TR Forums

Sun Nov 18, 2007 8:18 pm

Vrock had some excellent comments on the Radeon HD3870 review.
http://techreport.com/ja.zz?id=298056

What resolutions should be tested for graphics cards and games? I've listed some possibilities.

5:4 (1.25) aspect ratios:

 Width Height MPixels
  1280  1024    1.31

This is an odd-ball aspect ratio.

4:3 (1.33) aspect ratios:

 Width Height MPixels
   640   480    0.31
   800   600    0.48
  1024   768    0.79
  1280   960    1.23
  1600  1200    1.92
  2048  1536    3.15

Each of these has 56% to 64% more pixels than the previous resolution. The 1152x864 1.00 MP resolution that works well for 17" CRTs is about halfway between 1024x768 and 1280x960.

16:10 (1.60) aspect ratios:

 Width Height MPixels
  1280   800    1.02
  1440   900    1.30
  1680  1050    1.76
  1920  1200    2.30
      *     *   3.08*
  2560  1600    4.10

Each of these has 27% to 36% more pixels than the previous resolution. I haven't seen the 3.1 MP resolution, but that's about halfway between 1920x1200 and 2560x1600.

16:9 (1.78) aspect ratios:

 Width Height MPixels
  1280   720    0.92
  1366   768    1.05
  1920  1080    2.07

These are oddly spaced.

Here is my suggestion for resolutions to be tested:

 Width Height  MPixels Aspect
  1024   768    0.79    4:3
  1280   800    1.02   16:10
  1280  1024    1.31    5:4
  1680  1050    1.76   16:10
  1920  1200    2.30   16:10
  2048  1536    3.15    4:3
  2560  1600    4.10   16:10

Each of these has 28% to 37% more pixels than the previous resolution. This evenly-spaced (logarithmic) distribution includes the most popular resolutions for 16:10, 4:3 and 5:4 aspect ratios.

Is seven too many test points? Could we drop 1024x768, or are there still too many folks attached to this old resolution?

If you want to know what performance is likely to be for one of the other resolutions, it is very probably the same as for a resolution with the same number of pixels. 1366x768 = 1152x864 = 1280x800, for example. Other resolutions could be interpolated from the excellent X-Y charts that Damage includes in TR reviews. 1600x1200=1.92 MP is about 1/3 of the way between 1.76 and 2.30 MP.

I believe that the 1920x1200 or 1920x1080 screen resolutions will be the most important in the next decade because of the global proliferation of video devices and HD video content created for the 1920x1080 resolution. If I had to pick only one resolution at which to run benchmarks, it would be 1920x1200.

Edit: Replaced elegant HTML tables with clunky fixed-pitch code blocks.

crazybus · Sun Nov 18, 2007 8:30 pm

Personally, I think 1280x800 (or the equivalent 16:9 resolution) should be the minimum resolution. With my 20" lcd that's the lowest resolution I'd want to run at. 2048x1536 seems to be an odd resolution to test at as I'm not sure how many people actually game at that resolution.

Sargent Duck · Sun Nov 18, 2007 8:31 pm

No vote?

JustAnEngineer wrote:

Here is my suggestion for resolutions to be tested:
<table><tr><td> Width </td> <td> Height </td> <td> MPixels </td> <td>Aspect</td></tr> <tr><td>1024</td> <td>768</td> <td>0.79</td> <td>4:3</td></tr> <tr><td>1280</td> <td>800</td> <td>1.02</td> <td>16:10</td></tr> <tr><td>1280</td> <td>1024</td> <td>1.31</td> <td>5:4</td></tr> <tr><td>1680</td> <td>1050</td> <td>1.76</td> <td>16:10</td></tr> <tr><td>1920</td> <td>1200</td> <td>2.30</td> <td>16:10</td></tr> <tr><td>2048</td> <td>1536</td> <td>3.15</td> <td>4:3</td></tr><tr><td>2560</td> <td>1600</td> <td>4.10</td> <td>16:10</td></tr> </table> Each of these has 28% to 37% more pixels than the previous resolution. This evenly-spaced (logarithmic) distribution includes the most popular resolutions for 16:10, 4:3 and 5:4 aspect ratios.

Is seven too many test points? Could we drop 1024x768, or are there still too many folks attached to this old resolution?

Your list is good, but I fear for poor Damage. I'd drop the bottom and top resolutions to ease Damages work load.

I was kinda pondering this the other day, And I think it would come down to what everybody has for monitors.

