AMD - Piledriver chitchat

[CAT-THE-FIFTH]

I predicted what would happen. Plenty of those posters lack the mental capability to comprehend anything I was saying.

Did he actually think people would not notice his conclusion after looking at the results??

If you notice,it is French and German sites which tend to spin less things. The English language review sites,especially the US ones seem,terrified to take on Nvidia.

Their own results don't line up with their conclusions at all.

AMD is to blame too. They need to be more proactive in marketing and making sure Nvidia does not get away from seeding reviews like that.

They have let things get this way.

Also,another question to ask,do any of these people who run the sites,give out investment advice?? We just don't really know at times.

However,astroturfing will only get Nvidia so far. Intel is gunning for them in their most profitable market with MIC and Tegra is a litany of broken promises.

People also forget that Nvidia is being paid money by Intel each year until 2016,ie,$300 million each year. This is buried in their yearly profits.

It makes me wonder if thing are so rosy for them,if we take that income out.

[Jake12321]

I like AMD processors, this is because they are better value for money than Intel however if i had the money i would definately go with intel

[The Hand]

This is probably old news for you guys but i'll post anyway!
http://www.extremetech.com/computing...lobalfoundries
If there are to be FX processors in 2015 they will likely to be based on 28nm too. Well with Intel 14nm cpus coming it does make me think the medium/high desktop arena is looking bleak for AMD. Price is king though I suppose. I wouldn't say no to a nicely priced 8 core FX chip on 28nm!

[watercooled]

It's old news that the 'leaked' desktop roadmap to 2015 has been confirmed by AMD as a forgery. 28, 22, 20, 14 are more names for the nodes than descriptive values - GloFo 28nm is in many ways already denser than Intel 22nm.

28SHP listed in that article is likely also a mistake, or referring to older documents - 28SHP was a SOI process, scrapped by GloFo when AMD decided to go with bulk CMOS i.e. 28HPP for Kaveri. Although I'm not certain on that one - several sources including some AMD documents seem to refer to Kaveri on 28SHP, but there's no reference to it by GloFo that I can find. Either it's a mistake, it now means something other than SOI, or Kaveri is SOI (unlikely).

Edit: Just to add to that, 28nm was always likely for Carrizo, using 28nm for a single processor line would be unusual. Also, we're some way off seeing anything 'high performance' from Intel on 14nm. Broadwell is already later than expected, may not even come to desktop, and if it does it may be BGA-only.

[watercooled]

I've just spotted Basilisk referenced on SA forums, as Carrizo successor. Not much info about it, but maybe something to keep an eye out for?

[DanceswithUnix]

TSMC 20nm is in production.

http://www.fudzilla.com/home/item/34...20nm-lte-modem

[watercooled]

I wonder if we'll see any RV740-style pipecleaner 20nm GPUs in the near future then?

That didn't really happen on 28nm, but it seems a bit soon for a complete refresh after Hawaii, 750Ti, mass-GK110, etc.

[DanceswithUnix]

I wonder if we'll see any RV740-style pipecleaner 20nm GPUs in the near future then?

That didn't really happen on 28nm, but it seems a bit soon for a complete refresh after Hawaii, 750Ti, mass-GK110, etc.

That would be nice.

If they can knock out a really expensive but energy efficient top end board with huge die and masses of shaders, then I'm sure the cryptocoin miners will buy them, but I won't

[CAT-THE-FIFTH]

GCN is so inefficient:

http://www.anandtech.com/show/7800/a...deon-e8860-gpu

[Bonebreaker777]

GCN is so inefficient:

http://www.anandtech.com/show/7800/a...deon-e8860-gpu

What kind of performance we are talking about either in solo or "pseudo-crossfire" configuration in exchange for the 37w consumption?

[CAT-THE-FIFTH]

What kind of performance we are talking about either in solo or "pseudo-crossfire" configuration in exchange for the 37w consumption?

Some more details:

http://www.techpowerup.com/198212/am...edded-gpu.html

The 3DMark11 score is higher than an HD5770:

http://media.bestofmicro.com/S/R/326...k-graphics.png

It also has twice the VRAM too.

Its only a bit faster than an HD7750 but drops TDP from 55W to 37W,and has twice the VRAM of normal HD7750 cards too.

[Bonebreaker777]

I could imagine that kind of performance thanks to the consumption more widespread that ever before (entry level and mid range don't have the cooling facilities as high-end) and thanks to the possibility to pair up with the APU the performance would increase further.
Hope it supports Mantle.
All is left to hope for a fair price to ensure easy choice.

EDIT: Just read the article. And I was hoping that such GPU will make it's way to the main stream in form of notebooks. Not following the article it won't. Least the new ToughBooks will have something beefier that before

[scaryjim]

It's basically just a 7850M* downclocked to 600MHz from 675MHz. I guess it might be that the design has been ported from a high performance process node to a low power process node to squeak the TDP down, hence the lower clock speed? It's a slow wide approach to parallel compute and graphics - not dissimilar to Intel's approach with Iris tbh (much larger number of EUs but lower clock speeds). It also explains why the Kaveri IGP clocks stay the same regardless of the configured TDP - there's obviously a sweet spot for clocking each GCN design where it draws very little power (I'd love to know what TDP a 7750 would have on the power-optimised GF process Kaveri is being fabbed on!).

*which itself is a downclocked 7770

[CAT-THE-FIFTH]

This profile mentions 20NM Excavator:

http://www.linkedin.com/in/kvnagesh?_mSplash=1

I saw it mentioned on SA forums.

[CAT-THE-FIFTH]

Intel Reddit AMA:

http://www.reddit.com/r/IAmA/comment..._at_intel_ama/

They have fantastic people. I cannot underscore this enough, with the resources they have the fact that they're able to compete in the same ballpark we do shows their quality. Sadly for all of us, execution is key. We want to see an exciting marketplace as much as you do.

Thanks for this . We at AMD (especially on the GPU side) have a intense amount of respect for the engineers over at Intel. What Intel has done with their recent CPU architecture, along with the constant advances in fabrication technology, they deserve a lot of credit for "keeping the ball moving forward" in our industry.

To support jecb's argument, you often hear of negative press going on between the two companies, but that kind of animosity is largely isolated to the legal, marketing, and upper management levels. The engineers at most companies tend to have many good friends working for competitors, and while we might throw in a friendly jab every now and then, it's almost a universally friendly community.

Thanks for this AMA. It's always cool to hear what it's like on the blue team .

[The Hand]

This profile mentions 20NM Excavator:

http://www.linkedin.com/in/kvnagesh?_mSplash=1

I saw it mentioned on SA forums.

VERY interesting find there, I was expecting 28nm Excavator next year but 20nm?! I'm almost certainly going to get one if this is the case, as AMD know how to price their CPUs.