A plea from a PC gamer - Best Single card solution right now?

Page 4 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Janooo

Golden Member
Aug 22, 2005
1,067
13
81
There is nothing wrong with SLI/CF, it's perfectly seamless if you know how to use it. SLI more than CF in this case, because of customizable game profiles.

But you don't need either one for playing at 1920x1080. One GTX 480 will do. And you get customizable profiles (huge in terms of getting the best IQ for different games), physx (bleh), and cuda (yay). Heat and noise are a non-issue. The only negative is power consumption. If it's too rich for your blood get a GTX 470.

If you aren't willing to put up with the power consumption and don't care for the "nvidia package" then just get a 5870. It's a great card there is no question about it. If I were you and was going with ATI I'd actually get a 5970, but you said it's out of your budget.
FYI, you are contradicting yourself.
Higher power consumption generates more heat. The heat becomes an issue as well.
 

thegimp03

Diamond Member
Jul 5, 2004
7,420
2
81
I bought a 5870 recently and couldn't be happier. I run everything on max settings at 1920x1200 resolution.
 

Triglet

Senior member
Nov 22, 2007
260
0
76
Everyone always says they're running everything on max settings at blah blah resolution. While this may be true, the bigger issue is what kind of framerates you're getting (and in my opinion, what minimum framerates you're getting) at whatever settings/resolution. I play 90% FPS's, and could care less about average or max frames -- minimum's are what matters.

For example, let's say you're playing BC2 with the 5870 at max settings @ 19x12 (assume a badass cpu). For me, the minimum frames that kind of setup gives is not acceptable. Especially if you start enabling edge detection and slamming the AA quality all the way up.

Essentially, what's acceptable is ultimately in the eyes of the beholder. What was acceptable for me when I started gaming and where my expectations are now are two completely different things. At the end of the day, buy the best you can afford -- if that's a 5870 awesome. If that's a 480 even better!
 

taltamir

Lifer
Mar 21, 2004
13,576
6
76
Everyone always says they're running everything on max settings at blah blah resolution. While this may be true, the bigger issue is what kind of framerates you're getting (and in my opinion, what minimum framerates you're getting) at whatever settings/resolution. I play 90% FPS's, and could care less about average or max frames -- minimum's are what matters.

agreed, frankly I don't think there is any point to even measuring max framerate. And average framerate is nearly worthless (but not entirely)... still min framerate is the only metric I go by

And framerate is not an ideal way to describe it... i prefer "max time to render frame" measured in milliseconds.
 

Skurge

Diamond Member
Aug 17, 2009
5,195
1
71
agreed, frankly I don't think there is any point to even measuring max framerate. And average framerate is nearly worthless (but not entirely)... still min framerate is the only metric I go by

And framerate is not an ideal way to describe it... i prefer "max time to render frame" measured in milliseconds.

Personally, I'd rather have good averages and low minimums than high minimums and low averages.

I rather have a blip to 15fps every odd scene, but average 45 then keep 25fps minimum with an average of 30fps.

Dirt 2 as an example, if I can keep an average of 45 (which I do :D) and have a minimum of 15fps that happens when say I hit a water splash I would be much happier than not having any slowdowns during water splashes but having to play the game in semi-slow motion at an average of 30fps.


Just my 2 cents, basically I spend more time playing at the avg speed than the minimum. So it more important to me.
 

GaiaHunter

Diamond Member
Jul 13, 2008
3,732
432
126
What all we want is graphics card providing a smooth gameplay and that deviates the least from that smoothness.

Higher minimums aren't exactly a perfect indication of smoothness, like an average isn't.

Going to 15 fps from 60 fps avg is about the same as going to 7 fps from 60 fps avg - we will all notice the smoothness was lost.

What we need to know is how much time does a card spend not reaching smoothness conditions.
 

Skurge

Diamond Member
Aug 17, 2009
5,195
1
71
What all we want is graphics card providing a smooth gameplay and that deviates the least from that smoothness.

Higher minimums aren't exactly a perfect indication of smoothness, like an average isn't.

Going to 15 fps from 60 fps avg is about the same as going to 7 fps from 60 fps avg - we will all notice the smoothness was lost.

What we need to know is how much time does a card spend not reaching smoothness conditions.

.feel free to correct me by lets say you have 15fps minimum, 60fps avg.

The gap between between the minimum and avg is so large you have to assume it spends very little time at or anywhere near 15fps.
 

Cookie Monster

Diamond Member
May 7, 2005
5,161
32
86
Well if you want the best single GPU card, its the GTX480. The second down is obviously the HD5870.

For a GTX480, overclocking will be a breeze (via MSI afterburner, a fantastic tool IMO) and potentially net you another 15~25% performance with the stock cooler. Changing its fan profile can really help keep the temps down from the stock fan profile. If you could keep the heat in check (additional cooling), this card could be pushed even more.
 

JAG87

Diamond Member
Jan 3, 2006
3,921
3
76
FYI, you are contradicting yourself.
Higher power consumption generates more heat. The heat becomes an issue as well.

I didn't say it doesn't get hot. I said it's a non-issue because the noise is well under control.
 

Voo

Golden Member
Feb 27, 2009
1,684
0
76
.feel free to correct me by lets say you have 15fps minimum, 60fps avg.

The gap between between the minimum and avg is so large you have to assume it spends very little time at or anywhere near 15fps.
Just from the numbers it could be anything: 15fps minimum and 60avg can be 50% 15fps and the other 50% 105fps or every other combination you can imagine.

A high average frame rate may be great if you're interested in benchmarking, but if it starts to lag everytime there's some action on the screen that's just unacceptable imho. Who cares about 100fps average, if my gameplay is compromised because I get lags in the worst moments? Though like was already said often enough: Just the numbers aren't meaningful, if the minimum framerate is just a sole outlier there can be many reasons for it (and that's not a real problem), but if it that happens repeatedly that's bad.
 

Skurge

Diamond Member
Aug 17, 2009
5,195
1
71
Just from the numbers it could be anything: 15fps minimum and 60avg can be 50% 15fps and the other 50% 105fps or every other combination you can imagine.

A high average frame rate may be great if you're interested in benchmarking, but if it starts to lag everytime there's some action on the screen that's just unacceptable imho. Who cares about 100fps average, if my gameplay is compromised because I get lags in the worst moments? Though like was already said often enough: Just the numbers aren't meaningful, if the minimum framerate is just a sole outlier there can be many reasons for it (and that's not a real problem), but if it that happens repeatedly that's bad.

I guess the only way would be graphs. . . ugh, I gate reading graphs.
 

GaiaHunter

Diamond Member
Jul 13, 2008
3,732
432
126
.feel free to correct me by lets say you have 15fps minimum, 60fps avg.

The gap between between the minimum and avg is so large you have to assume it spends very little time at or anywhere near 15fps.

My point is about all the rage minimum frame rates get these days.

Who cares if the minimum frame rates are higher if they still don't provide smoothness?

Just looking at 3 sets of numbers, min, max and avg doesn't tell all the story, unless the minimum reaches smooth playing (lets say min is 40) or the avg is unplayable.

In those 2 situations is easy to draw proper conclusions.

Now if min as are 15 for a card and 1 for another but the averages are similar, either the benchmark isn't properly done or you need a graph to see what is happening (maybe you can see driver bug or some area where a bottleneck in the card happens). If the first card spends 10% of the time at around 15 fps and the second only reaches 1 fps in one point, I'm not exactly sure which would be the best card.