Graphics Chips And Test Configuration
by
Tino Kreiss
We used an Intel Core i7-920 CPU for our testing purposes, boosting its clock speed to 3.8 GHz in order to circumvent any potential CPU bottlenecks.
This time, our technical tables are more comprehensive to help us detail how our retail test cards differ from the clock rates for standard reference cards. You'll find three different generations of manufacturing technology represented here as well: 65, 55, and 40 nm all exert influence on maximum power consumption and operating temperatures for graphics chips. These tables also illustrate how overclocking on specialty models boosts fill rates and raw computing power for their graphics chips.
| Nvidia Graphics Cards | ||||||
|---|---|---|---|---|---|---|
| Manufacturer and Chipset | Code Name | Graphics RAM | GPU clock | Shader | RAM clock | SPs |
| EVGA GTX 295 Hydro Copper (GeForce GTX 295) | 2 x GT200b | 2 x 896 GDDR3 | 720 MHz | 4.0, 1,548 MHz | 2 x 1,080 MHz | 2 x 240 |
| GeForce GTX 295 | 2 x GT200b | 2 x 896 GDDR3 | 576 MHz | 4.0, 1,242 MHz | 2 x 999 MHz | 2 x 240 |
| Zotac GTX285 AMP Edition (GeForce GTX 285) | GT200b | 1,024 GDDR3 | 702 MHz | 4.0, 1,512 MHz | 2 x 1,296 MHz | 240 |
| MSI N285GTX SuperPipe OC (GeForce GTX 285) | GT200b | 1,024 GDDR3 | 680 MHz | 4.0, 1,476 MHz | 2 x 1,250 MHz | 240 |
| GeForce GTX 285 | GT200b | 1,024 GDDR3 | 648 MHz | 4.0, 1,476 MHz | 2 x 1,242 MHz | 240 |
| MSI N280GTX OC HydroGen (GeForce GTX 280) | GT200 | 1,024 GDDR3 | 700 MHz | 4.0, 1,400 MHz | 2 x 1,150 MHz | 240 |
| GeForce GTX 280 | GT200 | 1,024 GDDR3 | 602 MHz | 4.0, 1,296 MHz | 2 x 1,107 MHz | 240 |
| BFG GTX 275 (GeForce GTX 275) | GT200b | 896 GDDR3 | 648 MHz | 4.0, 1,440 MHz | 2 x 1,152 MHz | 240 |
| GeForce GTX 275 | GT200b | 896 GDDR3 | 633 MHz | 4.0, 1,404 MHz | 2 x 1,134 MHz | 240 |
| GeForce GTX 260 216SPs | GT200b | 896 GDDR3 | 576 MHz | 4.0, 1,242 MHz | 2 x 999 MHz | 216 |
| GeForce GTX 260 | GT200 | 896 GDDR3 | 576 MHz | 4.0, 1,242 MHz | 2 x 999 MHz | 192 |
| GeForce 9800 GTX+ | G92b | 512 MB GDDR3 | 738 MHz | 4.0, 1,836 MHz | 2 x 1,100 MHz | 128 |
| GeForce 9800 GTX | G92 | 512 MB GDDR3 | 675 MHz | 4.0, 1,688 MHz | 2 x 1,100 MHz | 128 |
| Manufacturer and Chipset | Memory bus | Fab technology | Transistors | Interface |
|---|---|---|---|---|
| EVGA GTX 295 Hydro Copper (GeForce GTX 295) | 2 x 448-bit | 55 nm | 2 x 1,400 million | PCIe 2.0 |
| GeForce GTX 295 | 2 x 448-bit | 55 nm | 2 x 1,400 million | PCIe 2.0 |
| Zotac GTX285 AMP Edition (GeForce GTX 285) | 512-bit | 55 nm | 1,400 million | PCIe 2.0 |
| MSI N285GTX SuperPipe OC (GeForce GTX 285) | 512-bit | 55 nm | 1,400 million | PCIe 2.0 |
| GeForce GTX 285 | 512-bit | 55 nm | 1,400 million | PCIe 2.0 |
| MSI N280GTX OC HydroGen (GeForce GTX 280) | 512-bit | 65 nm | 1,400 million | PCIe 2.0 |
| GeForce GTX 280 | 512-bit | 65 nm | 1,400 million | PCIe 2.0 |
| BFG GTX 275 (GeForce GTX 275) | 448-bit | 55 nm | 1,400 million | PCIe 2.0 |
| GeForce GTX 275 | 448-bit | 55 nm | 1,400 million | PCIe 2.0 |
