PowerColor confirms 69xx faster than GTX580

heflys20 said:
I recon it depends on what benchmarks you're going on. According to Techpowerup ( if the 6970 is 50% better than the 6870) it might gain a 9-15% lead over the 580. I can't speculate on whose drivers will be superior, or which will overclock better.



Again, what are you talking about?.....You said the GTX 580 was 75% faster at higher res. If it's 75% faster, then, obviously the 6870 is 75% slower. Regardless, nothing you've posted corroborates your 75% faster claim.
Click to expand...

Do this for me, 100 x (1 + 0.75) x (1 - 0.75)


Lord_Exodia said:
To actually be faster it would have to be at or more than 40% faster than 6870 technically no? Here's how I break it down.

6870 is 10% slower than 5870 which is 10% slower than GTX 480 which is 20% slower than GTX 580.
Click to expand...

I posted earlier, but I'll repeat:

580 is 143% the speed of the 6870.

So, if the 6970 is 43% faster than the 6870, the 6970 and the 580 are equal. Any faster and it will be greater.
 
heflys20 said:
I recon it depends on what benchmarks you're going on. According to Techpowerup ( if the 6970 is 50% better than the 6870) it might gain a 9-15% lead over the 580. I can't speculate on whose drivers will be superior, or which will overclock better.



Again, what are you talking about?.....You said the GTX 580 was 75% faster at higher res. If it's 75% faster, then, obviously the 6870 is 75% slower. Regardless, nothing you've posted corroborates your 75% faster claim.
Click to expand...

50% better will give it the lead but it's probably a single benchmark, the average will be around 30%.
The reason why I said drivers is because the 580 and 6970 will have similar performance. It will come down to who has better optimized drivers.
OCing sometimes depends on the starting core clock. I usually go with the card that has the lower core clock.
 
Lord_Exodia said:
To actually be faster it would have to be at or more than 40% faster than 6870 technically no? Here's how I break it down.

6870 is 10% slower than 5870 which is 10% slower than GTX 480 which is 20% slower than GTX 580.
Click to expand...

nope see if you have 10% slower than the nV gtx 580

Let x = amd 6890
Let y = nV gtx 580

Lets say nV gtx 580 gets 100 fps

So

X= Y - Y (10%)


So x = 90


So now 90 is what X is
So now if we need to find out how fast the 580 is

Y = X + X (10%)
Y = 99

Added to this if you have a data set such as different benchmarks the %'s can't be added together because the issue is magnified by the differential of each benchmark.
 
Nightbird said:
Do this for me, 100 x (1 + 0.75) x (1 - 0.75)
Click to expand...

What is this supposed to prove? That his original point was incorrect? He stated that the 50% improvement over the 6870, which was presented by the Powercolor rep, would still make the 6970 inferior to the 580. Why else would he quote the original post, and state:

That doesn't make it faster , it makes it slower
Click to expand...

If we were to utilize the 43% accumulated from the Metro bench, and not the proposed 50%, the 6970 would still be faster.

So, again, what are you showing me?
 
Okay so if it's 50% faster than 6870 then it's effectively only 7% faster than GTX 580, as the GTX 580 is 43% faster than 6870.

7% isn't much and it's a small enough lead that they will trade blows from game to game, but if amd releases at a aggressive price then it's surely gonna be a win win situation.

The video isn't clear if she's saying 30-50% faster than the 6800 series could she mean 40% across the board, there's too many ways to be able to tell where it'll be. Also is this in synthetic benchmarks or real world testing.

I'm gonna throw a curve ball here. For some reason in many of the reviews I noted that GTX 580 is SLOWER in synthetic benchmarks than it is in real world benchmarks. E.G. the first review that released showed it being either 14 or 16% faster at the most over 480. In a few of the real world tests it was anywhere from 14-24% faster than GTX 480. I thought that was really strange.

Anyway the performance is surely close either way and the price is where the spotlight needs to be now. Here's hoping... :)
 
heflys20 said:
If we were to utilize the 43% accumulated from the Metro bench, and not the proposed 50%, the 6970 would still be faster.

So, again, what are you showing me?
Click to expand...

The metro bench is just 40%, the 43% is from the aggregate. I'm pointing out that 580 being 50% faster has no source in his post.

----- scratch this, I don't understand your comment. 50% faster than the 6870 would make the 6970 faster than the 580.

