GTX480 -> 2-Way SLI Scaling -> Almost 100%!!!!!

Page 3 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

luv2increase

Member
Nov 20, 2009
130
0
0
www.youtube.com
Yes, you are suppose to take the 3-way results as indicative of how 3 x 5870 performed in tri-Fire because that is how they performed. You can caveat it with prospective improvements in drivers, and fully expect to see performance gains in later drivers, but to someone who just bought or is thinking of buying 3 x 5870's that it what they are going to see when they hook it up! That is an influence on the buyer's decision and very relevant.


No no no no no.... If people took the way 3 x 5870s performed in Tri-Fire with 9.8A, 9.10 RC6, 9.11, and 9.11 Hotfix, people today would still say that 3 x 5870s ARE NO BETTER THAN 2 X 5870S... Literally, at that time, 3 x 5870s were no better than 2 x 5870s. I lived it. I know what I'm talking about.

Immature drivers are immature drivers no matter how you look at it. It wasn't until 2.5 months after the 5870 was released that they truly began to shine. By your thinking, we should go back to September/October and look at those benchmarks and say that is how the 5870 performs even though the benchmarks of today show a COMPLETELY different story.

wow..
 

Daedalus685

Golden Member
Nov 12, 2009
1,386
1
0
No no no no no.... If people took the way 3 x 5870s performed in Tri-Fire with 9.8A, 9.10 RC6, 9.11, and 9.11 Hotfix, people today would still say that 3 x 5870s ARE NO BETTER THAN 2 X 5870S... Literally, at that time, 3 x 5870s were no better than 2 x 5870s. I lived it. I know what I'm talking about.

Immature drivers are immature drivers no matter how you look at it. It wasn't until 2.5 months after the 5870 was released that they truly began to shine. By your thinking, we should go back to September/October and look at those benchmarks and say that is how the 5870 performs even though the benchmarks of today show a COMPLETELY different story.

wow..

No, what we are saying is that until we have better drivers (or better cards, or whatever the problem is) what we have is what we have.... Mind you, without a source I'll have to take your word on triple crossfire scaling in the past.

You cannot go forward in time and get new drivers and you likely cannot make your own. Right now the scaling is what it is, tomorrow it will be what it will be. To compare today based on tomorrow with a gut feeling is asinine.
 

NoQuarter

Golden Member
Jan 1, 2001
1,006
0
76
No no no no no.... If people took the way 3 x 5870s performed in Tri-Fire with 9.8A, 9.10 RC6, 9.11, and 9.11 Hotfix, people today would still say that 3 x 5870s ARE NO BETTER THAN 2 X 5870S... Literally, at that time, 3 x 5870s were no better than 2 x 5870s. I lived it. I know what I'm talking about.

Immature drivers are immature drivers no matter how you look at it. It wasn't until 2.5 months after the 5870 was released that they truly began to shine. By your thinking, we should go back to September/October and look at those benchmarks and say that is how the 5870 performs even though the benchmarks of today show a COMPLETELY different story.

wow..

Lol.. no.. by my thinking, in November 3x5870's were no faster than 2x5870 and wasn't worth pursuing. In December 3x 5870's were no faster than 2x5870 and wasn't worth pursuing. Today, 3x5870's are faster than 2x5870 and is (maybe..) worth pursuing. We can only move forward on the timeline, I'm not going to rewrite history like you are trying and say in Nov it's a good idea to buy 3x5870 because in June becomes faster.

By *your* thinking no one should have bought a 5870 because 7 months later the GTX 480 is faster than it. No one should have bought a 3DFX Voodoo because 15 years later the GTX 480 blew it away.

These products and drivers did not exist at the time and could not be compared.
 
Last edited:

luv2increase

Member
Nov 20, 2009
130
0
0
www.youtube.com
To compare today based on tomorrow with a gut feeling is asinine.

Quite the contrary. Due to the historical events when it comes to GPU drivers and GPU performance, to think that what we have today with a GPU that wasn't even released 3 months ago is all we are going to get is asinine.


