AMD Carrizo Pre-release thread

[Abwx]

Well no matter what you think...at least on that picture all the GPU clocks are quite a bit higher. (136 Mhz on the top model is quiiiiite a bump)
The top range also has a higher max CPU clock.

That's why I'm saying that this picture must be faulty as hell...even if they fix up some things...then why ruin the gains again by lowering the Mhz once again? ESPECIALLY if you have a 95W TDP...which is something that especially the Athlons don't even get close.

If this picture is 100% accurate...then AMD just lost the little bit of common sense they had left. xD

Dont know the technical reason of thoses choices, could be that thoses chips, set apart the higher clocked ones, are dedicated to OEMs, i noticed that Asus sell quite a few AMD based home DTs and all use only 65W Richlands or Kaveris, 95W is definitly an enthusiast class TDP range that is well above TDPs of mainstream home or office PCs.

 

[cbn]

Given enough flash bloat, yes. But they get by pretty well otherwise. Think about it: they're faster than an E1 or E2 CPU. Lots of people out there are buying that stuff new.

And I've seen complaints about those E1 and E2 machines being too slow (even with a modern hard drive).

Anyway, my hope is that AMD can improve on this situation. With the benefit of Moore's law providing a greater xtor budget for processor designers a person shopping for a value desktop should not have to settle for a downclocked cat core box after coming from Core 2.

 

[mrmt]

The Stilt set the record straight.

Wow, so the Kaveri "refresh" is just another set of bins? AMD marketing is reaching new heights.

 

[Abwx]

Wow, so the Kaveri "refresh" is just another set of bins? AMD marketing is reaching new heights.

By ignoring posted infos one can write any bs he wants, if theses were bins there wouldnt be production candidates, productions ready waffers, and industrial productions that has not started, why bother?.

Here the thing so that everybody can see who is posting just to spread FUD given that this slide was already posted several times in this thread and just a few post ago :

 

[Erenhardt]

Wow, so the Kaveri "refresh" is just another set of bins? AMD marketing is reaching new heights.

The are definitely closing gap to intel marketing... :sneaky:

 

[Arachnotronic]

Wow, so the Kaveri "refresh" is just another set of bins? AMD marketing is reaching new heights.

Isn't that what Haswell Refresh was (sans the 4790K w/ improved packaging)?

 

[mrmt]

Isn't that what Haswell Refresh was (sans the 4790K w/ improved packaging)?

Definitely, and that's why, if not for the 4790k and the Pentium G3258, we wouldn't even bother with this so-called "refresh". Intel at least bothered to launch a new MB with more storage options, but in the end it's the same silicon.

 

[soresu]

The godavari schedule mentions software stack versions (0.7.0 - 1.0.2) tied to certain dates.

If this is just a new stepping/packaging refresh, then why does it mention new SW stack? Or is this just tied to the open source release of the HSA stack on Linux?

Perhaps SW stack refers to Mantle 1.0 public release too.

Edit: Date closely ties in with glNext announcement at GDC.

 

[erunion]

Isn't that what Haswell Refresh was (sans the 4790K w/ improved packaging)?

But Intel didn't give it a new code name and a new generation's nomenclature.

 

[III-V]

But Intel didn't give it a new code name and a new generation's nomenclature.

They did. Haswell Refresh is the internal codename, AFAIK. "Haswell" itself is a codename. The "public" name for Haswell is something like "4th Gen Intel Core i3/i5/i7 processor."

Doesn't really matter if AMD had called Richland "Trinity Refresh;" it's just semantics.

 

[erunion]

They did. Haswell Refresh is the internal codename, AFAIK. "Haswell" itself is a codename. The "public" name for Haswell is something like "4th Gen Intel Core i3/i5/i7 processor."

Doesn't really matter if AMD had called Richland "Trinity Refresh;" it's just semantics.

I wouldn't consider "Haswell Refresh" a new codename. Its immediately obvious that its still Haswell.
The same is true for the official names. I7-4770 -> i7- 4790. The relationship is obvious.

Compare that to A10-7850k -> A10-8850k. The name implies a new generation.