Lord_Exodia said:
Okay so if it's 50% faster than 6870 then it's effectively only 7% faster than GTX 580, as the GTX 580 is 43% faster than 6870.
Click to expand...

150 / 143 = 104.9%, so about 5% faster.
 
heflys20 said:
LOL. Then why was he agreeing with you!? :p
Click to expand...

I was just doing the math, the people that are adding and subtracting when they should be multiplying and dividing have thrown around too many numbers for me to track.
 
Nightbird said:
The metro bench is just 40%, the 43% is from the aggregate. I'm pointing out that 580 being 50% faster has no source in his post.

----- scratch this, I don't understand your comment. 50% faster than the 6870 would make the 6970 faster than the 580.



150 / 143 = 104.9%, so about 5% faster.
Click to expand...

In that game yes, but not over all, ok lets do this.....

Can't believe I'm doing this right now.
http://techreport.com/articles.x/19934/8

6870 gets 50 fps

50% faster would make it 75 fps

the gtx 580 gets 121

http://techreport.com/articles.x/19934/9

6870 gets 20 fps

50% faster gets ya 30 fps

the gtx 580 gets 36 fps



Now go through the rest of that review and any other review and you will see there are some games if we take the 50% as always happening the 6980 would come out just ahead of the 580, but most of the time its slower and quite a bit.
 
HAWX 2 and Lost Planet 2 are games that are notorious for making AMD cards look bad....What is that supposed to prove?
 
and, to add further confusion, Nightbird would be disappointed if the 6970 isn't 1.67X faster then the 6870, due to 167% more electricity used. :p

I'm going to bed now. Hopefully, basic math teachings will be over tomorrow. :D

Edit: Razor1, lol at using Hawx 2 and LP2 benchmarks as examples of a general performance increase. You couldn't have picked two more controversial. :p
 
heflys20 said:
HAWX 2 and Lost Planet 2 are games that are notorious for making AMD cards look bad....What is that supposed to prove?
Click to expand...


go through the other ones if you don't take initiative f*ck it, I really don't care.:). Sorry normally I'm more patient but like my drinks!
 
razor1 said:
hand picked benchmarks designed to make Nvidia look good
Click to expand...

Take all of them, favorable and unfavorable, or take the most popular games that are unlikely to favor one specific vendor over another, and then average them please. I wouldn't mind using this instead.

Tamlin_WSGF said:
and, to add further confusion, Nightbird would be disappointed if the 6970 isn't 1.67X faster then the 6870, due to 167% more electricity used. :p

I'm going to bed now. Hopefully, basic math teachings will be over tomorrow. :D

Edit: Razor1, lol at using Hawx 2 and LP2 benchmarks as examples of a general performance increase. You couldn't have picked two more controversial. :p
Click to expand...

Good night :)
 
Tamlin_WSGF said:
and, to add further confusion, Nightbird would be disappointed if the 6970 isn't 1.67X faster then the 6870, due to 167% more electricity used. :p

I'm going to bed now. Hopefully, basic math teachings will be over tomorrow. :D

Edit: Razor1, lol at using Hawx 2 and LP2 benchmarks as examples of a general performance increase. You couldn't have picked two more controversial. :p
Click to expand...


I just did the first two in that benchmark suite, the rest you guys can go through. I already did, and I'm a happy drunk, please don't make me do the rest, the typing is aggravating. And please pick any benchmark suite doesn't really matter, I just googled them and picked the first few suites, and yeah all the benchmarks in one suite,
 
Last edited:
heflys20 said:
That link proves your 75% difference theory incorrect.....But thanks for proving my point....I guess.
Click to expand...

LOL. It is obvious some games gtx 580 or even 480 will be faster than the 6970, just because some games are way too optimized for Nvidia card, and the same goes with some games with AMD cards. One could swing the math the way the want, but I would go with over all average than picking out some games that advantage one over the other.
 
Last edited:
heflys20 said:
Again, what are you talking about?.....You said the GTX 580 was 75% faster at higher res. If it's 75% faster, then, obviously the 6870 is 75% slower. Regardless, nothing you've posted corroborates your 75% faster claim.
Click to expand...

Brah...

FPS 6870 = 100.
FPS GTX580 (+ 75%) = 175.
75% slower = -75% or 0.25.

