Graphics Card Comparison Guide

JaniceZ writes "These days, there are so many graphics card models that it has become quite impossible to keep up with the different configurations. Therefore, we decided to compile this guide to provide an easy reference for those who are interested in comparing the specifications of the various desktop GPUs in the market as well as those already obsolescent or obsolete."

Me dumb

[Monte]

Let's say I didn't know anything about graphics cards. How does this help me?

We really wouldnt need this type of thing

[Crashmarik]

If the manufacturers didn't go out of their way to completely confuse the issue to the point where there are no definitive answers to the question.

Re:Article content is medicore at best

[feanor512]

The Rojakpot guide provides fillrate and memory bandwidth so you can directly compare cards without doing the math. These are the two most import factors in the performance of the cards. It also has quite a few more cards.

They missed FPS

[Mishra100]

I hear video cards being rated in FPS(frames per second) in certain video games all the time. If person A can get 100 more FPS out of Doom 3 using an ATI at the same cost as a Nvidia, usually they are going to go for that.

So my question is why didn't they include this in there? They have a lot of good data but I just wish that someone would run all the video tests on each card and check out the FPS data on certain popular games and produce them in a nice chart similar to this one.

When the mainstream magazines fail,

[TuxPaper]

the web picks up the slack. Back in the day when I bought computer magazines (at least 2 years ago), I've always wanted comprehensive charts of the latest graphic cards listed in magazines. Occassionally, if they were doing a graphics card special issue, there'd be maybe 7 cards compared.

However, this comparison guide is hardly a "easy reference". It's on the web, so give it some features. I want to sort, filter out columns, have side-by-sides comparisons, comments/ratings by users (or staff), etc.

I live in Japan now, where I can pick up a monthly computer magazine, and they have a section dedicated to charts on the latest CPU, Video, HDD, Motherboards, and Chipsets. The video chart, for comparison, has 14 different specs, all listed on one row, making it far easier to compare than this site.

The only advantage to the charts at the site in this story is that it will/does include old cards. But, as with other commenters in this thread, I say this story certainly feels like a cheap ad.

My solution

[YuriGherkin]

I recently came up with a good way to compare video card with a bang-for-buck type analysis.

* I went to the latest review of VGA cards at Tom's Hardware [tomshardware.com].

* I chose the top 12 video cards from ATI and nVIDIA

* I created a spreadsheet which calculated the relative rankings of each card across about 30 different tests for a range of games/benchmarks. i.e. the top scoring card in a category got 100% and the remaining 11 cards were expressed as a fraction of the top score.

* I averaged the rankings for the 30 categories

* I used a local hardware search tool to find the current "buy it today" best prices for each of those cards.

* I divided the average ranking by the price to get a bang/buck ratio that can help to compare the cards. i.e. so a card that averaged 90% but costs AUD$600 would have a lower final score than a card that came in at 50% but only cost AUD$200

Unfortunately, the spreadsheet is at work but the 6600GT was a clear winner in terms of bang-for-buck.

All these 12 cards were good, and most of them were the only ones remaining in the extreme tests like high-res DOOMIII with AA sort-of-tests. So, even if a card only came in at 50% average, it was still able to work with all the latest games at reasonable frame rates.

Re:Article content is medicore at best

[Cylix]

I can't tell if it's supposed to be humorous or arrogant.

Choices are bad! Oh noes... I have to read a bit before spending over a few hundred dollars.

To be honest, if you are not intending to spend more then a hundred dollars your choices aren't so great. Older cards tend to depreciate to their relative value as per their capabilities.

With newer cards there is of course the tax for being a newcomer and the cost of the performance.

When I play this game, I start by determining how much I want to spend and then purchase the better performer in that range. At the point I'm buying a graphics card where all the questions are usually answered. (A new card for me is in the 220-250 range)

Those buying bleeding edge are either going to have to make a blind decision or put in some research time. This is pretty much the way all transactions above impulse buy work.

It's not a curse, it's not a problem introduced in the graphics card market... it's just how things work.

It's kinda silly to just point and buy something. (Though I did just buy my last car that way.. which was silly... but the price was right)

Re:Thank God

[Gentlewhisper]

I will avoid the nVidia cards like the plague.

Google for "6600GT infinite loop" and you will see what I mean. I have a 6600GT OC from BFG sitting *right here* on my desk, and why isn't it in my computer? Well, it seems like nVidia is staffed with whores who can't even code a device driver properly!

An equivalent ATI card *just works*. Period.

Infinite loop indeed, what did they do, outsource their driver development to munkeys in India? Total waste of my money, never again.. nVidia.

Re:Thank God

[Anonymous Coward]

I have the same card sitting on my desk...oh wait, in my computer I mean. I have had no issue with the card and it works beautifully.

Plugged it in, installed the latest drivers and it "*just works*".

I had more heat related issues with the 9800 Pro, pawned that card off and bought the BFG 6600GT OC. Been happy ever since.

Re:YAG3DGCC!

[Jeff DeMaagd]

I agree. Most of the 3D benchmarks push the idea that the only buying considerations for buying are FPS and dollars. It would be nice if they considered the other stuff like you mentioned.

I personally would like to see a database like what storagereview.com does, and be able to cull out rediculous crap like graphics cards that take two slots. I'd also like to know what sound pitches the on-board fans use and how loud they are.

It seems like the people that buy the high end stuff buy without regard to practicality. It was bad enough that high end cards used two slot spaces but didn't need it for circuitry. Then there are rediculous modders that eliminated almost all useful PCI slots so they can use a CPU cooler on a GPU.

Re:Article content is medicore at best

[Seumas]

My gaming PC has a $600 videocard in it. I'm thinking of getting away from the annual process of building two new souped-up PCs to keep up with top of the line gaming and moving to console gaming once the new XBOX and PS3 come out.