Say, the max resolution of a:
19" (say 1280x1024)
22"
22" Wide Screen
24" Wide Screen (1920x1200)

Now, I know max resolutions do change from monitor to monitor, but if we looked at what the "common" max resolution of a monitor is.

So if you had a "generic" 22" monitor, you could see what frame rates you'd get at the max resolution of that monitor. Of course, it'd be nice to test maybe higher, but how many people here game on something larger than 24"? I know TR members have bigger monitors, but they are in a minority.

Sun Nov 18, 2007 8:34 pm

I would be willing to drop the two lowest resolutions from the list for the majority of testing (1024x768 0.79 MP and 1280x800 1.02 MP). The 2560x1600 4.10 MP resolution needs to stay there for folks trying to play games on the monitor that I have. :wink:

danny e. · Sun Nov 18, 2007 8:41 pm

the only thing I'd say is 1920x1200 is what everyone should be on by now.

$600 for a nice 24" is really not too much.. and monitors should last you several years.

so, more 1920x1200 scores would be nice.

lethal · Sun Nov 18, 2007 8:43 pm

hmm something that would be more reasonable IMO would be to clip the bottom 2 for high end tests and cut the top two for midrange/budget reviews from the 7 res list suggested.

Sun Nov 18, 2007 8:45 pm

crazybus wrote:
2048x1536 seems to be an odd resolution to test at as I'm not sure how many people actually game at that resolution.

Good point. However, it is evenly-spaced (on a logarithmic scale) between the extremely popular 1920x1200 resolution and the 2560x1600 resolution that is the maximum that the 3007WFP will handle. 2048x1536 is supported by at least some of the games that I've tried. If we dropped the 3.1 MP resolution from the list, there'd be a pretty big gap between 2.3 and 4.1 MP.

lethal wrote:
clip the bottom 2 for high end tests and cut the top two for midrange/budget reviews from the 7 res list suggested.

That's a very good suggestion. I like it. That would include 1280x1024, 1680x1050 and 1920x1200 in all reviews and extend the resolution scale in either direction as needed. I'd bet that those three resolutions cover the majority of new monitors being sold to gamers.

Voldenuit · Sun Nov 18, 2007 9:08 pm

Too many data points. Don't forget we might want to see results with and without AA as well.

I'd suggest just:
1024x768 (budget cards only)
1280x1024
1680x1050
1920x1080
2560x1600 (high end cards only)

For everything else, we can interpolate.

There is such a thing as too much info. Also, I think it's worthwhile to focus on widescreen displays in the higher resolutions as these are more common in high end gaming setups these days. It'll also show up games that fail on widescreen support.

Sun Nov 18, 2007 9:32 pm

Voldenuit wrote:
Too many data points. Don't forget we might want to see results with and without AA as well.

That's another excellent point.

Width Height MPixels Aspect
1024     768    0.79    4:3  Used when needed for low-end cards
1280     800    1.02  16:10  Used very sparingly for low-end cards
1280    1024    1.31    5:4  Included in most reviews
1680    1050    1.76  16:10  Included in most reviews
1920    1200    2.30  16:10  Included in most reviews
2048    1536    3.15    4:3  Used extremely rarely, if at all
2560    1600    4.10  16:10  Used when needed for high-end cards

The middle resolutions should be tested first and the results of those tests should determine which, if any, additional tests should be run.

If the combination of graphics card and game settings generate plenty of frames per second, there's no point in testing it at lower resolutions. If the results are too slow to be playable at 1280x1024, try it at 1024x768. If it's still unplayable, stop there. If it's okay at 1024x768, only then try the middle resolution of 1280x800. Similarly, if a game is already unplayable at 1920x1200, there's no point in testing it at higher resolutions. If it's good at 1920x1200, try it at 2560x1600 next.

We could get this down to as few as two tests for certain extreme combinations. Test at 1280x1024--plays like a slideshow. Retest at 1024x768--still too slow. Stop. The same could be true at the high end. Test at 1920x1200--smooth as silk. Test at 2560x1600--still rockin'. Stop.

The idea is to only add test points to the performance vs. resolution curve as long as the results provide useful information. One advantage of plotting frame rates at these evenly-spaced resolutions on an X-Y chart is that you would be able to visually interpolate and extrapolate to easily make comparisons when tests are omitted due to being outside of the useful range.

Edit: Replaced elegant HTML table with a clunky fixed-pitch code section.

CampinCarl · Sun Nov 18, 2007 10:10 pm

[quote="danny e."]the only thing I'd say is 1920x1200 is what everyone should be on by now.