175 * 0.25 = 43.75

75% faster is not 75% slower, you weren't using a different "base" fps to calculate the "75% slower" portion of the equation... You need to use the result of the first equation (6870 * 1.75)

You're thinking in like... Diablo 2 percents... :)
 
Again, how can a card be viewed as 75% faster, yet the card that it's 75% faster than isn't, as result, 75% slower? Typically, when someone tells me something is 75% faster, I assume the opposite is 75% slower, or 1/4 th speed. Regardless, the original argument was that the 6970's 50% (not even 43%) increase over the 6870 proposed by the PR rep meant that 580 would be "faster" than the 6970. When, if we utilize the math you provided, it equates to a 43% difference; thus, if the 6970 is 50% faster than the 6870, it beats the 580.

That's the whole point. Please, no mas!
 
you can't do it that way :cool: raw data a %'s correlate in only one direction either higher or lower not both.
 
Last edited:
razor1 said:
you can't do it that way :cool: raw data a %'s correlate in only one direction either higher or lower not both.
Click to expand...

Yes, I agree that your math is correct. But even using the correct math, your 75% faster claim is still nowhere near the truth. Where are your links that prove this? The previous link you posted shows it's around 40-45% faster on average, depending on the game and resolution.
 
jsrivo said:
Yes, I agree that your math is correct. But even using the correct math, your 75% faster claim is still nowhere near the truth. Where are your links that prove this? The previous link you posted shows it's around 40-45% faster on average, depending on the game and resolution.
Click to expand...


I just took a few reviews at the highest reses and highest AA and AF, it might not be 75% overall but in any case that wasn't my main point again I'm drinking, but I know 50% higher than a 6870 isn't going to make the 6980 faster than a gtx 580, I think it if those numbers are real (which I don't, and I can't comment which direction, higher or lower) are not correct.
 
I'm not sure why everyone is nitpicking the specifics there was nothing specific about what the rep said.

She said the 6900 series is about 30-50% faster than the 6800 series.

Here are several possibilities

the 6950 is 30% faster than 6850 and the 6970 is also 30% faster than the 6870

Or it could be something in between lets say 40%

or 39% faster across the board

or 1 is 35% faster and the other is 41% faster.

It could be the 6950 is 30% faster and the 6970 is 50% faster but that is not what she said. I think we're getting ahead of ourselves here. Is it that much faster in synthetic benchmarks or real world benchmarks, are they using Morphological AA when testsing etc.. so many possibilities The wait isn't very long and we'll get some juicy info soon.
 
That 50% higher than a 6870 might be a best case scenario anyway. I'd say GTX580 and 6970 performance would be a few percent within each other. Let's just hope the price will be good.
 
So with all this math, the 6950 will be close to the 5870. And the 6970 will be in the same ballpark as the Gtx 480.
The leaked specs is in no way putting the 6970 ahead of the Gtx 580, math or no math!
 
heflys20 said:
Seems some can't accept the possibility that the 6970 might be faster than the 580 and potentially cheaper.
Click to expand...

Doesn't bother me one bit.:cool:

I fully expect that the 6970 might be a bit better as a single card and be less capable in multi-monitor multi-GPU configs.
 
razor1 said:
75% doesn't mean 75% slower, look at the link that I linked and Nightbird's post.
Click to expand...

That would be well and done if you had stated that they 6870 was 75% as fast as the 580. You clearly stated that they 580 was 75% faster than the 6870 which is an absurd statement.
 
ya'll retarded...the card's not out yet. Who cares about numbers for a card that's not out yet!?
 
fattypants said:
Unless, of course, it scales like the 6870 and not like the 5870.
Click to expand...

Show me where the 6000s scale better in CF/Eyefinity setups than the 5000s.CF does very well with single monitors and 2 GPUS but beyond that tends to have a fair number of issues at least with the 5000s.
 
heatlesssun said:
Only 2 cards and no EF numbers I saw here.
Click to expand...
I'll let you google then; that link shows the 6800 series clearly to scale better than the 5800 series in two reviews. The big point is that 6800s do scale better in multi-gpu configs.

There are a few things that people debate about the 6870s and 6850s. Crossfire scaling is not one of them.
 
Last edited:
wtf at all the bad math in this thread?

"Okay so if it's 50% faster than 6870 then it's effectively only 7% faster than GTX 580, as the GTX 580 is 43% faster than 6870. "

If a 580 is 43% faster than a 6870, and a 6970 is 50% faster than a 6870, the 6970 is NOT 7% faster than a 580, its 4.8% faster.