I will give you an analogy. Let's take a 1997 Mustang Cobra as an example. Right off the parking lot, a new 1997 Mustang Cobra has 305 horsepower @ the flywheel and delivers X amount of performance. Are we supposed to only think that X amount of performance is all that car will "ever" achieve? No. Let's say I now put a Kenny Bell Supercharger on it. I put 4:10 racing gears in it. I gave it a nice aluminum driveshaft and safety loop of course. I then put a racing suspension on it. Now, let's say our theoretical car has 550 horsepower @ the wheels. Now, our car delivers (X + 80%) of performance.

Are we supposed to still say that the 1997 Mustang Cobra after all the enhancements only delivers X amount of performance? No.

In the example:
1997 Mustang Cobra = GTX480/HD5870
Performance enhancements = GPU driver improvements
 

Skurge

Diamond Member
Aug 17, 2009
5,195
1
71
Quite the contrary. Due to the historical events when it comes to GPU drivers and GPU performance, to think that what we have today with a GPU that wasn't even released 3 months ago is all we are going to get is asinine.


I will give you an analogy. Let's take a 1997 Mustang Cobra as an example. Right off the parking lot, a new 1997 Mustang Cobra has 305 horsepower @ the flywheel and delivers X amount of performance. Are we supposed to only think that X amount of performance is all that car will "ever" achieve? No. Let's say I now put a Kenny Bell Supercharger on it. I put 4:10 racing gears in it. I gave it a nice aluminum driveshaft and safety loop of course. I then put a racing suspension on it. Now, let's say our theoretical car has 550 horsepower @ the wheels. Now, our car delivers (X + 80%) of performance.

Are we supposed to still say that the 1997 Mustang Cobra after all the enhancements only delivers X amount of performance? No.

In the example:
1997 Mustang Cobra = GTX480/HD5870
Performance enhancements = GPU driver improvements

Car analogies are not a good idea mate.
 

NoQuarter

Golden Member
Jan 1, 2001
1,006
0
76
Quite the contrary. Due to the historical events when it comes to GPU drivers and GPU performance, to think that what we have today with a GPU that wasn't even released 3 months ago is all we are going to get is asinine.


I will give you an analogy. Let's take a 1997 Mustang Cobra as an example. Right off the parking lot, a new 1997 Mustang Cobra has 305 horsepower @ the flywheel and delivers X amount of performance. Are we supposed to only think that X amount of performance is all that car will "ever" achieve? No. Let's say I now put a Kenny Bell Supercharger on it. I put 4:10 racing gears in it. I gave it a nice aluminum driveshaft and safety loop of course. I then put a racing suspension on it. Now, let's say our theoretical car has 550 horsepower @ the wheels. Now, our car delivers (X + 80%) of performance.

Are we supposed to still say that the 1997 Mustang Cobra after all the enhancements only delivers X amount of performance? No.

In the example:
1997 Mustang Cobra = GTX480/HD5870
Performance enhancements = GPU driver improvements

Sigh.. another stupid car analogy.. sounds fun..

You are thinking of buying a 1997 Mustang Cobra, but at the time there are NO aftermarket parts for it. You hear there may be some in the works but you can't be sure. You buy it anyway, and race a dude with a faster car and lose. But you claim in 5 months your car will probably be faster than his and insist you won the race. Everyone laughs at you. Because you didn't win the fucking race. Try again in 5 months and hope those aftermarket parts come out and that he didn't get upgrades too.
 
Last edited:

luv2increase

Member
Nov 20, 2009
130
0
0
www.youtube.com
Car analogies are not a good idea mate.

Why is that mate? The GTX480 and Mustang Cobra are both craved by their respective aficionados. They are both known for something we call "high performing". The word "speed" can be attributed to them. People do customize each one. People do increase the performance of each one. They both can be finely tuned. They are both expensive in their rightful respective class.

Why wouldn't they be a good analogy? LOL
 

thilanliyan

Lifer
Jun 21, 2005
12,040
2,254
126
@thilan
You point to where I said said people with 480s and 5870s will have more than one. You can't do it. I simply stated "MANY". You are about the fourth person who took what I said completely out of context.

This is what you wrote:
"those who buy the 480 and 5870 are probably going to opt for 2 of them and possibly even 3 or 4!"