$600 for a nice 24" is really not too much.. and monitors should last you several years. /quote]

I don't mean to be a jerk, but, was that sarcasm? Because $600 is far too much for most people (enthusiasts included) to spend on a monitor. One would think that most people spend less than $1200 on their computer, so thinking that half of that cost on the monitor alone is 'really not too much' is just disproportionate.

Well, that's just what I think.

Entroper · Sun Nov 18, 2007 10:15 pm

Voldenuit wrote:
Too many data points. Don't forget we might want to see results with and without AA as well.

I'd suggest just:
1024x768 (budget cards only)
1280x1024
1680x1050
1920x1080
2560x1600 (high end cards only)

For everything else, we can interpolate.

There is such a thing as too much info. Also, I think it's worthwhile to focus on widescreen displays in the higher resolutions as these are more common in high end gaming setups these days. It'll also show up games that fail on widescreen support.

This hits the nail on the head. Of course, exceptions can be made when circumstances call for it, but I'd be very satisfied with reviews testing the three middle resolutions always, and one other depending on the market segment. (I'm already quite satisfied, of course. This would just be even better.)

I really want to echo the cost of monitors, too. For some of us, a 19" 1280x1024 LCD IS a pretty nice display. And a display is something that's going to last you a long time -- I upgraded from a 19" CRT that was 5 or 6 years old. I expect not to upgrade this display for a long time, and when I do, I won't be spending $500 or more on a new one.

The suggestion that you shouldn't spend more than $200 on a GPU if you don't spend $500 on a display is silly, IMO. Displays last 5 years or more regardless of what you spend on them. Graphics cards have much shorter (useful) lifespans. If your choice is between an 8600 GTS for $150 that will need to be replaced in 18 months vs. $250 for an 8800 GT which will last far longer, I think the 8800 GT wins even if you have a 19" monitor.

crazybus · Sun Nov 18, 2007 10:43 pm

Entroper wrote:
The suggestion that you shouldn't spend more than $200 on a GPU if you don't spend $500 on a display is silly, IMO. Displays last 5 years or more regardless of what you spend on them. Graphics cards have much shorter (useful) lifespans.

Which is why it's worthwhile to spend more on a nice display. The display is an integral part of the user experience so it's not a place where you should be cutting corners. A $$$ graphics card decreases in useful value a whole lot quicker than a nice display. Apparently the masses don't follow this logic as we can see with the proliferation of crappy LCDs.

dragmor · Sun Nov 18, 2007 10:43 pm

1024x768 - Minimum worth playing on a desktop
1280x1024 - Standard 17" & 19"
1680x1050 - Close enough to cover 1600x1200, standard 20" & 22".
1900x1200 - high end
2560x1600 - for top end cards only (JAE owns the only monitor at this res).

JustAnEngineer wrote:
I would be willing to drop the two lowest resolutions from the list for the majority of testing (1024x768 0.79 MP and 1280x800 1.02 MP). The 2560x1600 4.10 MP resolution needs to stay there for folks trying to play games on the monitor that I have.

The few people with the 30" are probably stuck playing at 1280x800 because their graphics setup has no hope at that resolution.

Bensam123 · Sun Nov 18, 2007 10:50 pm

1600x1200 is my prefered resolution. Although, I play almost all games at 1280x1024 for performance reasons. 1024x768 is the minimum tolerable resolution.

crazybus · Sun Nov 18, 2007 10:52 pm

JustAnEngineer wrote:
<table><tr><td> Width </td> <td> Height </td> <td> MPixels </td> <td>Aspect</td></tr> <tr><td>1024</td> <td>768</td> <td>0.79</td> <td>4:3</td> <td> Used when needed for low-end cards</td></tr> <tr><td>1280</td> <td>800</td> <td>1.02</td> <td>16:10</td> <td>Used very sparingly for low-end cards</td></tr> <tr><td>1280</td> <td>1024</td> <td>1.31</td> <td>5:4</td> <td>Included in most reviews</td></tr> <tr><td>1680</td> <td>1050</td> <td>1.76</td> <td>16:10</td> <td>Included in most reviews</td></tr> <tr><td>1920</td> <td>1200</td> <td>2.30</td> <td>16:10</td> <td>Included in most reviews</td></tr> <tr><td>2048</td> <td>1536</td> <td>3.15</td> <td>4:3</td> <td>Used extremely rarely, if at all</td></tr><tr><td>2560</td> <td>1600</td> <td>4.10</td> <td>16:10</td> <td>Used when needed for high-end cards</td></tr> </table>
The middle resolutions should be tested first and the results of those tests should determine which, if any, additional tests should be run.