| GeForce GTX 260 216SPs | 448-bit | 55 nm | 1,400 million | PCIe 2.0 |
| GeForce GTX 260 | 448-bit | 65 nm | 1,400 million | PCIe 2.0 |
| GeForce 9800 GTX+ | 256-bit | 55 nm | 754 million | PCIe 2.0 |
| GeForce 9800 GTX | 256-bit | 65 nm | 754 million | PCIe 2.0 |
| Manufacturer and Chipset | Pixel Fill Rate Gpixel/s | Texture Fill Rate Gtexel/s | Memory Bandwidth GB/s | Die Size mm² |
|---|---|---|---|---|
| EVGA GTX 295 Hydro Copper (GeForce GTX 295) | 2 x 20.2 | 2 x 46.1 | 2 x 121.0 | 2 x 470 |
| GeForce GTX 295 | 2 x 16.1 | 2 x 36.9 | 2 x 111.9 | 2 x 470 |
| Zotac GTX285 AMP Edition (GeForce GTX 285) | 22.5 | 51.4 | 165.9 | 470 |
| MSI N285GTX SuperPipe OC (GeForce GTX 285) | 21.8 | 49.8 | 160.0 | 470 |
| GeForce GTX 285 | 20.7 | 47.4 | 159.0 | 470 |
| MSI N280GTX OC HydroGen (GeForce GTX 280) | 22.4 | 51.2 | 147.2 | 576 |
| GeForce GTX 280 | 19.3 | 44.1 | 141.7 | 576 |
| BFG GTX 275 (GeForce GTX 275) | 18.1 | 41.5 | 129.0 | 470 |
| GeForce GTX 275 | 17.7 | 40.5 | 127.0 | 470 |
| GeForce GTX 260 216SPs | 16.1 | 36.9 | 111.9 | 470 |
| GeForce GTX 260 | 16.1 | 36.9 | 111.9 | 576 |
| GeForce 9800 GTX+ | 11.8 | 47.2 | 70.4 | 276 |
| GeForce 9800 GTX | 10.8 | 43.2 | 70.4 | 330 |
| ATI Graphics Cards | ||||||
|---|---|---|---|---|---|---|
| Manufacturer and Chipset | Codename | Graphics RAM | GPU Clock | Shader | Memory clock | SPs |
| Radeon HD 4890 | R790 | 1,024 MB GDDR5 | 850 MHz | 4.1 | 4 x 975 MHz | 800 |
| Radeon HD 4870 X2 | R700 (2 x RV770) | 2 x 1,024 MB GDDR5 | 750 MHz | 4.1 | 4 x 900 MHz | 2 x 800 |
| Radeon HD 4870 | RV770 | 512 MB GDDR5 | 750 MHz | 4.1 | 4 x 900 MHz | 800 |
| Radeon HD 4850 | RV770 | 512 MB GDDR3 | 625 MHz | 4.1 | 2 x 993 MHz | 800 |
| Radeon HD 4770 | RV740 | 512 MB GDDR5 | 750 MHz | 4.1 | 4 x 800 MHz | 640 |
17
Comments
Sponsored
Latest Graphics Cards News
- 20/05 – Qualcomm Hires Former AMD CTO Demers
- 18/05 – Nvidia Responds to GeForce 600 Series V-Sync Stuttering Issue
- 18/05 – Behold: Here Are The First Unreal Engine 4 Screens
- 17/05 – Nvidia Debuts GK110-based 7.1 Billion Transistor Super GPU
- 17/05 – GE Announces Kepler Graphics Card for Military and Aviation
Latest Graphics Cards reviews
- 18/05 – Best Gaming CPUs For The Money: May 2012
- 10/05 – Best Graphics Cards For The Money: May 2012
- 10/05 – GeForce GTX 670 2 GB Review: Is It Already Time To Forget GTX...
- 09/05 – Video Teaser: Radeon HD 7800 Series Remains A Solid Value
- 08/05 – Benchmarking AMD's 768-Shader Pitcairn: Not For Public...
Looks like another Nvidia fanboi article.
The claim that ATI stock is hard to get is just rubbish.
Try getting any stock of the paperlaunched latest Nvidia cards eh?
TSMC isn't able to produce any in sufficient numbers for Nvidia so the cards are just smoke and mirrors.
How much did the NV marketing department pay you to write this rubbish?
Nothing, Tom's is independent in reviews.
Im sure Tom`s Hardware is giving us an independent review, But I can see why some people might think theres a slight swing, Some important facts wasnt stressed enough.
