AMD: Volcanic Islands R1100/1200 (8***/9*** series) Speculation/ Rumour Thread

Nice stuff!

Maybe Devastator and Scrapper are Trinity and Richland (the former is hilarious :)

If you're looking for a VLIW5 with low numbers surely there'd be Bobcat APU in there.
 
Something I've overlooked before - CodeXL already gave us the information SweClockers is saying, more or less, since it doesn't only put it above Pitcairn (M290X), but above Tahiti Pro too (W8000)
 
Something I've overlooked before - CodeXL already gave us the information SweClockers is saying, more or less, since it doesn't only put it above Pitcairn (M290X), but above Tahiti Pro too (W8000)
Firepro W8000 is a bit more crippled than Tahiti Pro, as it has its memory interface reduced to 256bit.

I would also expect AMD to take advantage of 8Gb memory modules as in the W9100 and PS4 to equip the W8100 with 8GB of VRAM, twice that of the W8000, which could be enough to warrant the higher model number even if GPU performance alone isn't.
 
http://www.cadence.com/Community/bl...-20nm-16nm-finfet-and-3d-ic-technologies.aspx
While TSMC has four "flavors" of its 28nm process, there is one 20nm process, 20SoC. "20nm planar HKMG [high-k metal gate] technology has already passed risk production with a very high yield and we are preparing for a very steep ramp in two GIGAFABsTM," Sun said.

And again some year later TSMc again disagrees with them selves announcing rampant 16nm FF/FF+ production
http://www.cadence.com/Community/bl...-20nm-16nm-finfet-and-3d-ic-technologies.aspx

20nm, 16nm and 10nm Progress
While TSMC has four "flavors" of its 28nm process, there is one 20nm process, 20SoC. "20nm planar HKMG [high-k metal gate] technology has already passed risk production with a very high yield and we are preparing for a very steep ramp in two GIGAFABsTM," Sun said.

Accordingly to TSMC we're actually see very short lived 20nm lineup in HPm so all this delays and waitings for newly AMD redesigns might come up in vain to end user. We once more expected much more proactive redesigns from AMD aimed at lower mainstream 3D segments as we witnessed with GM107 which actually doesnt have price or performance per watt counterpart. Instead we actually got nothing in return for more than 30 month of ffailed expectations except slight PitcairnXT to CuracaoXT redesign which presumably only served internally with polishing up heatspots in turn resulting with better yields, which actually didn't reflect at all with presumably better price or TDP improvements.
 
I wonder why AMD stopped to publish die shots?

AMD hasn't regularly published die shots. RV770 was interesting because it was such a clear shot of the actual chip.
If it's not the common case where AMD does nothing or uses a different chip's photo, it's a blurry or obfuscated picture.

It used to be much better with CPUs and early APUs, but that died as AMD's market presence did.
 
Wonder how hard it really is to make your own die shots? Sounds like a good click-bait way for tech websites? Remove die from substrate, use some acid, take picture, add website watermark. Profit!

These guys (http://zeptobars.ru/en/) do it for smaller ICs, but AFAIK not for flip chip stuff.

(Edit: these zepto guys are apparently boiling highly concentrated sulfuric acid at very high temperatures to do their magic. That may be a first clue as to why not everybody is doing it.)
 
Last edited by a moderator:
AMD hasn't regularly published die shots. RV770 was interesting because it was such a clear shot of the actual chip.
If it's not the common case where AMD does nothing or uses a different chip's photo, it's a blurry or obfuscated picture.

It used to be much better with CPUs and early APUs, but that died as AMD's market presence did.

They still publish die shots for APUs.
 
Which high-quality Kaveri shot are you aware of?

With Richland, I saw a graphic that showed a Trinity die shot that was copy-pasted to the other side of a comparison picture.
That may have been a joke, but an understandable one.
 
Which high-quality Kaveri shot are you aware of?

With Richland, I saw a graphic that showed a Trinity die shot that was copy-pasted to the other side of a comparison picture.
That may have been a joke, but an understandable one.

Kaveri's die shot wasn't very clear, that's true. But there were die shots for Llano, Trinity, and I assume that Richland does look identical to Trinity. There are also die shots for Bobcat, Kabini and Beema, as well as Zambezi and Vishera.
 
The little brother of the W9100 is launched: Look for me it is just an Hawaii cores ( look like the 290 ) http://www.amd.com/en-gb/products/graphics/workstation/firepro-3d/8100#

Bolstering the AMD Professional Graphics product family with yet another productive workstation solution, the AMD FirePro W8100 features a best-in-class 8 GB of GDDR5 memory, up to 4.2 TFLOPS of single precision compute performance, and unmatched double precision compute performance of up to 2.1 TFLOPS.



Full article
http://www.guru3d.com/news_story/amd_launches_firepro_w8100_professional_graphics_card.html

In addition, there will be a sneak preview of the S9150 thoses next days on the International Super Computing 2014
 
Last edited by a moderator:
The little brother of the W9100 is launched: Look for me it is just an Hawaii cores ( look like the 290 ) http://www.amd.com/en-gb/products/graphics/workstation/firepro-3d/8100#
Hmm so the first number indicates the "class" the second the generation or something. Seeing a W8100 is superior to a W9000 in every way can't say really I agree with that but I guess it makes sense for some...
In any case this should be quite a competent workstation card (so the large price hike in the class is justified), just a bit slower than the W9100, and with a lower TDP to boot - so the latter probably only really makes sense if you need 16GB of VRAM.
 
Back
Top