AMD AMA - Starts at 12PM EST on Wednesday May 9

[SOSKI_KROTA]

Hi James, love to see AMD having great success!

1) What do you think about using infinity fabric in videocards? Does it have many perspectives? How much software optimization will be needed?

2) What do you think about TSMC's Wafer on Wafer technology?
https://www.pcgamesn.com/nvidia-amd-tsmc-3d-gpu
Do you think it will be implemented in the near future? What about cooling solutions for there?

3) What do you think about AMD's premium package cooling solutions for high end CPU's?
I always thought that would have really positive effect because you can market it much more aggressively.

Go AMD!

 

[dcominottim]

Broad strokes, the phases look something like this:

• Architecture development
• Product outline definition
• Product modeling and synthesis
• First tape out
• Bring up validation
• Refinement and improvement for performance, yield, cost, features
• Final Tape out
• Platform Validation
• Mass Production

The timing varies depending on a lot of factors, but overall that's a 3-5 year process, depending on a number of factors.

Wow, that's such an amazing answer! Would love to know the average duration of each phase, but your answer was formidable already. Thank you so much for the AMA and all your replies. You're a great guy and professional, and it's also amazing to see this level of engagement with the community from AMD. Congrats!

Oh, and by the way, can I buy an AMD Ryzen T-Shirt with global shipping from somewhere?

 

[ThatBuzzkiller]

I'm not in tech marketing, but we do evaluate different instruction extensions to help improve our products, based on area, efficiency, adoption rate, implementation cost, etc. I have no info on RoCM, sorry.

Awesome, are you guys specifically considering or focusing on AVX-512 and transactional memory support though ?

 

[dcominottim]

We're looking at ways to make it simpler, for sure. It's mostly a cost and alignment problem, which means it's solvable with time.

Despite the hassles, I think end users prefer to have some kind of cross-compatibility than a forced purchase of new platforms. How we set the expectations was what tripped us up.

You're definitely right about what end-users prefer. I chose AM4 instead of Intel's upcoming CFL 8C because of long-term compatibility. Thanks for being so pro-consumer on that one! And your effort to provide boot kits has been nothing short of amazing.

 

[PingPongun]

There can be found some news about 2800u and 2600u, which, according to videocardz.com, are described by YM2800C3T4MFB and YM2600C3T4MFB,

If I understood it correctly

(Y-means Ryzen; M-mobile/D-desktop/E-embedded; next is name(last digit can by exchanged with U or X or B-pro mobile); next is TDP marker:{ C3-45W(35w-54w), C4-15w(12w-25w), C6-35W, C5-65W, BB-65W, BC-95W, BG-105W, A8-180w}(c is present only with APU!); next core type?{T-mobile; M-desktop; AM-desktop pro(one additional 0 in name omitted; U-Threadriper}; core count( 2- dual core; 4-quadro core;6;8; C-12cores; G-16cores); MFB and OFB(?)-appears only in APU/ last sign in Threadripers and desktops indicates generation(?){E-1st gen; F-2nd gen})

It means that these new Ryzens would be quite the same as embedded V1807B and V1756B (YE1807C3T4MFB and YE1756C3T4MFB) (TDP=45W(35-54W); Base freq.=3,35/3,25GHz; Dual-channel 64-bit DDR4 Up to 3200 MT/s; Vega 11/8; GPU freq.=1,3GHz)


1) Would (2800/2600u and V1807B/V1756B) be exactly the same chips or would they be slightly improved/would they really be capable of using such a nice(for mobile) RAM.

2) If they really are going to use vega 11/8 without any VRAM, wouldn’t it be a nice idea to launch also the same CPU chip, but merged with only vega 2or3( vega 11/8 may be too week to many more demanding games, so it must be used an additional GPU for them and the power of internal GPU isn’t used, the weaker internal GPU’s should lower costs, so also consumer price, so it can reach higher market share) or launch an version with equivalent( maybe stronger)of VEGA GH/GL

 

[jsimenhoff]

And that's a wrap folks!

We'd like to give big thanks to our guest, James Prior, for tirelessly answering as many questions as he could muster. Thank you also to the rest of the team at AMD for helping us organize the AMA.

And last but not least, thanks to all of you for submitting your questions. Your technical expertise made our AMA a smashing success. Until next time!

Do you have a special guest in mind for the next AMA? PM me with any requests and we'll try to make that happen.