AMD GPU'14 Event Detailed, Announces Radeon R9 290X

[Prima.Vera]

Forgive the ignorance, but WHEN will the cards be available for testing and benching? I mean in the office before public release.

 

[the54thvoid]

I'm just amused that certain peeps are saying AMD's very own PR slide showing the Firestrike scores is unreliable. I stated I wanted the new card to beat Titan, so I could go and get a card I could tamper with more than my current castrated card. So to me, the things at first sneak peak is a disappointment, not a victory to Nvidia.

Folk really ought to stop reading posts in such combative ways. Only a dick would want AMD to fail with a new piece of tech and they would have no place to be on a tech forum like this.

 

[RCoon]

I'm just amused that certain peeps are saying AMD's very own PR slide showing the Firestrike scores is unreliable. I stated I wanted the new card to beat Titan, so I could go and get a card I could tamper with more than my current castrated card. So to me, the things at first sneak peak is a disappointment, not a victory to Nvidia.

Folk really ought to stop reading posts in such combative ways. Only a dick would want AMD to fail with a new piece of tech and they would have no place to be on a tech forum like this.

Seconded, I want R9 290X to be a viable alternative to all the voltage locked 780's/Titan's, so we can get back to serious overclocking on water without having to mod this and that so we dont hit arbitrary limits.

 

[HumanSmoke]

I'm just amused that certain peeps are saying AMD's very own PR slide showing the Firestrike scores is unreliable.

It's not a road any company would want to go down IMO. Short term strategic gain generally isn't offset by long term loss of goodwill

I stated I wanted the new card to beat Titan, so I could go and get a card I could tamper with more than my current castrated card. So to me, the things at first sneak peak is a disappointment, not a victory to Nvidia.

I wouldn't really even see it as a disappointment. If on the balance of benchmarks it trades blows with the GTX 780/Titan then that is well and good -depending upon price, because there are going to be apps (OpenCL etc) and games where the Radeon will be faster in any case. In the end it doesn't really matter what the performance over and above the opposition- or its own product stack- is unless the price is right. Price at $600 or $650 as has been mentioned in various places and it pretty much doesn't change anything ( Zotac and PNY are both in that price range from a quick Google) except getting the option of another feature set

Folk really ought to stop reading posts in such combative ways.

Seconded

Only a dick would want AMD to fail with a new piece of tech and they would have no place to be on a tech forum like this.

Seems more a case of one group using the available information to ascertain a possible/probable level of performance, and another group being a little defensive over the findings. I don't see a lot of people wishing AMD away...more a case tempering expectation with the information given by the vendor.

There's always been an irrational train of thought that the next architecture is going to be the one- the [insert huge percentage] quantum leap reminiscent of the ~1994-99 era- and it just ain't going to happen. Even if the technology breakthroughs were available, it is still two companies carving up a cake that used have fifty with their hands in it.

 

[BiggieShady]

I'm surprised no one is commenting on AMD's response to nVidia's "GeForce Experience" or whatever it's called.

So please, you people, spare me these comments of how useless this or that is, what is bloatware or not how this and that is a data mining conspiracy (which might very well be true, but seriously now, just unplug your internet and be done with it if you're that desperate). I'd like to read more valid opinions than that on what is supposed to be a very decent forum...

You have a funny notion of indecency.

I'm sorry you didn't like my comment. I wish you that all comments you ask for to be the ones you actually like ... let's cross our fingers and wish it really, really, hard .... if it doesn't work for some reason, look at the good side - at least you are not that surprised anymore.

 

[Xzibit]

I'm just amused that certain peeps are saying AMD's very own PR slide showing the Firestrike scores is unreliable.

I guess I was confused since everyone saying R9 280X = 7970 GHZ

Back in March Brandonwh64 posted this valid score (old drivers)

Unless i'm reading his data wrong hes using an original 7970 and boosting it to 7970 GHZ @ stock clocks the numbers just didn't match up to me.

That would mean the AMD PR slide is off by -400 to -600 with the R9 280X. Since it seamed they were that off to a supposed refresh. It raised an eyebrow.

Silly me. For not taking things at face value

 

[PopcornMachine]

Searched the thread with the term "crossfire" and found it really quick.

http://www.techradar.com/news/compu...u-the-radeon-r9-290x-1183950?src=rss&attr=all

That is speculation at this point. AMD said nothing about this.

Maybe true, but why wouldn't they say that at the big info press conference thing?

Seems more noteworthy than audio.

 

[NeoXF]

I guess I was confused since everyone saying R9 280X = 7970 GHZ

Back in March Brandonwh64 posted this valid score (old drivers)

http://img.techpowerup.org/130331/Capture027.jpg

Unless i'm reading his data wrong hes using an original 7970 and boosting it to 7970 GHZ @ stock clocks the numbers just didn't match up to me.

That would mean the AMD PR slide is off by -400 to -600 with the R9 280X. Since it seamed they were that off to a supposed refresh. It raised an eyebrow.

Silly me. For not taking things at face value

That's a 1050MHz, 1100MHz Turbo w/ stock 6000MHz VRAM HD 7970 GHz... so pretty much what we should expect from a 280X... sooo... yeah...
(ofc, you have to also take into account the 4,5GHz hexacore i7 platform it's running on)

And a 1000 point discrepancy is nothing to sneeze at... I imagine the R9-290X (a even more CPU-limited GPU) is being off-shot by even more in those claims.

 

[Nordic]

That is speculation at this point. AMD said nothing about this.

Maybe true, but why wouldn't they say that at the big info press conference thing?

Seems more noteworthy than audio.

I was simply answering your question of where he pulled that information from.

 

[PopcornMachine]

I was simply answering your question of where he pulled that information from.

Thanks for the info. I heard something different. Turns out AMD hasn't really said yet.

If they'd been more clear on important stuff, it would have saved all of us a lot trouble.

 

[TheoneandonlyMrK]

Seems more a case of one group using the available information to ascertain a possible/probable level of performance, and another group being a little defensive over the findings. I don't see a lot of people wishing AMD away...more a case tempering expectation with the information given by the vendor.

There's always been an irrational train of thought that the next architecture is going to be the one- the [insert huge percentage] quantum leap reminiscent of the ~1994-99 era- and it just ain't going to happen. Even if the technology breakthroughs were available, it is still two companies carving up a cake that used have fifty with their hands in it.

Guess what, Yet again i think your wrong, that next evolutionary step is going to be fairly soon(1-2years) a node down and will involve Tsv enhanced chips and additional memory near die plus some serial compute power perhaps , Btw this is only my opinion dont get too excited and thats not a vendor specific rumour imho.

 

[HumanSmoke]

Guess what, Yet again i think your wrong, that next evolutionary step is going to be fairly soon(1-2years) a node down and will involve Tsv enhanced chips and additional memory near die plus some serial compute power perhaps , Btw this is only my opinion dont get too excited and thats not a vendor specific rumour imho.

Well I certainly hope you're right. But knowing how the IHV's operate, they seldom extend themselves unless they need to, and mono/duopolies tend to incrementally advance in order to extract the most from their herd. As little as 5-6 years ago the determining factor was how big and complex you could make a GPU, and you were limited by the fab tooling. That really isn't the case anymore - litho tools have a reticle size that allows for 800+mm² (858mm for the ASML TWINSCAN for example), yet the only company taking advantage of the litho advancement in o.a. size is Intel ( Xeon Phi is supposedly 650-700mm²), and pretty much only to advance their standard/integrated product line - basically the same tactics Intel used twenty years ago with ProShare.

Anyhow, always good to here another's opinion. Here's to a brighter future