SMP-Oriented Video Card Round-up

Jason Mitchell writes "I just noticed that 2CPU.com has posted a rather large video card round-up. They ran game and application benchmarks on a dual Athlon MP and Xeon workstation and also did some unique qualitative testing pertaining to s-video output quality. It's a good read."

Comments

Why?

[floamy]

Why only the older video cards no 9700pro/gffx?

Re: SMP-Oriented Video Card Round-up

[Dunark]

Too bad they didn't mention one of the bummers about the Matrox G550: It only supports video playback to the S-Video output when you set your whole desktop to 1024 x 768 16-bit color. This is a major disappointment if you're used to running your display at 1600 x 1200 24-bit.

s-video

[Mononoke]

The resolution of the s-video output was limited to 800x600...

I'll bet the s-video output was limited to 400 lines of NTSC video. There is no XXXbyXXX measurement of NTSC video.

Yeah, I know what he meant (ie: The highest resolution that could be downconverted to NTSC was 800x600.) but most people won't, and that's the whole point of a review.

Next time you want to compare s-video outputs, use the proper tools and terms.

Video Card Reviews

[argmanah]

My personal preference as far as hardware review sites is Tom's Hardware Guide (formerly http://sysdoc.pair.com). He gives much more insight into testing methodology and has access to a greater variety of hardware than the article linked to in the story. He also does more testing than game framerates, like Solidedge and 3D Studio Max benchmarks.

In addition, Tom sorts his results! The results in the story's article aren't sorted by performance, so if I want to find the card that performed the best in any specific benchmark, I have to scroll up and down the chart to see which number is highest.

Admittedly, your mileage may vary on a system with multiple processors, but in the end, this is a video card test, isn't it?

MX?

[dolo666]

I'm fairly happy with my Geforce 4 MX, which is a big step up from my old Geforce 1 ddr. I've been utilizing the video capture feature of it and you can download a movie of me outrunning cops in Vice City, HERE.

I'm just curious... what is so bad about MX that it only cost me $112 Canadian dollars to get the card? I find that it gives me pretty good fps in Quake 3.

But Doom 3 will be another story, methinks. :)

Re:MX?

[htmlboy]

It's reasonably fast at what it does, so it will run Doom 3 at a decent framerate.

in a .plan update from a few months ago, john carmack said this:

Do not buy a GeForce4-MX for Doom.

Nvidia has really made a mess of the naming conventions here. I always thought it was bad enough that GF2 was just a speed bumped GF1, while GF3 had significant architectural improvements over GF2. I expected GF4 to be the speed bumped GF3, but calling the NV17 GF4-MX really sucks.

so a gf4mx will run doom, but it won't be pretty.

Re:Quality of S-video outputs

[MBCook]

A standard NTSC TV only has 525 lines or so (I think it's 625 for PAL), but that includes the VBI (where closed caption stuff is sent, among other things, IIRC). So if your watching DVDs using some software DVD player, then what's going on is this:

1. DVD Video is usually about 480 lines high (I think)
2. This gets convered to 600 or 768 or some other resolution for display on the monitor
3. The TV-Out hardware has to take that number, and turn it back into the resolution of the TV, whatever that may exactly be

So basically the problem is that you're blowing it up, then shrinking it back down, and you lose a little quality each time.

Now if your card is good (or if you have a hardware DVD only card like a ReelMagic (great cards)) then things are different. When you use the TV out on such cards (or if a normal card can put the DVD out directly to the TV skipping the middle) then you only have one scaling, the same scaling that a DVD player would do. This gives you a much better picture. My ReelMagic Hollywood+ rivals most standard DVD players (up to about $150 maybe?) in my eyes.

So basically the problem is that it's much easier (I think) to simply decode the DVD onto the framebuffer and output that (the first method) than to bypass the framebuffer and output directly to the TV (the second method). And let's face it, while it was great to watch DVDs on your PC and hook it to your TV years ago, you can now get a decent new DVD player for under $80, so it's not a feature in high demand.

Re:read the article asshole

[Lurgen]

I read the article you fool.

Q3 featured strongly in their list of benchmarks, especially when they measured the impact of anti-aliasing and dual-head setups.

2D performance is almost irrelevant, and if you understood most of their benchmarks you'd know that despite them not being 3D games, they were still mostly 3D related.

Re:That sucked!

[Arandir]

What the f&#k were they thinking, including an antique Matrox in the list? And that Radeon 7500...? OK, they were nice a year ago, but who cares!

Okay, time to burn some of my limitless karma...

Since when is one year ago ancient? Just because something is older than the last time you changed your underwear doesn't make it ancient. I'm getting pretty sick and tired of you munchkins running the video market. It makes it tough for the rest of us who want a solid stable video card instead of whatever the prepube crowd wants this week.

These aren't "crappy old cards", they're superb modern cards that have been around long enough to prove their merit. Maybe you should go read some reviews about last months cards. Or are those still too archaic for you? Maybe you need to wait till next week to read about this week's cards.

Some of us have better things to do than to buy a new video card everytime the industry says "jump". Some of us has dropped out of the constant upgrade rat race that you kids insist on playing.