Polaris performance numbers perf/$

[Araxie]

But no, he didn't.

You're contradicting yourself. First, you state that there is no reference GTX 970 (you're correct, it's only a reference spec). So, you state that we can only compare overclocked cards (mostly correct, there are some reference clocked cards of each, though they are the minority). But then you cite Alkaladur's post, which actually compared a reference clocked GTX 970 to an ASUS Matrix, which has a 21.6% base overlock of the core, and a factory memory overclock (which not all cards do). This isn't the comparison that you think it is, and now I have to ask:

Are you deliberately cherry picking to fit a theory, or are you genuinely unaware that it was not an apples to apples comparison?

nah I never cherry pick, I just have to say, even if in the reviews doesn't show the clock of the cards you can easily asume that even if the name say "GTX 970" that card aren't running at "reference" clocks , that card used is probably some kind of Asus/EVGA/MSI/Gigabyte/Zotac/etcetcetc custom cooled card running at a minimum 1350mhz boost clock, Im failing here by assuming how a card behave out of the box?. that's why I do a big emphasis in do comparisons..

let me add some more complex things to the mix, let me add this Colorful iGame GTX 970, why this card? just picked some X card with a 1402mhz average boost clock (according review). and some good numbers:

Spoiler: Summary 1080

Spoiler: Summary 1440

Colorful iGame GTX 970 4 GB Review

Now I want to year you, How do you analyze that?

 

[Alkaladur]

You're right in that 4k is where it's most obvious. And your cherry picked game showed a 94% gain. And my cherry picked game showed a 38% gain. Again, the reason that you test multiple games, and not just one game is to get an overall picture of how a card performs. And the CPU limitation does matter, just not the way that you intend it to.

Let me clarify. My games run at a certain speed, let's say, 100 miles per hour. In order for me to justify spending $300 on a new GPU, I want to go 200mph. Are you with me? Going from a 970 (100mph) to a 980ti (135-145mph) is NOT 200mph, even if I were to change my CPU (not changing the CPU, so the CPU limitation stays, regardless of what it is). So, if AMD comes out with a new card that runs as fast as the 980ti...will it get me to 200mph? Nope.

I cannot explain it any simpler than that. My entire point was that I don't upgrade until I get 2x benefit. I showed a 1.35x-1.45x benefit, so I'm staying pat. You argued a 1.6x benefit with no data to back it up (still not 2x). You then reduced your argument to 1.5x (data doesn't back this up). You then showed one particular game with one specific scenario where it ALMOST eeked out 2.x. So unfortunately, despite your best efforts to cherry pick data in your favor, you couldn't demonstrate 1) that my 1.35x-1.45x data was incorrect, and 2) that the card would be a 2x upgrade for me.

But, you said that was your last attempt, so I'll make this mine. I bear no animosity towards you. And I'm not going to claim to be a GPU expert. All I'm going to say is that the data before me convinced me that an upgrade at this time isn't worthwhile. Like you, I thought it was 1.6x at LEAST, but then the data changed my mind.

60% advantage is because of shader count.

2816/1664

69% advantage, clock for clock

50% actual performance in games is commonplace

Strange of you to expect double performance, never has there been a enthusiast gpu 2x as fast the high end gpu of the same generation

 

[KazeoHin]

60% advantage is because of shader count.

2816/1664

69% advantage, clock for clock

50% actual performance in games is commonplace

Strange of you to expect double performance, never has there been a enthusiast gpu 2x as fast the high end gpu of the same generation

He's saying he does not want to upgrade until there is an option that fits his performance needs at the price he is willing to pay.

 

[Alkaladur]

He's saying he does not want to upgrade until there is an option that fits his performance needs at the price he is willing to pay.

Considering the 970 successor is expected to deliver roughly 980ti performance for the same price as the 970, he's gonna wait a long time

 

[N4CR]

Nice picture of the Polaris 11 die, must be a joint venture between Canon and AMD.

AMD allowing Canon to photograph their upcoming GPU die with Canon's state-of-the-art 128k pixel CCD sensor

It'll be some poor chinese workers' 1mp flip phone or similar...