Being in the market for a new graphics card im getting alot of conflicting reviews from different people, But I still hold Tom`s reviews quiet high.
reynod get real pal, your obviously getting defensive because your an ATI 'fanboi' as you put it. if you actually read the article it says the 4770 is a great card but is hard to get hold of, which is the truth. and regardless of that fact the 4770 is more of a mid range card which wouldnt have a living chance against the 295 or the 4870x2 so whats the point about giving it a full glowing refernece. the simple fact is that the 295 is the best card available, just see the results for yourself.
ohh and another pointer this article is in relation to the best of the best which nvidia currently holds the title for. now if we were talking about the best for value then it would be a differnt story
It's hard to call it the best of the best when there are none to be bought becasue the 295 and 280 are a halo part ... and not available in any volume.
That's because they can't actually make them ... its a sham.
I call that simply wrong.
Oh ... and I have Nvidia cards mainly.
The average results are skewed in favour of Nvidia because of the inclusion of the results from "The Last Remnant" game which is quite apparently not working on ATi cards....this is obvious to anyone so why did you do that unless you were deliberately trying to favour Nvidia?
Also never seen anyone use "The Last Remnant" as a bench mark before...you must have looked long and hard to find one that was so unfair.
I would accept the review if Chris did it.
I am sure it would have been more balanced.
There are a number of overclocked X2 ATI cards out there too ... funny bout that.
This is just bad analysis, plain and simple. It does no service to the author, to TG and to the readers. Come on folks, this is elementary logic 101:
1) EITHER there is a problem with the Last Remnant test, OR we must accept the GTX295 is actually 636% faster than the 4870x2 (303 aggregate frames versus 41), and that the 4850 is 17% faster than the 4870x2 (48 frames versus 41).
2) In spite of this rather obvious issue the author simply included the test and made no comment about it. The appropriate action is clearly to either exclude the test, or to include it with an explanation of the discrepancy. The author did neither.
The error was compounded because he used the flawed data to draw (similarly flawed) conclusions. For example he claims that the EVGA295 is "29.8% faster" than the 4870x2 based on the aggregate framerate data. In total the EVGA295 polls 620 more frames than the x2, but 262 of the difference (or 42% of the outperformance!!) is attributable to this single flawed test!
3) I have to conclude that EITHER he did not notice the anomalous result (in which case he is incompetant) OR he did notice it and ignored it (in which case he is being misleading).
Which is it to be?
I would like to say this is a very flawed but unbiased article, but once again Toms is about as impartial as Hitler. ONE ATI card. And the benchmarks have been intentionally slanted. Again. But this time with all the subtelty of North Korea on a nuking spree. Where the hell are the non-reference HD4850X2s? At least those are actually available on teh interwebs! Not a single HD4890"XT" ("true" OC, the ATI-verified higher-binned parts) - they're not even acknowledged on the chart! Those are close to a GTX285 at stock and have a much higher OC ceiling. Where's the Sapphire Atomic HD4870X2? That's not just liquid-cooled, you can even throw the CPU into the pre-built WC loop! And that unit is actually available, and it has significantly higher stock clocks and OC headroom!
You didn't even try to hide it this time Toms. Yes, nVidia has the fastest cards. No, in the real world this doesn't really mean much - exactly where do you go to buy these things, and what bank do you have to knock off to afford them?! 99% of people like window-shopping these "halo" parts but they won't influence their actual purchases - the value-for-money in their price bracket is what affects them, and for most people it ain't gonna be a $800+ halo part!
And no 2560x1600 for high end cards?! This is pretty ridiculous.
Wow, this Roundup is not balanced!
I would say it is from an Nvidia Fanboy, the Frames are not the same i experience with my 4870 x2 and i got the one from MSI.
this is my last visit to Toms... time to find another source of components reviews that is REAL. I thought CPU mag was bad lol.
I agree with the percentage of people here, I am looing for a GFX card and this by no means gives me a good idea of what cards are available and comparative results other than the Nvidia market. Id have liked to have seen a broader range of ATI cards such as the 4850 x2 and more OC cards.
Um, I think you lot have missed the point of this article.
This article is not about comparing numerous cards to determine the best. This article is about taking the established top-tier cards and looking at their crazy, crazy cooling and finding the real gem.
The reason there's only 1 ATI card in this list is simply because ATI only has 1 card that can compete with NVIDIA's simply better lineup. This is about the best of the best, and ATI only have 1 entry in that category.
Plain, simple, stop moaning when you've missed the point.
LePhuronn - well said and totally agree with you. I cannot believe how many people feel 'threatened' because their favoured brand/make of graphics hardware perhaps didn't appear too favourably in their limited shortsightedness. It's not about 'which is best', or 'value for money'. As LePhuronn states the article is about taking a look at top-tier cards and their specific cooling systems... not a buyers guide!