"Typically, when someone tells me something is 75% faster, I assume the opposite is 75% slower, or 1/4 th speed. "

This is seriously just wrong. If item A is 100 and item B is 175, item B is 75% more than A and A is 42.86% less than B. You can't flip them around, percentages are relative to something.
 
razor1 said:
That doesn't make it faster ;), it makes it slower. On average the 580 gtx is 75% faster then the 6870 at higher settings.
Click to expand...

I really would like to know what kind of GTX 580 you talking about... does it even exist :eek:
 
heatlesssun said:
Doesn't bother me one bit.:cool:

I fully expect that the 6970 might be a bit better as a single card and be less capable in multi-monitor multi-GPU configs.
Click to expand...
lol 6xxx series rock when it comes to cf its not anything like 5xxx series so yeah you might get bothered after all :D
 
bexamous said:
wtf at all the bad math in this thread?

"Okay so if it's 50% faster than 6870 then it's effectively only 7% faster than GTX 580, as the GTX 580 is 43% faster than 6870. "

If a 580 is 43% faster than a 6870, and a 6970 is 50% faster than a 6870, the 6970 is NOT 7% faster than a 580, its 4.8% faster.

"Typically, when someone tells me something is 75% faster, I assume the opposite is 75% slower, or 1/4 th speed. "

This is seriously just wrong. If item A is 100 and item B is 175, item B is 75% more than A and A is 42.86% less than B. You can't flip them around, percentages are relative to something.
Click to expand...

This only makes sense if the 6870 is indeed 100, and the 580 is indeed 175. And, then, based on your math, that wouldn't be 75%! It would be around 42-43%, as you yourself indicated. The original poster, who started all this, stated that the 580 is 75% faster than the 6870. For some reason, folks keep trying to tell me that 75% increase doesn't equate to ......75%. That doesn't make an ounce of sense. Either it's 75% or it's 42-43%! Adding 75 to 100 isn't a 75% increase! 100 to 400 is a 75% increase!

Edit: BTW, 75% of 175 doesn't equate to 43%. No, it would equate to 43.75, which isn't a percentage, but the numerical value of 25% of 175. The numerical difference between 100 and 175 is 75%.

Can we please get off the math? It's getting old.
 
Last edited:
heflys20 said:
This only makes sense if the 6870 is indeed 100, and the 580 is indeed 175. And, then, based on your math, that wouldn't be 75%! It would be around 42-43%, as you yourself indicated. The original poster, who started all this, stated that the 580 is 75% faster than the 6870. For some reason, folks keep trying to tell me that 75% increase doesn't equate to ......75%. That doesn't make an ounce of sense. Either it's 75% or it's 42-43%! Adding 75 to 100 isn't a 75% increase! 100 to 400 is a 75% increase!

Edit: BTW, 75% of 175 doesn't equate to 43%. No, it would equate to 43.75, which isn't a percentage, but the numerical value of 25% of 175. The numerical difference between 100 and 175 is 75%.

Can we please get off the math? It's getting old.
Click to expand...

No.

Lets say we have two numbers,
100 and 150.

150/100 = 1.5 or 150%. (% is calculated by 1.5 x 100%)

150 is 1.5x of 100, as in 100 x1.5 = 150.
150 is 150% of 100, as in 100 x(150%/100%) = 150
150 is 0.5x more than 100, as in 100 + (100 x 0.5) = 150
150 is 50% more than 100, as in 100 + (100 x [50%/100%]) = 150

100/150 = .667 or 66.7% (rounded, % calculated by .667 x 100%)

100 is .667x of 150, as in 150 x .667 = 100 (rounded, won't repeat this again)
100 is 66.7% of 150, as in 150 x(66.7%/100%) = 100
100 is .333x less than 150, as in 150 - (150 x.333) = 100
100 is 33.3% less than 150, as in 150 - (150 x [33.3%/100%]) = 100


Ok, even more simply, say a graphics card costs 100$, but one day due to a shortage there is a 50% price hike. The card costs 150$ now. :( Now, suppose the shortage ends, and there is a 50% price cut, 150$ - 50% = 75$.

150 is 50% more than 100, but 75 is 50% less than 150, not 100.

Not doing this because it really matters here, but you need to know this when you are investing in retirement and getting a mortgage. Say you are investing, and your stocks increase by 100% and then drop by 100%. YOU ARE NOT WHERE YOU STARTED. :D
 
You must log in or register to reply here.
Back
Top