No one worth their salt will leak Polaris spec. I know it will be showed it off within the next month though..

 

[Daniel_Chang]

nah I never cherry pick, I just have to say, even if in the reviews doesn't show the clock of the cards you can easily asume that even if the name say "GTX 970" that card aren't running at "reference" clocks , that card used is probably some kind of Asus/EVGA/MSI/Gigabyte/Zotac/etcetcetc custom cooled card running at a minimum 1350mhz boost clock, Im failing here by assuming how a card behave out of the box?. that's why I do a big emphasis in do comparisons..

That's where you made a mistake. You assumed. TechPowerUp's charts use all reference-clocked cards. And when they do a review of a specific model, they compare it not only to those cards, but also the reference of that version (like how the ASUS Matrix was also compared to the reference 980ti). So yes, his comparison was of a grossly overclocked 980ti vs. a reference 970. And it STILL didn't meet his original claimed delta!

let me add some more complex things to the mix, let me add this Colorful iGame GTX 970, why this card? just picked some X card with a 1402mhz average boost clock (according review). and some good numbers:

Now I want to year you, How do you analyze that?

I don't. The 980ti isn't in that comparison, and it's important to pull data from the same comparison. Because they update these charts with different drivers tested. For example, in my original source, it showed a 40.8% delta between the 970 and 980ti at 1440p. However, a follow-up post by Alkaladur showed a roughly 42% delta at 1440p. New drivers had increased the delta.

So, if I plugged the numbers from your links into data imported from another review, it may increase or decrease the delta, but the comparison wouldn't be apples to apples. It would not be valid.

He's saying he does not want to upgrade until there is an option that fits his performance needs at the price he is willing to pay.

Right, however, we did get into a debate about specific numbers. I made a claim, and the claim was not widely accepted, so it was on me to back up that claim. I feel that I have. The data supports my conclusion.

Considering the 970 successor is expected to deliver roughly 980ti performance for the same price as the 970, he's gonna wait a long time

 

[Alkaladur]

That's where you made a mistake. You assumed. TechPowerUp's charts use all reference-clocked cards. And when they do a review of a specific model, they compare it not only to those cards, but also the reference of that version (like how the ASUS Matrix was also compared to the reference 980ti). So yes, his comparison was of a grossly overclocked 980ti vs. a reference 970. And it STILL didn't meet his original claimed delta!



I don't. The 980ti isn't in that comparison, and it's important to pull data from the same comparison. Because they update these charts with different drivers tested. For example, in my original source, it showed a 40.8% delta between the 970 and 980ti at 1440p. However, a follow-up post by Alkaladur showed a roughly 42% delta at 1440p. New drivers had increased the delta.

So, if I plugged the numbers from your links into data imported from another review, it may increase or decrease the delta, but the comparison wouldn't be apples to apples. It would not be valid.



Right, however, we did get into a debate about specific numbers. I made a claim, and the claim was not widely accepted, so it was on me to back up that claim. I feel that I have. The data supports my conclusion.

This should match 'reference' 970 clocks, so is as fair a comparison as well ever be able to make really, compare to the same aib 970 posted above by araxie and you should see 50% improvement is the norm at 1440p+

Colorful iGame GTX 980 Ti 6GB Review

The colorful 970 posted above by araxie is at 1418mhz boost compared to 1328 on this one.

 

[Araxie]

That's where you made a mistake. You assumed. TechPowerUp's charts use all reference-clocked cards. And when they do a review of a specific model, they compare it not only to those cards, but also the reference of that version (like how the ASUS Matrix was also compared to the reference 980ti). So yes, his comparison was of a grossly overclocked 980ti vs. a reference 970. And it STILL didn't meet his original claimed delta!



I don't. The 980ti isn't in that comparison, and it's important to pull data from the same comparison. Because they update these charts with different drivers tested. For example, in my original source, it showed a 40.8% delta between the 970 and 980ti at 1440p. However, a follow-up post by Alkaladur showed a roughly 42% delta at 1440p. New drivers had increased the delta.

