Nvidia's GTX680 gets thrashed by AMD's mid-range Radeon HD 7870 in GPU compute Sourc

[Goodman]

And the Inq is a tabloid, so I don't think it's a reputable source.

That & S/A they both make BS articles they are only good as a story telling & never as fact

People take these two (The Inquirer & SemiAccurate) as fact have serious issues IMO

 

[Crap Daddy]

It's interesting that this thread gets more attention than the one who announced Tom's hardware review leaks...

 

[VulkanBros]

It's interesting that this thread gets more attention than the one who announced Tom's hardware review leaks...

Not really........flies like cowshit......

 

[Sinzia]

Im waiting for the actual reviews, then I'll decide.... one of my GTX 470's is dying.

 

[Stephen.]

TL : DR

Weren't you trashing the 680 in the Gigabyte thread saying that if it didn't support DX 11.1 it would be a worthless card, and then saying you only wanted the card for photo editing? Now is compute your main concern?

I don't really understand the need to trash an unreleased card, besides, even if you prefer AMD cards, if the 680 ends up being faster than the 7970 and retails for $499 as the latest rumors suggest, that means you'll be able to get a 7970 cheaper, isn't that good for you, and for all costumers in general?

Oh well i just want what's best all round if you haven't noticed yet. And as per the DX 11.1 it was confirmed in the thread that it supports it so +1 for that so the GTX680 and HD7979 have almost a similar chance from me of getting bought, then i found this news on overclock.net.
And so you know i currently own a Point Of View TGT Tuned GTX460 Charged 1GB, so i am not trashing Nvidia for your information, i am trying to get clarifications as to why there are so much - things about Kepler coz Fermi did not have this much -things about it, AMD Cards seem to get less criticism recently, which has to account for something, Nvidia used to be where AMD Radeon are now.

And again to clarify i have used cards from both camps

 

[Kreij]

things about Kepler coz Fermi did not have this much -things about it,

You weren't paying attention. The original relase of Fermi got trashed from every point on the globe.

Just wait for the real reviews of the 680 and then your questions will be answered.

 

[phanbuey]

you weren't paying attention. The original relase of fermi got trashed from every point on the globe.

Just wait for the real reviews of the 680 and then your questions will be answered.

iiaahhh cannttt waiittttt.... Eeeeeeeeee

 

[Stephen.]

GK110 Specifications Approximated, why the hefty price for GTX680 GK104, it will make the GK110 way more expensive. Or if GK110 comes out and is priced at 500/550, it will make people who bought GTX680 feel reaped off...

 

[phanbuey]

GK110 Specifications Approximated, why the hefty price for GTX680 GK104, it will make the GK110 way more expensive. Or if GK110 comes out and is priced at 500/550, it will make people who bought GTX680 feel reaped off...

how do you think the people that just bought the 7970 will feel in about 4 days?

 

[Stephen.]

how do you think the people that just bought the 7970 will feel in about 4 days?

They'll feel bad, but when the GK110 comes out they'll feel very terribly bad , if there is anything like that.

 

[Benetanegia]

GK110 Specifications Approximated, why the hefty price for GTX680 GK104, it will make the GK110 way more expensive. Or if GK110 comes out and is priced at 500/550, it will make people who bought GTX680 feel reaped off...

Latest rumors say that Nvidia priced the GTX680 at $499 and that only means one thing, they are in for a price war, because being 15% faster (according to leaked review) matching the price of HD7970 would have been enough to compete. Pricing it lower suggests they want to fight and don't care about average price dropping a little over the time. Say in 5 months it goes down to $350-$400), they'd still be making a lot of money off them, and they would have a place for GK110 in the $500-$600 range.

Dual GK104 is going to be IMO, much faster than previous dual-GPU cards in relation to their single GPU cards, so no overlapping between GK100 and dual-GK104. Far lower TDP + dynamic clocks (it becomes 10x more important in a dual card with its "mixed" GPU usage) will make it use similar clocks as the single GPU, instead of ++25% lower clocks that have been used in previous iterations. Full chip enabled too. I don't see how the dual card (GTX690? 695?) could be any less than 80% faster than the single GPU card (GTX680) and based on how previous dual cards scaled, despite their lower clocks and crippled core, 90% seems achievable with GK104. And I seriously doubt GK110 will be more than 50% faster than GK104 in games. So probably GTX670 (+25%) GTX680 (+25%) GTX770 (+25%) GTX780 (+25%) GTX690

And everything is a moot point anyway. In 5-6 months 90% of the people who bought GTX680 for $500 (or $550 HD7970) at launch, will just be looking for another $500 card and if it offers 40% more performance they will most probably be glad to pay it.

 

[OneMoar]

why are you ............... reading the inquirer . lol...........

 

[Stephen.]

What is clear at this time though is that NVIDIA is pitching GTX 680 specifically for consumer graphics while downplaying compute, which says a lot right there. Given their call for efficiency and how some of Fermi’s compute capabilities were already stripped for GF114, this does read like an attempt to further strip compute capabilities from their consumer GPUs in order to boost efficiency. Amusingly, whereas AMD seems to have moved closer to Fermi with GCN by adding compute performance, NVIDIA seems to have moved closer to Cayman with Kepler by taking it away.

