from my system, but needless to see I messed up one time. Card still worked after that, but I would get random artifacting and freezing that wouldn't go away without a hard reset.