[Hardwarecanucks] GTX 780 Ti vs R9 290X; The Rematch

[raghu78]

It's even worse if HWC thinks they run the test with a custom R290X at 1Ghz without throttling, because their original reference review was with a REAL reference, which throttled even on Uber. So they are basically saying, back then, the 780Ti was ahead by 7% vs a throttling reference, and now its 6% versus a NON throttling simulated R290X.

Which is even MORE BS because a real reference R290X is now as fast or faster than the 780Ti. Things like the Sapphire Tri-X, as shown by computerbase.de are about 10% above reference R290X numbers.

Again, I am not questioning whether NV neglect, or shift their focus, that isn't my agenda (surprising to some!). My only agenda here is correcting such a BS site spreading pure fud.

They manage to turn games where R290X normally leads the 780Ti by far, into results that have the 780Ti in front. Contrary to every other reputable site. That should say a lot about their ethics.

SKYMTL's main purpose with this article was to defend Nvidia. He would not write an article which criticized Nvidia. He is pretty much written that article to do damage control among Kepler users and literally PR for Nvidia. Any article from that shill about Nvidia has a foregone conclusion. Its written to praise, worship, adore, eulogize or defend Nvidia.

Heck TPU which has a slight Nvidia bias is showing that 780 Ti lost performance against R9 290X / GTX 970. computerbase which is generally neutral shows a significant fall for 780 Ti against Maxwell / GCN. The US tech press is now paid and bought for and all they care about is their ad revenue, free launch events passes, free hardware etc. Pathetic.

 

[AnandThenMan]

SKYMTL's main purpose with this article was to defend Nvidia.

Seems like an awful coincidence. Kepler has been seeing sharp criticism for falling behind relative to the AMD products sold in the same time frame, now we see an article saying hey no worries Kepler is fine.

Can't help but feel the article is damage control.

 

[iiiankiii]

SKYMTL's main purpose with this article was to defend Nvidia. He would not write an article which criticized Nvidia. He is pretty much written that article to do damage control among Kepler users and literally PR for Nvidia. Any article from that shill about Nvidia has a foregone conclusion. Its written to praise, worship, adore, eulogize or defend Nvidia.

Heck TPU which has a slight Nvidia bias is showing that 780 Ti lost performance against R9 290X / GTX 970. computerbase which is generally neutral shows a significant fall for 780 Ti against Maxwell / GCN. The US tech press is now paid and bought for and all they care about is their ad revenue, free launch events passes, free hardware etc. Pathetic.

Remember, HWC was one of the only FEW sites that got a Titan Z for review from Nvidia. In that review, it had the Titan Z beating the 295x2...... which went against the majority of the conclusion from professional reviewers. Their results can be odd at times.... I can only wonder why...

 

[selni]

Seems like an awful coincidence. Kepler has been seeing sharp criticism for falling behind relative to the AMD products sold in the same time frame, now we see an article saying hey no worries Kepler is fine.

Can't help but feel the article is damage control.

No it's not a coincidence a site looked at something people care about, but that it's an article aimed at something topical is hardly suspicious.

Besides that, does anyone disagree with what they're actually showing? Ie Kepler not losing absolute performance or much relative performance vs GCN in a bunch of mostly older games and Maxwell vs Kepler improving a bit since release?

The other complaint about Kepler (worse relative performance in some newer titles compared to Maxwell/GCN) isn't addressed sure... but while this should be simple enough to show, it's harder to show why without going into more detail than I'd expect from a tech site.

 

[Gikaseixas]

Is that really fair? Couldn't it just be that Nvidia gets the most out of their cards they can get near launch and so there isn't much room left for improvement? Not every card has unlimited potential locked away by driver inefficiencies. I would bet when we get something like the Witcher driver it is less about squeezing more out of the hardware and more about squeezing down the image quality in some way we can't easily see to placate the mob.

Historically that's not the case, I've had many Nvidia cards in the past and I used to see pretty good improvement thru driver updates but to be fair you could be right.

 

[Ranulf]

Semi OT but Linus Tech Tips on the gtx 480 and drivers since 2010:

https://www.youtube.com/watch?v=iZUshOSWQRo

 

[DarkKnightDude]

As a former owner of a Kepler GPU, it was obvious to me that was underperforming in newer games over the last year.

 

[Gikaseixas]

There are just a million reasons why the 290x flopped, but people will continue to make excuses for it rather than actually look at what happened

A million reasons really?
The crappy cooler was the only reason it failed. It allowed higher than optimal temperatures, therefore limiting projected clock speeds.

If AMD had launch it with a proper cooler (one just like the OG Titan) it would avoid these kind of exaggerations people are making today.


Before I get accusations of being a AMD fanboy:
I had a 390x on my hands a few days back but it's performance wasn't enough for me so I just got a GTX 980ti since it's simply the best GPU in the market today. This baby gets here today, can't wait to put my hands on it

 

[boozzer]

Hey man, I tried to give AMD the benefit of the doubt and waited but as soon as I realized that the Fury X was not going to be up to par I bought the GTX 980 Ti. I need gaming performance at my resolution (2560x1600) and my future monitor is going to be 3440x1440. I can't sit around waiting for AMD to get their crap together.

that was more of a response to the other poster instead of you.

nothing wrong with buying for what you want/need. :thumbsup:

 

[3DVagabond]

So the GCN R9 290X got better over time but the Kepler GTX-780Ti didnt ??? If that is what you are saying then that is exactly what we are talking about

Also HWC tested vs. launch drivers. Those would have to be the slowest drivers for any card.

