GK104 running about 10-20% slower than Tahiti.

I can't seem to find that quote in the actual article, just someone commenting saying they read it in the article?

As for the actual article it doesn't sound all bad, if a $300 dollar card can even compete with a 7970 even in games it doesnt have an "advantage" in, that sounds like a win to me.
 
I think your thread title is your way of justifying your 7970 purchase. I bought one as well but I have to admit the 10-20% slower than Tahiti number is only part of the article. Unfortunatley the article doesn't state which Tahitit it will be 10-20% slower than, though if it is competing with Pitcairn in some benches then it alsmost certainly isn't the 7970. AMD have 3x Tahiti cards planned, 7970, 7950 and 7890, if it is 10-20% slower than the 7890 or even the 7950 then it will be slower than a GTX 580. If these numbers are correct then it couold possibly be at GTX 570 speeds unless it is a hardware accelerated PhysX title. The article clearly shows that when GK104 is used for hardware PhysX it will be much faster. The problem is that hardware accelerated PhysX games are a very rare breed.

http://en.wikipedia.org/wiki/PhysX

Reading what Charlie said I am glad I didn't wait for Kepler myself as it seems GK104 is only likely to be faster when it is used for hardware accelerated PhysX titles. For all other titles it looks likely to fit right in below Tahiti at the Pitcairn level. I have not played one single hardware accelerated PhysX title and thought "wow". I might have thought, nice, or not bad when I played Batman AA on my SLI GTX 460s but certainly not a big selling point IMHO.

If this article is correct and GK104 will be slower than Tahiti, or even Pitcairn in some cases then it fully explains the $300 price point. Reality is almost always far less spectacular than rumours in the world of the GPU enthusiast. Far too many people were eager to believe the $300 for HD 7970 beating performance rumour. Even I dared to hope the rumours were true but at the back of my mind I felt that at $300 it would come in below the GTX 580.
 
Last edited:
ICDP said:
I bought one as well but I have to admit the 10-20% slower than Tahiti number is only part of the article.
Click to expand...

At what point in the article does it even actually say this? If I search "benchmarks" on that page I cannot find it in the actual article.

EDIT: nv I found it.
 
Hmm, a $300 (mid-range) dollar part performing under a $550 (high-end) dollar part, odd indeed :p In the end, we'll just have to wait until release and compare dollar : performance of all available cards.
 
"In the same way that AMD’s Fusion chips count GPU FLOPS the same way they do CPU FLOPS in some marketing materials, Kepler’s 3TF won’t measure up close to AMD’s 3TF parts. Benchmarks for GK104 shown to SemiAccurate have the card running about 10-20% slower than Tahiti. On games that both heavily use physics related number crunching and have the code paths to do so on Kepler hardware, performance should seem to be well above what is expected from a generic 3TF card. That brings up the fundamental question of whether the card is really performing to that level?"


There it is.
 
What does this mean in the end? Is it cheating? Is it ethical? Is Kepler/GK104 going to be worth the money? Will it beat AMD’s 7970? These are all subjective decisions for you to make. What software will Nvidia show off as benchmarks to promote Kepler’s performance? That list is a little narrower. What will happen to sites that dare to test software that is not ‘legitimately accelerated’? No idea, but history offers some clues. One thing you can say for sure is that the information released prior to and with the card is unlikely to be the whole story. Legitimacy, performance, honesty, and ethics are unlikely to resemble the official talking points, and the whole truth is likely to be hidden from prying eyes for very partisan reasons. Big grains of salt around this one people, be very skeptical of everything you hear, and take nothing at face value.S|A
Click to expand...


LOL. Vindicated. I especially love the Arkham City patch after Tahiti is released and blows the doors off of GTX 580.
 
8du8 said:
I can't seem to find that quote in the actual article, just someone commenting saying they read it in the article?

As for the actual article it doesn't sound all bad, if a $300 dollar card can even compete with a 7970 even in games it doesnt have an "advantage" in, that sounds like a win to me.
Click to expand...

Here in lies the problem with ambiguous articles and statements/rumours. I take this 10-20% slower than Tahiti to mean it will have Pitcairn performance in non-PhysX titles. You take it to mean 10-20% slower than the 7970, but he didn't say 7970, he said Tahiti.