I'm not asking you to provide proof that ALL 5870/480 buyers buy two or more. But at the very least you should be able to back up your claim that MANY do. What does "MANY" mean to you though? 50%? 60%? 90%

I personally think your claim is wrong but I could be wrong myself. I'm sure you would have provided something to back up what you said after multiple people have asked IF YOU COULD. Since you haven't so far, I'm guessing your claim was just an assumption on your part and seems to be a wrong one at that (especially judging by the unscientific poll on multi-GPU in the other thread). Hence, you saying we should compare XFire to SLI when comparing a 5870 to a GTX480 is incorrect.
 

Skurge

Diamond Member
Aug 17, 2009
5,195
1
71
Why is that mate? The GTX480 and Mustang Cobra are both craved by their respective aficionados. They are both known for something we call "high performing". The word "speed" can be attributed to them. People do customize each one. People do increase the performance of each one. They both can be finely tuned. They are both expensive in their rightful respective class.

Why wouldn't they be a good analogy? LOL

Well the example you gave is more akin to overclocking than driver improvements.
You don't buy drivers, they come from the manufacturer and they don't end your warranty.

Plus a 1970s cobra is more like a riva TNT or something. You can soup up the mustang as much as you want, its not gonna beat a 911 turbo around the Nurburgring.

EDIT: My bad, I meant 1997 cobra
 
Last edited:

mfenn

Elite Member
Jan 17, 2010
22,400
5
71
www.mfenn.com
I don't know if many of you have seen this review by Guru3D or not. I'm sure it was probably posted on here before. Many people seem to "want" to focus on pitting a single 5870 up against a single 480. But, why would we do this when more likely than not, those who buy the 480 and 5870 are probably going to opt for 2 of them and possibly even 3 or 4!

OP, here's a better car analogy for you.

You're basically saying that the majority who want to spend more than $450 on a GPU end up spending $1000-$2000. :rolleyes::rolleyes::rolleyes::rolleyes::rolleyes:

That's like saying nobody buys a Mustang GT because a Shelby GT500 is only twice as much! Hell, by your logic, nobody buys a Mustang GT because a Porsche 911 GT3 is only 4 times as much.

Seriously, give it a rest.
 
Last edited:

Paratus

Lifer
Jun 4, 2004
17,522
15,567
146
Would it be wrong to suggest that this thread is now about pie? Makes as much sense as cars.
 

Puffnstuff

Lifer
Mar 9, 2005
16,187
4,871
136
The first thing I noticed after I installed my second 480 was that in jc2 my framerates doubled indicating 100% scaling.
 

tincart

Senior member
Apr 15, 2010
630
1
0
You guys are really blowing his way out of proportion to a very disturbing level. Seriously, show some maturity.

Funniest thing I have read all day. I actually find it much more entertaining to read your posts under the assumption that you are a master of satire.
 

luv2increase

Member
Nov 20, 2009
130
0
0
www.youtube.com
The first thing I noticed after I installed my second 480 was that in jc2 my framerates doubled indicating 100% scaling.

I noticed the same thing too. It's very cool indeed.


@tincart
I said what I said because it seems like it is darn near impossible for anyone to have serious discussions on this thread. People want to be sarcastic, throw insults, act childish, make fun of others, go off-topic and etc.. just because they don't agree with someone.

People need to have more decency.

A prime example is the second post of this thread. The individual asked me is I felt better now? Seriously, that is just rude and uncalled for. The sad part is that the poster has been a member since 2004 and has almost 2,500 posts. I can't stand negativity. People need to learn to smile and be considerate. I'm not saying I am perfect because I have a lot of work to do as well.
 
Last edited:

Madcatatlas

Golden Member
Feb 22, 2010
1,155
0
0
Car analogies are bad skywalker..
they JUST dont work. reasons pointed out again and again.
You may like cars AND GPUS, but you dont make comparisons between the two.


Its interesting to note from Lonyos crossfire/sli poll, that very few people have more than 1 card. in other words, tri crossfire or tri sli is for the very few.
 

luv2increase

Member
Nov 20, 2009
130
0
0
www.youtube.com
Car analogies are bad skywalker..
they JUST dont work. reasons pointed out again and again.
You may like cars AND GPUS, but you dont make comparisons between the two.


Its interesting to note from Lonyos crossfire/sli poll, that very few people have more than 1 card. in other words, tri crossfire or tri sli is for the very few.