 

[hawtdawg]

As a former owner of a Kepler GPU, it was obvious to me that was underperforming in newer games over the last year.

Same here. One of the most telling things that happened was with Witcher 3. Enough people complained, and Nvidia magically pulled 10% performance out of their ass in less than a week.

The real study that needs to happen, is someone comparing pre-game ready drivers to post-game ready drivers, not to see if they lost performance, but to see if Nvidia did anything at all to give them any performance.

 

[thesmokingman]

Also HWC tested vs. launch drivers. Those would have to be the slowest drivers for any card.

Another trick up the shill reviewers sleeve. There are so many ways to skew the numbers to fit an agenda.

 

[dogen1]

Unless you ran said hardware yourself and found the results to be accurate, then when you post your findings, are subsequently called out as biased anyway. But I digress..

I've seen that in terms of compute, Kepler seems to be at a disadvantage compared to Maxwell, GCN, and even Fermi. There are probably some deficiencies in how the shader cores are utilized that was fixed in hardware via Maxwell.

Yeah. IIRC, Kepler is more dependent on the driver to run at peak performance than GCN, Fermi, and probably Maxwell(not 100% sure).

 

[RampantAndroid]

Seems like an awful coincidence. Kepler has been seeing sharp criticism for falling behind relative to the AMD products sold in the same time frame, now we see an article saying hey no worries Kepler is fine.

Can't help but feel the article is damage control.

So....wait. The conspiracy theory for Kepler "getting worse" has been "NVidia wants all their NVidiots to buy Maxwell, so they make Kepler worse!"...but suddenly "No, it's not that. NVidia is just a bunch of idiots ignoring Kepler and they need some reviewer to defend them!"

Huh? Which is it?

 

[AnandThenMan]

Huh? Which is it?

Neither, nothing to see here everything is fine. Or maybe there is nuance to the situation. Did you ever consider Nvidia realized that throwing Kepler owners under the bus was not a good idea and they were afraid of the reputation hit they were taking? So they decided to do some damage control.

 

[Red Hawk]

So....wait. The conspiracy theory for Kepler "getting worse" has been "NVidia wants all their NVidiots to buy Maxwell, so they make Kepler worse!"...but suddenly "No, it's not that. NVidia is just a bunch of idiots ignoring Kepler and they need some reviewer to defend them!"

Huh? Which is it?

I don't think there's any big conspiracy behind Nvidia not optimizing for Kepler. I think to a certain degree Nvidia is aiming new graphics features to work best on Maxwell. But I don't think that necessarily means they're neglecting Kepler, or even Fermi; did you know that DSR has been extended to support Fermi graphics cards? You can make DSR work on a GTX 480. Try getting VSR to work on a Radeon HD 5870. If Nvidia's "neglect" of Kepler is bad, then AMD's neglect of anything Terascale-based is worse!

On a separate note, from what I can tell this HardwareCanucks review seems obviously biased and I'm not inclined to give it any weight.

 

[96Firebird]

People complain that Kepler is being neglected, so a review site decides to test that theory. Now that they have published the article, it is being called "damage control" by AMD fans. Shocker!

 

[Abwx]

People complain that Kepler is being neglected, so a review site decides to test that theory. Now that they have published the article, it is being called "damage control" by AMD fans. Shocker!

No comment of yours about the numbers in this review, instead you are relying to an ad hominem, i guess that you know that you would be hard pressed to prove that this review is not biaised, hence your off topic comment.

 

[Silverforce11]

People complain that Kepler is being neglected, so a review site decides to test that theory. Now that they have published the article (that has fake data compared to every other reputable site), it is being called "damage control" by AMD fans. Shocker!

fixed.

 

[96Firebird]

Unlike you guys, I don't blame everything I don't agree with as biased.

Continue believing the whole world is out to get you and anything that doesn't conform to your beliefs is wrong...

 

[Silverforce11]

So quick to resort to ad hominem!

It's not that I don't agree with their results. Every other major site also do not agree with their results.

If it was peer reviewed publication, such an article would have been thrown in the trash.

 

[tential]

A million reasons really?
The crappy cooler was the only reason it failed. It allowed higher than optimal temperatures, therefore limiting projected clock speeds.

If AMD had launch it with a proper cooler (one just like the OG Titan) it would avoid these kind of exaggerations people are making today.


Before I get accusations of being a AMD fanboy:
I had a 390x on my hands a few days back but it's performance wasn't enough for me so I just got a GTX 980ti since it's simply the best GPU in the market today. This baby gets here today, can't wait to put my hands on it

I guess you're going to ignore miners, power consumption, long term customer cost acquisition vs short term customer cost acquisition, customer price perception, used market flood of miner cards, over production due to miners of the R9 200 series and as a result a massive overstock of cards they couldn't sell due to the used market, and other things as to why the R9 290x failed.

I don't think you're an AMD fanboy... I just think like many people, you aren't looking at the big picture and only think about the things you care about and not everything possible in the process of running a business to make money.

 

[Goatsecks]

So quick to resort to ad hominem!

It's not that I don't agree with their results. Every other major site also do not agree with their results.

If it was peer reviewed publication, such an article would have been thrown in the trash.

Has any one tried replicating his results then?

That would be pretty conclusive.