If the combination of graphics card and game settings generate plenty of frames per second, there's no point in testing it at lower resolutions. If the results are too slow to be playable at 1280x1024, try it at 1024x768. If it's still unplayable, stop there. If it's okay at 1024x768, only then try the middle resolution of 1280x800. Similarly, if a game is already unplayable at 1920x1200, there's no point in testing it at higher resolutions. If it's good at 1920x1200, try it at 2560x1600 next.

We could get this down to as few as two tests for certain extreme combinations. Test at 1280x1024--plays like a slideshow. Retest at 1024x768--still too slow. Stop. The same could be true at the high end. Test at 1920x1200--smooth as silk. Test at 2560x1600--still rockin'. Stop.

The idea is to only add test points to the performance vs. resolution curve as long as the results provide useful information. One advantage of plotting frame rates at these evenly-spaced resolutions on an X-Y chart is that you would be able to visually interpolate and extrapolate to easily make comparisons when tests are omitted due to being outside of the useful range.

I agree that results are only applicable if the game is playable at that resolution, otherwise it's merely academic. I will contend that 1280x800 is actually a resolution I use. For example in Crysis it's the highest resolution that's actually playable with medium settings. And in other games I may use it when a lower resolution with AA looks better than a higher resolution without. The scaling on my screen doesn't actually look that bad.

mongoosesRawesome · Sun Nov 18, 2007 10:59 pm

crazybus wrote:
Entroper wrote:
The suggestion that you shouldn't spend more than $200 on a GPU if you don't spend $500 on a display is silly, IMO. Displays last 5 years or more regardless of what you spend on them. Graphics cards have much shorter (useful) lifespans.
Which is why it's worthwhile to spend more on a nice display. The display is an integral part of the user experience so it's not a place where you should be cutting corners. A $$$ graphics card decreases in useful value a whole lot quicker than a nice display. Apparently the masses don't follow this logic as we can see with the proliferation of crappy LCDs.

displays come down in price too, you know. it's easy to get a 22 inch monitor for less than $250. it was much more expensive just a few years ago.

I imagine that $600 display you bought will be 250 in the next couple years.

While i agree that monitors are important, i upgrade as i go and try not to buy absurdly high priced parts. When I do upgrade, I sell the remaining parts or give them to family. In the end, I've still spent less.

Entroper · Sun Nov 18, 2007 11:10 pm

mongoosesRawesome wrote:
I imagine that $600 display you bought will be 250 in the next couple years.

I didn't buy a $600 display, I spent about $200. And this was 2 years ago, when anything larger was getting out of the range of affordability. You're right, LCDs are advancing more rapidly than ever before, that's why I didn't feel it was worth going up the price curve. It also helped that I got a substantial upgrade from what I was used to at the time.

crazybus · Sun Nov 18, 2007 11:11 pm

mongoosesRawesome wrote:
displays come down in price too, you know. it's easy to get a 22 inch monitor for less than $250. it was much more expensive just a few years ago.

I imagine that $600 display you bought will be 250 in the next couple years.

While i agree that monitors are important, i upgrade as i go and try not to buy absurdly high priced parts. When I do upgrade, I sell the remaining parts or give them to family. In the end, I've still spent less.

I paid $520 for a 20" S-IPS lcd 2 years ago and they still go for around $400. Although 24" LCDs have near halved in price in the same time.

Sun Nov 18, 2007 11:12 pm

When doing a review on a top of the line $400+ card (or $800+ for SLI/crossfire) I don't think it makes any sense what so ever to test a card at anything less than 1600x1200, unless the game just doesn't run at that high of a resolution.

I like your list, but I would shorten it to 1600x1200, 1920x1200, 2048x1536, 2560x1600, and 1024x768 where 1600x1200 isn't playable.

When doing a review of midrange cards, though, they should only be tested on what they can play. I think damage does a good job of testing at relevant resolutions.

Sun Nov 18, 2007 11:26 pm

dragmor wrote:
2560x1600 - for top end cards only (JAE owns the only monitor at this res). The few people with the 30" are probably stuck playing at 1280x800 because their graphics setup has no hope at that resolution.