So, if I plugged the numbers from your links into data imported from another review, it may increase or decrease the delta, but the comparison wouldn't be apples to apples. It would not be valid.

In this case I was more speaking about GTX 970 Clock numbers, that's why I looked for a specific GTX 970 Review with a stated average 1402mhz average clock versus a "reference" clocked GTX 970. I was asking to analyze for that specific reason, not to directly compare to a GTX 980TI but instead Clocked reference 970 versus custom 970, and the results. so what are those results with a whopping ~7% performance increase at 1400mhz+ over a "reference" clocked GTX 970. what could be this?. is the GTX 970 poor scaling? or the "reference" also have a high boost clock..? those are the things to analyze and its why I always asume (my mistake? Yes, but mostly im right assuming that) that every GTX 970 analyzed in any review is over 1300mhz boost clock...

 

[Daniel_Chang]

This should match 'reference' 970 clocks, so is as fair a comparison as well ever be able to make really, compare to the same aib 970 posted above by araxie and you should see 50% improvement is the norm at 1440p+

Colorful iGame GTX 980 Ti 6GB Review

The colorful 970 posted above by araxie is at 1418mhz boost compared to 1328 on this one.

Key word, should. Forgive me for not taking your 50% estimate at face value give how this thread has gone so far.

Run the numbers. Show me. I've shown you repeatedly and so far, you've proven immune to hard data. It's your turn to crunch the numbers.

 

[Alkaladur]

Key word, should. Forgive me for not taking your 50% estimate at face value give how this thread has gone so far.

Run the numbers. Show me. I've shown you repeatedly and so far, you've proven immune to hard data. It's your turn to crunch the numbers.

Run them yourself I sent you the link.

100/65 > 150%

More than that I cannot do, instead I'll go to bed and leave you your strange expectations of double performance.

 

[Daniel_Chang]

Run them yourself I sent you the link.

100/65 > 150%

More than that I cannot do, instead I'll go to bed and leave you your strange expectations of double performance.

So, I just want to be clear on this.

Me: 980ti is roughly 35% faster than 970 (shows data).

You: No, it's a comfortable 60% faster (shows no data).

Me: No, it's roughly 40% (shows data, corrects prior error in math).

You: No, it's up to 50% (shows GTX 980ti grossly overclocked against stock 970, still doesn't match your numbers, which I had to run for you).

Me: ^ran above numbers for you to show that your data didn't back up your claim, used a counter example showing 30-35% gain to show the error in cherry-picking.

You: Well, maybe this will work.

Me: Show me.

You: No, too lazy, you do it.

Dude, if you can't even be bothered to attempt to prove your own point, there's no debate here. You essentially just forfeited. Like you, I'm calling it a night. Let me know when you can be bothered to cite and run valid numbers to prove your own point. I'm not doing your work for you anymore.

 

[Alkaladur]

So, I just want to be clear on this.

Me: 980ti is roughly 35% faster than 970 (shows data).

You: No, it's a comfortable 60% faster (shows no data).

Me: No, it's roughly 40% (shows data, corrects prior error in math).

You: No, it's up to 50% (shows GTX 980ti grossly overclocked against stock 970, still doesn't match your numbers, which I had to run for you).

Me: ^ran above numbers for you to show that your data didn't back up your claim, used a counter example showing 30-35% gain to show the error in cherry-picking.

You: Well, maybe this will work.

Me: Show me.

You: No, too lazy, you do it.

Dude, if you can't even be bothered to attempt to prove your own point, there's no debate here. You essentially just forfeited. Like you, I'm calling it a night. Let me know when you can be bothered to cite and run valid numbers to prove your own point. I'm not doing your work for you anymore.

...

1328mhz...54% advantage

Again, 60% improvement in shader number, actually 69.

69% shader advantage, clock for clock.

Please use brain. Thanks, and please shorten your posts, devoid of reason as they are you can at least spare me some brain lesions by being concise.

 

[Daniel_Chang]

...

1328mhz...54% advantage

Again, 60% improvement in shader number, actually 69.

Please use brain. Thanks, and please shorten your posts, devoid of reason as they are you can at least spare me some brain lesions by being concise.