My main reason is that I can get any decent strategy or RPG game that I want to play on the Mac. Everything else, I can get for the console. And rather than maintaining two expensive high-end gaming PCs for myself and my brother, we can buy one copy of a game and have one machine that'll last us years. Much better than dishing out $110 for two copies of Unreal or Counter Strike: Source on top of the videocards and everything else.

Mostly, I just want out of the rat race. I have a spare bedroom filled with a couple dozen monitors and linux, solaris and windows boxes and several laptops. And all I use is my Powerbook. I'm undecided as to whether I'll use just the laptop, the laptop and buy a desktop mac on x86 later or if I'll stick with my powerbook and have a windows desktop.

Really, the only reason I would keep the windows box now is the same reason I kept a windows box in the linux/solaris world. Because it runs the games. I think I can do without Counterstrike and a couple other PC games and settle for Halo and Madden and GT4 on the console. Or at least, I'm going to try.

A More Concise Table...

[ptcheezer]

I've always referred to this excellent pcvsconsole table of PC video card specs [pcvsconsole.com] for a very concise overview of the specs and performance of the various video cards. It really is worth a look if you want to see how much more horsepower is in the newer cards as compared to the older ones.

Re:Short list

[adam613]

That's not surprising. Most four-year-old hardware will work better on Linux than the brand-new equivalent, because various developers have had four years to write drivers.

What is impressive about nVidia is that their brand-new hardware works just as well under Linux as the four-year-old stuff.

Nvidia Linux support

[DrJimbo]

A couple years ago, I noticed a memory leak in OpenGL apps when using Nvidia drivers and an experimental Gentoo Linux kernel. I sent Nvidia an email about it around 10 pm Saturday evening.

I got a response about 20 minutes later which included a patch for the Linux kernel I was using. I recompiled my kernel with the patch and it fixed the leak.

It is too bad their drivers are closed source, but I have to say that their Linux support is outstanding and on a par with the best support I've experienced.

nVidia and compatibility

[King_TJ]

Yeah, I've always had good luck with nVidia drivers/cards and Linux - although not sure I'd say they're exactly "at the same level" as the Windows driver counterparts.

On my MythTV box, using a GeForce 4Ti 4600 card, I've run into lots of issues of nVidia changing around little details related to the card's ability to output in HDTV resolutions, to properly select or auto-detect which port the card is connected to (s-video, composite, or DVI/VGA) and other such things. It generally works well... don't get me wrong. But some of the optional parameters nVidia says you can specify in your XF86Config-4 file seem to shift around from version to version in their driver updates, and things that work fine in one release are broken again in the next. (Lately, I've had issues where the option to specify some "overscan" for composite or s-video output with an integer value between 0.0 and 1.0 seems to have no effect at all on my card.)

Re:Short list

[myslashdotusername]

If you want Linux compatibility, you want nVidia.

and if you want a graphic card with a Manufacturer's Warrenty you want to get an ATI. Seriously for certain card models ATI cards aren't that badly supported (even under linux) sure, most ati cards work better under windows than under linux (with a few exceptions) but if you're buying a gaming system and worrying about linux you're either A. dual booting or B. running some type of virtual machine set up, or C. planning on migrating down to 'linux' when you retire the system. and if you're building a 'workstation' you're going to be looking at the FireGL cards, and nvidia has nothing that compares to FireGL and the linux drivers for ATI products are Specifically written for the FireGL product lines..

so you be happy with your Nvidia card which has some dubious warrenty from some company who may or may not have a 'real' setup for providing warrenty service, and requires you to return the card to a retailer, who pays a $xx,xxx a year fee for the right to RMA defective products to the manufacturer. i think i'll stick with ATI, and if i need an ATI card to work under linux I'll be sure to grab up a FireGL card...

Re:Thank God

[Worminater]

did you bother turning agp fast write off?

Re:Nice charts, what happened to Matrox?

[Anonymous Coward]

Well now Matrox doesn't really do consumer market goods anymore, too much competition I guess they couldn't afford spending millions on research and developpement for a product with a 6 months or less lifespan. Now they do things like hi-end cards for busisiness purpose (medical imaging, surveillance systems, and this sort of thing) there are still in pretty good shape they just went for a more specialized market.

Re:But what about Linux drivers?

[simonecaldana]

How can a person use a Unix-like OS for any reasonable period of time without realizing even on a single user system how important the concept of a root user and root password is?

By using Mac OS X.

Where's Matrox's Parhelia?

[twicesliced]

I mean, sure, it wasn't the fastest thing on the block, but it had triple-monitor support and some other really nifty technologies in there (par for the course from Matrox). I know, not exactly a gaming card, but if you're going to include the G550, you can't leave the Parhelia out. Then there's the Px50 series too...

Re:Nice charts, what happened to Matrox?

[Jeffrey Baker]

Matrox are still the only people (of which I am aware) shipping a quad-head DVI video card. And it's passively cooled, to boot. It's the video card Bloomberg sends you with a new terminal. ATI and Nvidia would kill to have a reliable price-insensitve market like that, but they just can't make the quality goods like Matrox can. The other two are stuck in the l33t overclocker eXtreme universe.

Re:Article content is medicore at best

[feanor512]

I agree that benchmarks are important. I explained in another post that a card with a higher fill rate and more bandwidth isn't always faster.

If you look at an OpenGL game such as this one http://graphics.tomshardware.com/graphic/20031229/ vga-charts-06.html [tomshardware.com] the 5900U beats the 9800 Pro by a greater margin than the 9800 Pro beat it in that DirectX benchmark you linked.

The stats _are_ a good indicator of performance in the same family. The performance of the 9700 through the 9800XT scales mostly according to fillrate and memory bandwidth.