With that said, in discussing Kepler with NVIDIA’s Jonah Alben, one thing that was made clear is that NVIDIA does consider this the better way to go. They’re pleased with the performance and efficiency they’re getting out of software scheduling, going so far to say that had they known what they know now about software versus hardware scheduling, they would have done Fermi differently. But whether this only applies to consumer GPUs or if it will apply to Big Kepler too remains to be seen.

http://www.anandtech.com/show/5699/nvidia-geforce-gtx-680-review/3

SO IT'S TRUE GTX680 (GK104) IS NOT THAT GOOD IN COMPUTE COMPARED TO AMD's GCN.

 

[phanbuey]

What is clear at this time though is that NVIDIA is pitching GTX 680 specifically for consumer graphics while downplaying compute, which says a lot right there. Given their call for efficiency and how some of Fermi’s compute capabilities were already stripped for GF114, this does read like an attempt to further strip compute capabilities from their consumer GPUs in order to boost efficiency. Amusingly, whereas AMD seems to have moved closer to Fermi with GCN by adding compute performance, NVIDIA seems to have moved closer to Cayman with Kepler by taking it away.

With that said, in discussing Kepler with NVIDIA’s Jonah Alben, one thing that was made clear is that NVIDIA does consider this the better way to go. They’re pleased with the performance and efficiency they’re getting out of software scheduling, going so far to say that had they known what they know now about software versus hardware scheduling, they would have done Fermi differently. But whether this only applies to consumer GPUs or if it will apply to Big Kepler too remains to be seen.

http://www.anandtech.com/show/5699/nvidia-geforce-gtx-680-review/3

SO IT'S TRUE GTX680 (GK104) IS NOT THAT GOOD IN COMPUTE COMPARED TO AMD's GCN.

so what you're saying is that you're basing this on one benchmark...

 

[xenocide]

SO IT'S TRUE GTX680 (GK104) IS NOT THAT GOOD IN COMPUTE COMPARED TO AMD's GCN.

I'll take it in exchange for higher frame rates, less noise, lower temperatures, and lower power draw.

 

[Stephen.]

so what you're saying is that you're basing this on one benchmark...

Doesn't matter whichever benchmark, Nvidia has taken that rout on GTX680 (GK104), they have essentially sacrificed GPU compute in the GK104 in favor of Graphics(Gaming Mostly), so basically i expect any other benchmark for GPU compute to be relatively similar.

 

[Mussels]

the claim he makes about AMD gaining 30% (20% OC + 10% drivers) is lols, w1zz has proven that doesnt really happen these days.

 

[trickson]

This is just pure bullshit! W1zzard did a review and proves that it is!
http://www.techpowerup.com/reviews/NVIDIA/GeForce_GTX_680/32.html

Of course the 7870 is not listed but hell going off what I see in that review I think that is will not be getting Thrashed!

 

[Stephen.]

This is just pure bullshit! W1zzard did a review and proves that it is!
http://www.techpowerup.com/reviews/NVIDIA/GeForce_GTX_680/32.html

Checkout this:
http://www.anandtech.com/show/5699/nvidia-geforce-gtx-680-review/17

does well in only 1 benchmark out of 3, the card is not good at compute 1 out of 3 is not good enough.

 

[trickson]

Checkout this:
http://www.anandtech.com/show/5699/nvidia-geforce-gtx-680-review/17
does well in only 1 benchmark out of 3, the card is not good at compute 1 out of 3 is not good enough.

Well there it is there then. But do not say this too loud some might take that one BM as the HOLY GRAIL! Don't get me wrong here I love ATI cards (AMD now) but the evidence is clear. Some one must be getting payed!

 

[OneMoar]

Well there it is there then. But do not say this too loud some might take that one BM as the HOLY GRAIL! Don't get me wrong here I am an love ATI cards (AMD now) but the evidence is clear. Some one must be getting payed!

I smell a fan boi .......

 

[trickson]

I smell a fan boi .......

Of ATI? OR Nvidia? No on both accounts. I am a fanboy of performance. If you look at my specs you will see I use both AMD and Intel products. Though they are old and a bit dated they serve me very well!
The reason I said what I did is because there are some ULTRA fanboys here and they will take that ONE BM and and say that ATI is the best there is. Then beat the shit out of you and call you names and get all pissed off and well every thing goes to hell. I think that time will tell just how well they really are once W1zzard gets a hold of it!

 

[Stephen.]

http://www.anandtech.com/show/5699/nvidia-geforce-gtx-680-review/17

As always our final set of benchmarks is a look at compute performance. As we mentioned in our discussion on the Kepler architecture, GK104’s improvements seem to be compute neutral at best, and harmful to compute performance at worst. NVIDIA has made it clear that they are focusing first and foremost on gaming performance with GTX 680, and in the process are deemphasizing compute performance

 

[trickson]

http://www.anandtech.com/show/5699/nvidia-geforce-gtx-680-review/17

As always our final set of benchmarks is a look at compute performance. As we mentioned in our discussion on the Kepler architecture, GK104’s improvements seem to be compute neutral at best, and harmful to compute performance at worst. NVIDIA has made it clear that they are focusing first and foremost on gaming performance with GTX 680, and in the process are deemphasizing compute performance

This would seem the case from what I have seen. And since most here are gamers that would mean it is not getting thrashed, Right?

 

[Stephen.]

This would seem the case from what I have seen. And since most here are gamers that would mean it is not getting thrashed, Right?

The post title reads Nvidia's GTX680 gets thrashed by AMD's mid-range Radeon HD 7870 in GPU compute


In GPU compute yes its getting thrashed, and just for the record i am not a fan boy either. i have used both cards ever since i started purchasing GPU, and currently using GTX460 since 2010.