Ahh, insults. And, we're done here. I don't play that game.

 

[Alkaladur]

Ahh, insults. And, we're done here. I don't play that game.

You don't play that game because it won't run, you'd need the 50+% perf advantage for that.

 

[ChosenUno]

Sigh...

Here...

970 Strix overclocked to 1443/8002.
ASUS GeForce GTX 970 Strix review

980Ti Strix overclocked to 1492/8000.
ASUS GeForce GTX 980 Ti STRIX Review

Comparing Bioshock Infinite at 1440p: 134FPS/86FPS = 59.5% difference.
Comparing Hitman Absolution at 1440p: 101FPS/63FPS = 60.3% difference.
Comparing Tomb Raider at 1440: 161FPS/96FPS = 67.7% difference.
Comparing 3DMark Fire Strike: 20903/12495 = 67.3% difference.

Let's say the 980Ti has a 3% clock advantage (1492/1443=1.033, i.e. 3.3% difference) and that performance scaling with clock is linear. That still means on average over these 4 benchmarks the 980Ti still leads the 970 by 61.8%.

At the very minimum you should expect a 50% increase when both sides are clocked equally, because the 980Ti has 50% more ROPs, 50% more bandwidth, 69% more CUDA cores, and 50% more VRAM. If not then there's likely a bottleneck with the CPU.

 

[SpeedyVV]

Ahh, insults. And, we're done here. I don't play that game.

You don't play that game because it won't run, you'd need the 50+% perf advantage for that.

Someone once said "winning an argument on the internet is like winning the special Olympics. even if you are the winner, you are still a retard".

 

[Alkaladur]

Sigh...

Here...

970 Strix overclocked to 1443/8002.
ASUS GeForce GTX 970 Strix review

980Ti Strix overclocked to 1492/8000.
ASUS GeForce GTX 980 Ti STRIX Review

Comparing Bioshock Infinite at 1440p: 134FPS/86FPS = 59.5% difference.
Comparing Hitman Absolution at 1440p: 101FPS/63FPS = 60.3% difference.
Comparing Tomb Raider at 1440: 161FPS/96FPS = 67.7% difference.
Comparing 3DMark Fire Strike: 20903/12495 = 67.3% difference.

Let's say the 980Ti has a 3% clock advantage (1492/1443=1.033, i.e. 3.3% difference) and that performance scaling with clock is linear. That still means on average over these 4 benchmarks the 980Ti still leads the 970 by 61.8%.

At the very minimum you should expect a 50% increase when both sides are clocked equally, because the 980Ti has 50% more ROPs, 50% more bandwidth, 69% more CUDA cores, and 50% more VRAM. If not then there's likely a bottleneck with the CPU.

: ( so much wasted breath

It's not like you and I are invalidating classical physics, it doesn't take a PhD to reach the same conclusion we have, I just don't understand the reasoning at all...

Like, what if techpowerup's suite contains a bunch of cpu bound games that are old?

How does a gpu with 50+% more everything *not* perform 50%+ faster clock for clock? Lol

Even if you call it just 50% faster, its a great deal, I bought mine when a similar (gigabyte g1) 970 was for 350 new. I paid 610 for the 980ti G1, that's 75% more expensive for 50%+ performance.

Notbad.jpg

Someone once said "winning an argument on the internet is like winning the special Olympics. even if you are the winner, you are still a retard".

The paralympics are no laughing matter, and those people are impaired physically afaik, not mentally

 

[SpeedyVV]

: ( so much wasted breath

The paralympics are no laughing matter, and those people are impaired physically afaik, not mentally

Paralympians are not mentally ill, arguing on the internet is.

 

[Alkaladur]

Paralympians are not mentally ill, arguing on the internet is.

You're probably right, been arguing with a bright young mind who claimed the 290x has outperformed every Titan from the original to the X.

I know he's wrong.

I know I can present data to prove. Him wrong

I know he'll never concede he is wrong

So I'm dumb for trying really

 

[Daniel_Chang]

Sigh...

Here...