CounterStrike:Source, Guild Wars and Supreme Commander look great at 2560x1600 at maximum settings with 2X AA enabled on my Radeon X1950Pro 512MB. I would like to upgrade my graphics card in the next few months so that I can play the latest games at this resolution with 4X AA or higher. The demos for Unreal Tournament 3 and Crysis were too sluggish at 2560x1600, so I ran those at 1280x800. Those were the first two titles that I have encountered that I could not run satisfactorily at 2560x1600.

There are at least 4 monitors that support 2560x1600. I have the Dell UltraSharp 3007WFP. The newer 3007WFP-HC adds a wider color gamut. You can buy the same LCD panel in the Samsung SyncMaster 305T monitor. It's going to be a while before monitors of this resolution are extremely popular, but they are available in the mainstream retail channel and from the world's largest PC company.

P.S.: I forgot that you can get one from the world's most pretentious PC company, too. :wink:

danny e. · Sun Nov 18, 2007 11:36 pm

mongoosesRawesome wrote:

displays come down in price too, you know. it's easy to get a 22 inch monitor for less than $250. it was much more expensive just a few years ago.

I imagine that $600 display you bought will be 250 in the next couple years.

While i agree that monitors are important, i upgrade as i go and try not to buy absurdly high priced parts. When I do upgrade, I sell the remaining parts or give them to family. In the end, I've still spent less.

there is big difference between the good monitors and the cheap ones, though. it doesnt pay to skimp on monitors of all things.

i very much doubt the good 24" LCD's will be down to 250. I could see them dropping do around 350-400 range, but typically when all that starts to happen the quality goes as well.

for example.. just try to find a good quality 21-22"CRT now days.. you wont. they have the cheap $300 ones that are crap compared to either of mine. less expensive sometimes comes at the price of quality. this can be seen in the dvd-burner world also.. where its now a commodity product for $30.. but half of them dont work right or give out after a few months of use.

Sun Nov 18, 2007 11:43 pm

Bensam123 wrote:
1600x1200 is my prefered resolution. Although, I play almost all games at 1280x1024 for performance reasons. 1024x768 is the minimum tolerable resolution.

If you have a 4:3 CRT, you should use 1280x960 instead of 1280x1024. Otherwise, all of your on-screen characters will look shorter and/or fatter than they should.

DrDillyBar · Mon Nov 19, 2007 12:45 am

1280x1024 (1280x800 or 1366x768)
1600x1200 (1680x1050)
1920x1200 (1920x1080)
2560x1600

If you can't play the game @ 1024x768 there's something very wrong. Pick 3 or so as approperate.

I spent $900+ on my monitor when it was cutting edge, and consider $300ish for a video card approperate, considering I'm in the 1.8Mpixel category.

dragmor · Mon Nov 19, 2007 3:11 am

JustAnEngineer wrote:
dragmor wrote:
2560x1600 - for top end cards only (JAE owns the only monitor at this res). The few people with the 30" are probably stuck playing at 1280x800 because their graphics setup has no hope at that resolution.
CounterStrike:Source, Guild Wars and Supreme Commander look great at 2560x1600 at maximum settings with 2X AA enabled on my Radeon X1950Pro 512MB. I would like to upgrade my graphics card in the next few months so that I can play the latest games at this resolution with 4X AA or higher. The demos for Unreal Tournament 3 and Crysis were too sluggish at 2560x1600, so I ran those at 1280x800. Those were the first two titles that I have encountered that I could not run satisfactorily at 2560x1600.

Count me surprised, when I had the 7950GX2 is struggled at 1680x1050 with with 4 AA and AF.

JustAnEngineer wrote:
There are at least 3 monitors that support 2560x1600. I have the Dell UltraSharp 3007WFP. The newer 3007WFP-HC adds a wider color gamut. You can buy the same LCD panel in the Samsung SyncMaster 305T monitor. It's going to be a while before monitors of this resolution are extremely popular, but they are available in the mainstream retail channel and from the world's largest PC company.

The monitors are out there and cost about the same as the 24" did 2 years ago but I think it will be a very long time before its a standard gaming resolution. 30" is simply to large for most peoples setups and most of the people who want the extra large monitors are buying HDTV's.

flip-mode · Mon Nov 19, 2007 7:28 am

Entroper wrote:
Voldenuit wrote:
Too many data points. Don't forget we might want to see results with and without AA as well.

I'd suggest just:
1024x768 (budget cards only)
1280x1024
1680x1050
1920x1080
2560x1600 (high end cards only)

For everything else, we can interpolate.

This hits the nail on the head.

Yes, it does. This is a good resolution spread.