For the 100th time, that poll had nothing to do with people buying the extreme high end cards and SLI/CF.

I am actually shocked that so many people have had CF/SLI to be quite honest. I wouldn't think many people with non-extreme cards would be putting so many of them in multi-GPU configs.
 

tincart

Senior member
Apr 15, 2010
630
1
0
@tincart
I said what I said because it seems like it is darn near impossible for anyone to have serious discussions on this thread. People want to be sarcastic, throw insults, act childish, make fun of others, go off-topic and etc.. just because they don't agree with someone.

People need to have more decency.

No. You said what you said because pretty much everyone in this thread disagrees with you and they are providing good arguments against you.

Disingenuously painting yourself as the victim in a situation where most people justifiably think you're wrong is not a winning strategy, but that's just my opinion man. More that that, crying wolf indicates a certain lack of... decency.
 

MJinZ

Diamond Member
Nov 4, 2009
8,192
0
0
No doubt. I skews the numbers by a factor of 2-3 at least. I'd venture to guess that less than 10% of PC gamers have ever had SLI/CF.

I had Crossfire. It sat there looking pretty, and never played anything more than 3Dmarks.

Ah, the day Diablo 3 comes out...

Too bad I'd probably max out D3 @ 100FPS with a 5 year old graphics card.
 

luv2increase

Member
Nov 20, 2009
130
0
0
www.youtube.com
No. You said what you said because pretty much everyone in this thread disagrees with you and they are providing good arguments against you.

Disingenuously painting yourself as the victim in a situation where most people justifiably think you're wrong is not a winning strategy, but that's just my opinion man. More that that, crying wolf indicates a certain lack of... decency.


What did I say what I said?
 

thilanliyan

Lifer
Jun 21, 2005
12,040
2,254
126
No one would have disagreed with much in this thread if all it was about was how good Fermi's scaling is in 2-way SLI. However, you (luv2increase) for some reason decided to encompass high-end single card recommendations in it as well by proclaiming that many people who buy 5870s or 480s would buy more than one, when that is in fact not the case (or have you some evidence of that now?). Thank you for showing us that Fermi scales well with 2-way SLI and is highly recommended for people wanting to spend $1000+ on GPUs (NOT the people who are happy sticking with single high end cards).
 

luv2increase

Member
Nov 20, 2009
130
0
0
www.youtube.com
No one would have disagreed with much in this thread if all it was about was how good Fermi's scaling is in 2-way SLI. However, you (luv2increase) for some reason decided to encompass high-end single card recommendations in it as well by proclaiming that many people who buy 5870s or 480s would buy more than one, when that is in fact not the case (or have you some evidence of that now?). Thank you for showing us that Fermi scales well with 2-way SLI and is highly recommended for people wanting to spend $1000+ on GPUs (NOT the people who are happy sticking with single high end cards).


Point taken. They don't do SLI/CF testing in reviews for the heck of it. They do it because there is a market out there for it. They didn't create SLI/CF for the heck of it. They did it to improve performance. Many people who opt for the absolute top-end obviously want a lot of performance. Some of these people will obviously get 2 or more cards for SLI/CF.

The moral is that if you have plans on doing a 2-way config, your best bet is to go with the 480. If you plan on just getting a single top-end card with no plans in the future to get another one, then go for the 5870.
 

tannat

Member
Jun 5, 2010
111
0
0
Let's see:

80% scaling SLI
70 % scaling CF

10% better performance GTX480 than HD5870.
(In canned benches)

(1,1x1,8)/(1x1,7)=1,164

Yes, GTX480 SLI looks like a viable deal when it's less than 15% more expensive than HD5870 CF. At least for benchmarks.

Agreed
 

MagickMan

Diamond Member
Aug 11, 2008
7,460
3
76
Would it be wrong to suggest that this thread is now about pie? Makes as much sense as cars.

I've noticed that the frequency of the Green attacks have increased of late. Does it have anything to do with the glut in the retail channel? Maybe it's due to the fact that places like Newegg are giving away power supplies and hard drives just to unload GTX 4xx cards?

nV marketing honcho: "Go to the enthusiast sites and talk up our cards and drag ATI down, they're kicking our ass!"

;)