970 Strix overclocked to 1443/8002.
ASUS GeForce GTX 970 Strix review

980Ti Strix overclocked to 1492/8000.
ASUS GeForce GTX 980 Ti STRIX Review

Comparing Bioshock Infinite at 1440p: 134FPS/86FPS = 59.5% difference.
Comparing Hitman Absolution at 1440p: 101FPS/63FPS = 60.3% difference.
Comparing Tomb Raider at 1440: 161FPS/96FPS = 67.7% difference.
Comparing 3DMark Fire Strike: 20903/12495 = 67.3% difference.

Let's say the 980Ti has a 3% clock advantage (1492/1443=1.033, i.e. 3.3% difference) and that performance scaling with clock is linear. That still means on average over these 4 benchmarks the 980Ti still leads the 970 by 61.8%.

At the very minimum you should expect a 50% increase when both sides are clocked equally, because the 980Ti has 50% more ROPs, 50% more bandwidth, 69% more CUDA cores, and 50% more VRAM. If not then there's likely a bottleneck with the CPU.

Data is not valid for the following reasons:

-Tested 1 year apart on different driver revisions. Newer drivers will change the delta. Both cards need to be tested on the same driver.
-Tested on different systems. One uses an older Sandy Bridge-E w/8GB DDR3. The other uses a newer Haswell-E witrh 16GB DDR4. The deck was clearly stacked in your comparison.
-You cherry picked a few games out of the entire comparison.

Show me an apples to applies comparison (same test suite, same hardware/software, same percentage overclock) across numerous games and I'll concede. So far the only test that matches this data is the one I've presented where both cards were at reference, in the same system, with the same hardware/software, and across 21 games. That is how you properly test. Those results were 35-45% faster for the 980ti over the 970 depending on resolution (1080p, 1440p, or 2160p).

 

[Daniel_Chang]

You're probably right, been arguing with a bright young mind who claimed the 290x has outperformed every Titan from the original to the X.

I know he's wrong.

I know I can present data to prove. Him wrong

I know he'll never concede he is wrong

So I'm dumb for trying really

This sounds like my debate with you.

I've shown you data to prove you wrong. Your own data proves you wrong (after I do the numbers, since you've admitted you won't crunch the numbers). You'll never concede that you're wrong. And to top it off, you then went into insults because you couldn't admit to being wrong. This is why our debate ended.

 

[Araxie]

Data is not valid for the following reasons:

-Tested 1 year apart on different driver revisions. Newer drivers will change the delta. Both cards need to be tested on the same driver.
-Tested on different systems. One uses an older Sandy Bridge-E w/8GB DDR3. The other uses a newer Haswell-E witrh 16GB DDR4. The deck was clearly stacked in your comparison.
-You cherry picked a few games out of the entire comparison.

Show me an apples to applies comparison (same test suite, same hardware/software, same percentage overclock) across numerous games and I'll concede. So far the only test that matches this data is the one I've presented where both cards were at reference, in the same system, with the same hardware/software, and across 21 games. That is how you properly test. Those results were 35-45% faster for the 980ti over the 970 depending on resolution (1080p, 1440p, or 2160p).

Agreed... only under same conditions can direct comparisons be made..

 

[Alkaladur]

This sounds like my debate with you.

I've shown you data to prove you wrong. Your own data proves you wrong (after I do the numbers, since you've admitted you won't crunch the numbers). You'll never concede that you're wrong. And to top it off, you then went into insults because you couldn't admit to being wrong. This is why our debate ended.

No, my debate with you revolves around the fact that you're looking at situations in which the gpu is not being fully stressed, if the delta between 980 and 980ti is 30% clock for clock, you're clearly not stressing the 50% + everything on the chip

Because this isn't irritating right ? You're being arrogant, and you're simply choosing to ignore everything that has been posted, both by me and by others who took the time to try and find reviews comparing the cards at similar clocks. I have no incentive to contribute to the discussion really, it's not like it affects me whether you buy a 980ti, or a x70, or polaris, or anything really. Like everyone else i'm just replying to your posts