Entroper wrote:
The suggestion that you shouldn't spend more than $200 on a GPU if you don't spend $500 on a display is silly, IMO. Displays last 5 years or more regardless of what you spend on them.

Exactly. Just think of running any of the recent games on at 1280x1024 on a Geforce 6800 Ultra fer crap sake. That may have seemed a stupidly incongruent pairing of GPU and Display when the 6800 launched, but today that card could not run any of these games with all the details on at acceptable frame rates.

High end GPUs eventually become incapable of pushing high-end resolutions. For that reason getting a video card that seems far far too powerful for your monitor is not so silly as some make it out to be.

mattsteg · Mon Nov 19, 2007 8:40 am

flip-mode wrote:
Entroper wrote:
Voldenuit wrote:
Too many data points. Don't forget we might want to see results with and without AA as well.

I'd suggest just:
1024x768 (budget cards only)
1280x1024
1680x1050
1920x1080
2560x1600 (high end cards only)

For everything else, we can interpolate.

This hits the nail on the head.
Yes, it does. This is a good resolution spread.

Entroper wrote:
The suggestion that you shouldn't spend more than $200 on a GPU if you don't spend $500 on a display is silly, IMO. Displays last 5 years or more regardless of what you spend on them.
Exactly. Just think of running any of the recent games on at 1280x1024 on a Geforce 6800 Ultra fer crap sake. That may have seemed a stupidly incongruent pairing of GPU and Display when the 6800 launched, but today that card could not run any of these games with all the details on at acceptable frame rates.

High end GPUs eventually become incapable of pushing high-end resolutions. For that reason getting a video card that seems far far too powerful for your monitor is not so silly as some make it out to be.

Sure, but you're talking about a 3.5 year old card (and that wasn't a really incongruent pairing - look at TR's 6800 review(s) and there were games tested at that resolution which were already starting to challenge the 6800).

Video cards are largely "now" purchases. If you don't get good value out of a high end card before the next product refresh cycle, you probably should have gotten a cheaper card and replaced it sooner with another cheaper card that outperforms the first high end card for less total outlay.

Entroper · Mon Nov 19, 2007 9:31 am

mattsteg wrote:
Video cards are largely "now" purchases. If you don't get good value out of a high end card before the next product refresh cycle, you probably should have gotten a cheaper card and replaced it sooner with another cheaper card that outperforms the first high end card for less total outlay.

I think playing today's games with maxed settings at your monitor's native resolution, with nary a framerate hiccup in sight = good value.

lordT · Mon Nov 19, 2007 9:48 am

1280x800 and 1920x1200 are the only resolutions I am interested in.

mattsteg · Mon Nov 19, 2007 9:55 am

Entroper wrote:
mattsteg wrote:
Video cards are largely "now" purchases. If you don't get good value out of a high end card before the next product refresh cycle, you probably should have gotten a cheaper card and replaced it sooner with another cheaper card that outperforms the first high end card for less total outlay.

I think playing today's games with maxed settings at your monitor's native resolution, with nary a framerate hiccup in sight = good value.

Maybe, although defining "value" without even considering cost is kind of silly. In any case, it all depends on your sense of value, but that's a reasonable measure of the performance side of things. If you're sporting a lower resolution display that mid-range cards will run everything on, it's silly to pay the inflated cost of a high-end card "for the future" when, by the time you'll actually need it, a new midrange card may outperform it for less than the price difference between midrange and high-end. You don't buy something that obsolesces as rapidly as a high-end video card unless it's going to give you a notable benefit over the next <1 year.

flip-mode · Mon Nov 19, 2007 10:29 am

mattsteg wrote:
Video cards are largely "now" purchases. If you don't get good value out of a high end card before the next product refresh cycle, you probably should have gotten a cheaper card and replaced it sooner with another cheaper card that outperforms the first high end card for less total outlay.

I totally agree with that. But I think that given a fixed amount of cash and a "must buy now" scenario that I'd rather have a smaller monitor and more powerful card than a bigger monitor and a card that can't drive it. And I say this in the context of a machine where gaming is a primarily use. If gaming is a secondary or tertiary activity on the machine then I'd go for a bigger display and cheaper card.

It'd be a tough call for me to make honestly.

Still, given that you mention that even at launch the 6800 Ultra ran in to trouble at 12x10, that makes it sound even more reasonable to use such a display with a high-end card.

TR Forums

What resolutions should be benchmarked?

What resolutions should be benchmarked?

Revised list of resolutions to be tested

Re: Revised list of resolutions to be tested

Who is online