So what we have is news that GK104 will handilly beat the 7970 in some (rare) games but be beat by Pitcairn in others. :)
 
Yeah, I found it. Still a $300 vs a $550 one and its only 10-20% slower? Thats not bad at all, especially if it overclocks well.

Hopefully we wont have to wait too much longer to find out for sure how this thing is going to perform. Also would be nice to get more info on higher end Kepler cards.

EDIT: @ICDP Even still 7950 is hardly any cheaper than the 7970. As long as the gk104 OC's well it still sounds to me like $300 is a good value.
 
So Nvidia's mid-range card will compete with AMD's mid-range card in most games, and in ones that support the Physx API it will trounce the 7970 and GTX 580? I guess we'd all better start hoping for Physx adoption then.
 
Wasn't it said that the gk104 will be on par with the GTX 580 since it is midrange..not sure why this is surprising. I honestly found the rumors of it beating the HD 7970 to be really surprising, would be nice, but I am not expecting it to happen.
 
xIronCrossx said:
Wasn't it said that the gk104 will be on par with the GTX 580 since it is midrange..not sure why this is surprising. I honestly found the rumors of it beating the HD 7970 to be really surprising, would be nice, but I am not expecting it to happen.
Click to expand...

The article posted today at semiaccurate says:

GK104 will positively fly in physx enabled games (batman: AC)

GK104 will be slower than pitcairn in all other games (~around 6970 performance)

Widespread adoption of physx will be a tough sell since nearly 99% of AAA titles are multi platform. The wild card is overclocking, hopefully these things will oc like crazy. And really you can't complain since its 300$. I was initially disappointed but if it is indeed 300$, I say hell yeah. Would be a great card at that price.
 
Last edited:
Well, when has semi accurate been correct? (50% of the time maybe?)

I don't believe these rumors or the ones saying it will rofl pwn the HD 7xxx series. I just wanna see some damn reviews :(
 
xIronCrossx said:
Well, when has semi accurate been correct? (50% of the time maybe?)
Click to expand...
I don't understand why people rely on that site at all for anything concrete. The name of the site itself is "SEMI - ACCURATE." By definition, most of what he prints will have some grain of truth while the rest is pure bullshit.
 
I still think GK104 is going to be a great card for the money. I wouldn't believe all the rumors though. SA is very esoteric about this information and it leads me to question his sources.

Wish GK104 was here now.
 
Semiaccurate was pretty right on the money with Fermi though, but his sources often have biases or occasionally spew BS like claiming ATI's 2900XT will beat the 8800GTX.
 
Just thought I'd point out, as some of you seem to have missed it, that GK104 looks to be able to run not only Hardware accelerated Physx titles very well, but also games using the software PhysX API, of which there are over 100.

This could mean, according to Charlies article, that in over 100 games GK104 could be faster than 7950/70, but in the rest more like 7870 performance.

I could be wrong, but this quote from the article points to this imo:

One thing we can say is that the list of benchmarks shown off by Nvidia where Kepler has an overwhelming advantage all support PhysX. This is not to say that they are all hardware/GPU PhysX accelerated, they are not, most use the software API.

This is important because it strongly suggests that Nvidia is accelerating their own software APIs on Kepler without pointing it out explicitly. Since Kepler is a new card with new drivers, there is no foul play here, and it is a quite legitimate use of the available hardware.
Click to expand...

I dont believe this allows Nv to claim overall performance figures, however I have not yet seen anything from Nvidia claiming these either, only going from SA articles.
 
So much depends on definition of terms. If "10-20% less than Tahiti" means 7970, then the midrange Kepler is on par with a GTX580. That's pretty sweet in my book, and there's still the performance and flagship parts to come.
 
Commander Suzdal said:
So much depends on definition of terms. If "10-20% less than Tahiti" means 7970, then the midrange Kepler is on par with a GTX580. That's pretty sweet in my book, and there's still the performance and flagship parts to come.
Click to expand...

Indeed, so much is open to interpretation, he does state it will "probably lose to a Pitcairn in some games", so is he referring to low end Tahiti? I don't really know, either way the rumours that GK104 would murder the HD 7970 for $300 have always seemed far fetched to me, even though I did hope it was possible.

IMHO PhysX never really took on, the number of major AAA titles released over the past year with PhysX can be counted on one hand. If Nvidia are pouring more R&D into that dead (or dying) horse then I can't see it being a big selling point to be honest.
 
You must log in or register to reply here.
Back
Top