I'm the last person you can accuse of not understand the number crunching, or not doing it, if you had taken the time and effort to think through exactly what you're trying to evaluate you would look at the data that best reflects situations in which gpu performance is best being tested
In the future, more games will show the 50%+ advantage, because it will be more commonplace for the software to saturate the gpu

AotS doesn't stress VRAM, so it's a great one, compared 970 to 980ti in that

Reference cards here

5.3 tflops for the 980, meaning it must be running at..... 5300/(2048x2)=1293mhz

6.4tflops on the 980ti, 6400/(2816x2)= 1136mhz

Have a good day

 

[rinaldo00]

At the very minimum you should expect a 50% increase when both sides are clocked equally, because the 980Ti has 50% more ROPs, 50% more bandwidth, 69% more CUDA cores, and 50% more VRAM. If not then there's likely a bottleneck with the CPU.

Except theoretical gains do not always translate to real world gains, or 2 cards in SLI would give twice the performance.

 

[rinaldo00]

The paralympics are no laughing matter, and those people are impaired physically afaik, not mentally

paralympics is physically handicapped, special Olympics (what he said in his first post) is mentally handicapped.

 

[Alkaladur]

Except theoretical gains do not always translate to real world gains, or 2 cards in SLI would give twice the performance.

That is true but SLI is not the same as considering a single gpu 50% larger

 

[rinaldo00]

That is true but SLI is not the same as considering a single gpu 50% larger

Still, you can't just multiply the increase in 'parts' and say without testing that you will get the same increase in performance.

 

[Alkaladur]

Still, you can't just multiply the increase in 'parts' and say without testing that you will get the same increase in performance.

I didn't. It just so happens that to demonstrate the advantage you need to stress it sufficiently, and not many games will do that at 1080p, 1440p+ it becomes much more apparent

 

[NKD]

When did this thread turn in to 970 vs 980ti. Lol

 

[SpeedyVV]

paralympics is physically handicapped, special Olympics (what he said in his first post) is mentally handicapped.

Uhmm, this makes me think about something!

Maybe we could have like a SUPER-Olympics. Para-Olympians vs Special-Olympians.

I wonder who would win?

Sort of like AMD vs NVidia.

 

[ChosenUno]

Data is not valid for the following reasons:

-Tested 1 year apart on different driver revisions. Newer drivers will change the delta. Both cards need to be tested on the same driver.
-Tested on different systems. One uses an older Sandy Bridge-E w/8GB DDR3. The other uses a newer Haswell-E witrh 16GB DDR4. The deck was clearly stacked in your comparison.
-You cherry picked a few games out of the entire comparison.

Show me an apples to applies comparison (same test suite, same hardware/software, same percentage overclock) across numerous games and I'll concede. So far the only test that matches this data is the one I've presented where both cards were at reference, in the same system, with the same hardware/software, and across 21 games. That is how you properly test. Those results were 35-45% faster for the 980ti over the 970 depending on resolution (1080p, 1440p, or 2160p).

NVIDIA GeForce GTX 980 Ti Video Card Roundup – ASUS, MSI and Zotac - Page 6 of 13 - Legit Reviews

I'm gonna stick with 1440p results since the 970 can hit VRAM limits in 4K. Will also be comparing results between 980Ti Gaming 6G and 970! Amp Omega.

BF4: 94.5/58.5=61.5%

FO4: 93.7/59.1=58.5%

GTAV: 87.9/58.6=50%

SW:Battlefront: 115.3/75=53.7%

3DMark FS: 18122/11566=56.7%

3DMark FS Ex: 8439/5432=55.4%

PNY GeForce GTX 980 Ti Review: Page 5 of 7
This review compares a stock 980Ti against a Gaming G1 970, one of the more aggressive SKUs for the 970 out there. It boosts to 1329MHz (SOURCE) compared to 1202MHz of the stock 980Ti (SOURCE).

