AMD: Southern Islands (7*** series) Speculation/ Rumour Thread

[AlexV]

AS may be obvious, I took the liberty of splitting the interesting discussion about the human optical system and whatnot into a separate thread that will hopefully do it proper justice. For any complaints, please contact willardjuice, and for true anger start up a blog hating on the site, IHVs, and existence in general. I do hope it doesn't come to that!:smile:

 

[neliz]

We need to do something because it's going to take a month before we get some leaked slides.

 

[AlexV]

We need to do something because it's going to take a month before we get some leaked slides.

Sure. Start by using that thread, which will make it obvious to the people looking it up that they'll get oodles of info about perceivable framerates. I know I wouldn't be looking for such info in a next GPU speculation thread, so it would be wasted if left here, especially when the slideware hits.

 

[hoom]

So has anyone spotted some new juicy bits of mis-translation yet?
This sig is feeling a bit old

 

[DarthShader]

mistranslation? Here's some via XS from Chiphell:

All the HD7XXX series products are just rebranding products. (Double confirmed)

Tahiti has met more problems than GF100. It's a big challenge for Tahiti to get an early release, due to very large die size, 384 bit vram bandwidth, very low yield, new tech like PCI-E 3.0 and higher vram clock.

The top 28nm graphics card based on single-GPU (not necessarily Tahiti) should be able to get P45,000+ in 3DMark Vantage.

http://www.xtremesystems.org/forums...-HD7XXX-Series-Latest-leaked-details-Chiphell

I suggest a grain of salt the size of the meteor that made the dinosaurs go extinct.

 

[MfA]

Just looked back in the news archives a bit ... volume production for 40nm was announced in November and AMD shipped the first 40nm GPUs in March, expecting anything this year was always hopelessly optimistic when volume production was only just announced.

 

[Kaotik]

Just looked back in the news archives a bit ... volume production for 40nm was announced in November and AMD shipped the first 40nm GPUs in March, expecting anything this year was always hopelessly optimistic when volume production was only just announced.

We, however, don't know wether RV740 was in production when 40nm was in volume production - it did face a lot of problems, didn't it?

 

[iMacmatician]

mistranslation? Here's some via XS from Chiphell:

All the HD7XXX series products are just rebranding products. (Double confirmed)

Tahiti has met more problems than GF100. It's a big challenge for Tahiti to get an early release, due to very large die size, 384 bit vram bandwidth, very low yield, new tech like PCI-E 3.0 and higher vram clock.

The top 28nm graphics card based on single-GPU (not necessarily Tahiti) should be able to get P45,000+ in 3DMark Vantage.

http://www.xtremesystems.org/forums...-HD7XXX-Series-Latest-leaked-details-Chiphell

I suggest a grain of salt the size of the meteor that made the dinosaurs go extinct.

I have no words...

Two questions though, how does the GTX 580 and the HD 6970 score in 3DMark Vantage, and how high would the clocks of either chip (including memory clocks) need to be to reach P45000+?

 

[neliz]

I have no words...

Two questions though, how does the GTX 580 and the HD 6970 score in 3DMark Vantage, and how high would the clocks of either chip (including memory clocks) need to be to reach P45000+?

Please ignore that.. please.. for the love of god ignore that!

 

[homerdog]

mistranslation? Here's some via XS from Chiphell:

http://www.xtremesystems.org/forums...-HD7XXX-Series-Latest-leaked-details-Chiphell

I suggest a grain of salt the size of the meteor that made the dinosaurs go extinct.

But it's DOUBLE CONFIRMED

 

[Alexko]

But it's DOUBLE CONFIRMED

I sense a new meme being born…

 

[CarstenS]

+1

 

[UniversalTruth]

But it's DOUBLE CONFIRMED

Double confirmed through emails with @amd.com addresses?

Lie should be repeated many times in order to have even a slight chance to become truth.

Please ignore that.. please.. for the love of god ignore that!

Yeah, because those things in chiphell are too much nonsense (BS).
I'll take your advice.

 

[hoom]

Right, that'll do it

 

[suryad]

Lol that's what I'm doing too. I'm currently on Assassins Creed Brotherhood and then might have a go at Mafia 2. Crysis 2, Lost PLanet 2 and Battlefield 3 are all waiting though.

I may have to pick up Rage is the AMD issues are sorted by the time I'm ready. I really should be able to play it perfectly at 1080p and locked 60fps on my 4890 so if I can't I'll have to pass, at least until I get an upgrade.

In fact I am on Brotherhood right now and have finished Mafia 2 and have Witcher 2 still to finish up and F1 2010 to keep me from upgrading

That said, there was some blurb I read on the web about DX 11.1 or something like that. These cards at this point in the game will probably not support that spec? Can anyone care to shed more light about what is going to be in DX 11.1 vs DX 11?

 

[Arnold Beckenbauer]

http://twitter.com/#!/AMDRadeon/status/135036137763913728

Awesome! RT @_LBasso loved the new XDR2 on HD7000. Definitely going for that @AMDRadeon
11 Nov via HootSuite

Bad joke or what?

 

[caveman-jim]

Charlie @SA says that Intel confirmed PCI-E 3 on SB-E was working by using prototype AMD boards.

 

[caveman-jim]

http://twitter.com/#!/AMDRadeon/status/135036137763913728

Bad joke or what?

Bad PR more like.

 

[neliz]

Charlie @SA says that Intel confirmed PCI-E 3 on SB-E was working by using prototype AMD boards.

:thumbsup:

 

[CarstenS]

Charlie @SA says that Intel confirmed PCI-E 3 on SB-E was working by using prototype AMD boards.

But, he needs to DOUBLE CONFIRM it. *SCNR*

Could very well be the case, but why didn't Intel then have PCIe 3.0 officially supported? Or can you just not validate something with something else that still isn't final/validated itself?