 

[PG]

Am I really the only one who remembers Devil's Canyon as the code name for the Haswell refresh cpus?

http://www.anandtech.com/show/8227/devils-canyon-review-intel-core-i7-4790k-and-i5-4690k

 

[jpiniero]

Am I really the only one who remembers Devil's Canyon as the code name for the Haswell refresh cpus?

http://www.anandtech.com/show/8227/devils-canyon-review-intel-core-i7-4790k-and-i5-4690k

Intel only considers the 4790K and 4690K as Devils Canyon.

 

[Spawne32]

I wouldn't consider "Haswell Refresh" a new codename. Its immediately obvious that its still Haswell.
The same is true for the official names. I7-4770 -> i7- 4790. The relationship is obvious.

Compare that to A10-7850k -> A10-8850k. The name implies a new generation.

Exactly, I would have at least presumed they would have went with something like a 7890 code name for a refresh of the 7850k.

 

[Shehriazad]

All we can do at this point is wait, anyway.

Whatever happens to Carrizo and the Kaveri refresh in the future is just something we can't know since AMD isn't really telling us everything in detail, anyway.

Sure we know some things...but we all know that companies like AMD can switch around/change everything at a whim.

And leaks like that picture of Godavari may be a lot of things...but reliable? Nuh.

 

[DrMrLordX]

The Kaveri refresh parts are (presumably) already shipping, and availability should come in the next few months, so we'll see what's up soon enough. Some of the stats seem legit and others, maybe not so much.

 

[Abwx]

The Kaveri refresh parts are (presumably) already shipping, and availability should come in the next few months, so we'll see what's up soon enough. Some of the stats seem legit and others, maybe not so much.

Slide say industrial production starting march 25, since it s related to AMD s production this means binning and packaging, so shipping, and eventual leaks, will occur on early april, product may launch only at Computex time.

 

[Shehriazad]

Well...let's just hope that they are competitive with their own products at least...since we already know that they won't be competing with Intel xP

 

[coercitiv]

From the [FONT=&quot]AMD AHCI controller driver release notes:
[/FONT]

[FONT=&quot]Revision History[/FONT] [FONT=&quot]1.3.001.0245[/FONT]

[FONT=&quot]- Fix the SB700 A12 boot hang issue if PM is enabled[/FONT]
[FONT=&quot]- Enable MSI + CCC as one option[/FONT]
[FONT=&quot]- Fix the long timer DPC WHCK failure for WNGlitchfree HD Video Playback on AC[/FONT]
[FONT=&quot]- Added Carrizo support[/FONT]

[FONT=&quot]
[/FONT]

 

[Shehriazad]

intredasting!

 

[formulav8]

I'm still waiting to hear specifically whether AMD will have Excavator desktop chips for FM2+ or not before I buy my next system. I definitely will if there will be. My Core2 x4 could use a bump up.

 

[inf64]

I'm still waiting to hear specifically whether AMD will have Excavator desktop chips for FM2+ or not before I buy my next system. I definitely will if there will be. My Core2 x4 could use a bump up.

I doubt there will be FM2+ Excavator parts. There were such plans but due to lower clocks on Carrizo (due to design and process related choices they have made) it got canned. iGPU seems to be a good deal better than Kaveri but even with IPC improvements that Excavator brings it's just not enough to counter a (wide) gap in clocks.

 

[DrMrLordX]

From the [FONT=&quot]AMD AHCI controller driver release notes:
[/FONT][FONT=&quot]
[/FONT]

Wow AMD, way to send mixed signals.

 

[Shehriazad]

I have to admit that this AHCI driver release note is a tad awkward. Still doubting to see FM2+ Excavator.

That said...something I didn't quite think about before...wouldn't Carrizo be nearly perfect for the AM1 socket?

That one focuses on entry level/emerging markets and have low TDPs. I feel like a Carrizo AM1 would totally rip apart any current chips on that platform due to the fact that Carrizo thrives in a low tdp environment.

 

[ShintaiDK]

Its a unified driver across platforms isnt it? I dont see the speculation.