3DMark FS: 17380/11488=51.3%

Heaven: 2618/1438=82.1%

Bioshock Infinite 1080p: 161/94=71.2%
Bioshock Infinite 1440p: 105/76=38.2%

GTAV 1080p: 148/102=45%
GTAV 1440p: 108/70=54.3%

MetroLL 1080p: 119/59=101%
MetroLL 1440p: 79/53=49.1%

Thief 1080p: 110/49=124.4%
Thief 1440p: 75/51=47.1%

TR 1080p: 213/111=91.9%
TR 1440p: 128/78=64.1%

 

[KazeoHin]

Welcome to AMD's marketing department: Where even when the topic is about an AMD product on an AMD forum, we are still talking exclusively about Nvidia products!

 

[Maddness]

Welcome to AMD's marketing department: Where even when the topic is about an AMD product on an AMD forum, we are still talking exclusively about Nvidia products!

Yeah. It's a bit of a sad state of affairs. I guess everyone is excited to get some news on next gen cards. So they are just trying to pass the time talking about anything.

 

[Daniel_Chang]

NVIDIA GeForce GTX 980 Ti Video Card Roundup – ASUS, MSI and Zotac - Page 6 of 13 - Legit Reviews


PNY GeForce GTX 980 Ti Review: Page 5 of 7
This review compares a stock 980Ti against a Gaming G1 970, one of the more aggressive SKUs for the 970 out there. It boosts to 1329MHz (SOURCE) compared to 1202MHz of the stock 980Ti (SOURCE).

Your first link compared cards with the following core overclocks:
GTX 970 - 4.95%
GTX 980ti - 11.78%, 12.16%, and 12.53%

So, that one had the deck stacked in favor of the 980ti, though not as bad as all previous posts. Your second link, however, was interesting. It's not valid data because the deck was again stacked, but this time against the 980ti (I'm trying to be fair). So it's enlightening to see such large deltas. They got some very interesting results. Here are my observations on that one:

1. The GTX 970 scales from 1080p to 1440p better than the 980ti in every single game except GTA V. That's a red flag.
2. The GTX 970 performs comparable or better in 2 games at 1440p over 1080p. Doesn't that seem odd?
3. These numbers conflict with the numbers from TechPowerup in the few games that were benchmarked by both sites.

So, in your opinion, which tests do you think are more valid? The Techpowerup tests that show performance scaling consistent with other reviews from other websites, or this Proclockers site with performance scaling that doesn't line up with anything? Either way, kudos for being the first to find tests that shows a 980ti having consistently >50% performance over a 970 without resorting to the 980ti being ovcerclocked to a higher degree than the 970.

 

[Daniel_Chang]

Welcome to AMD's marketing department: Where even when the topic is about an AMD product on an AMD forum, we are still talking exclusively about Nvidia products!

Yeah. It's a bit of a sad state of affairs. I guess everyone is excited to get some news on next gen cards. So they are just trying to pass the time talking about anything.

Sorry about that guys. I'm to blame there. With a rumored $300 Polaris comparable to a GTX 980ti, I was using the 980ti as a benchmark to determine if an upgrade from my 970 to this rumored Polaris was worthwhile. I decided "not likely" based on some hard data. And since then, we've had a user contesting my data (but not providing any of his own to counter it). The last page has one user actually providing real data, and that's why the discussion is continuing. It's always enlightening to get the point of view from others.

 

[harmattan]

So very glad I have alerts turned on for this Polaris thread

 

[razor1]

The plot thinkens normally I don't post fudzilla stuff, but the pricing seems right.

AMD Polaris 10 has 390/390X performance

 

[Pieter3dnow]

The plot thinkens normally I don't post fudzilla stuff, but the pricing seems right.

AMD Polaris 10 has 390/390X performance

You posted an article that I did't realize was new
But still nothing in there worth mentioning
The only thing that seems plausible is the performance bracket ...

 

[schizo]

That pricing would be disastrous. Next-generation, same performance, exact same price. Oh, and it runs cooler and is more power-efficient-- but who really gives a crap about that on the desktop?

A very low-power GPU with 390x-level performance would be interesting for laptops, but that assumes GP106 doesn't compete in that space. Nah. Prices have to come down.

 

[razor1]

You posted an article that I did't realize was new
But still nothing in there worth mentioning
The only thing that seems plausible is the performance bracket ...

May 3rd dude , anycase the performance crap doesn't matter too much.