AMD’s Lisa Su Confirms Zen 3 Coming in 2020, Talks Challenges in Notebooks

At CES this week, Dr. Lisa Su sat down for a roundtable discussion with several tech publications. AMD came to CES this week with some major announcements about its Ryzen 4000 Mobile family of APUs, and Dr. Su confirmed and commented on several more points during the talk.

Anandtech has a transcript of the discussion, which is worth reading in full, but I want to touch on two specific issues that Dr. Su discussed. First, she confirmed that we will see Zen 3 in 2020, though she didn’t give a specific date or any performance information.

It’s generally expected that Zen 3 will be the last AM4/DDR4 refresh for AMD and that the company will move to new DDR5 platforms in 2021. There have been rumors that AMD targeted IPC for improvements with Zen 3 rather than attempting to lift clock speed and talk of a unified chiplet structure that would combine all eight cores around a centralized 32MB L3 cache, rather than dividing the L3 into two 16MB chunks. Optimizations like this would supposedly drive a 1.17x IPC improvement over the Ryzen 3000 family based on the Zen 2 architecture. None of these rumors are confirmed, so we’ll have to see what happens on this front.

AMD’s Mobile Strategy

Mobile has always been AMD’s weakest market, but the company appears to have a genuine opportunity with the Ryzen 4000 family. Intel’s 10th Gen Comet Lake family tops out at six cores, while AMD has squeezed eight into a suite of 15W Ryzen CPUs. Now, granted, we don’t know what frequencies these chips will hold under load, but AMD’s base frequency with 8 cores is 1.8GHz, 2GHz with the 8C/8T 4700U, and 2.1GHz on the 4600U (6C/12T). The 4500U (6T/6C) has a base clock of 2.3GHz, and the 4300U (4C/4T) rises to 2.7GHz.

Intel shows a rather different pattern. The 6C/12T i7-10100U has a base clock of just 1.1GHz. Dropping to the 105100U (4C/8T) immediately improves the base clock up to 1.8GHz. The Core i3-10110U (2C/4T) has a 2.1GHz base clock. We use base clock when discussing TDP, in all cases, because Intel derives its TDP figures from base clock. What we see here is that Intel is cutting base clock much harder than AMD when it adds cores. Moving from 4C/4T to 6C/12T triples Intel’s thread count, but cuts its base clock by 1.48x. AMD shifts from 4T – 16T — quadrupling its thread count — and only cuts base clock by about a third. We can compare against the current Ryzen 3000 Mobile family as well, though not perfectly — both the 3700U and the 3500U are 4C/8T parts, and AMD isn’t shipping one of those this time. It’s interesting, however, that the Ryzen 5 4500U maintains the same 2.3GHz base clock speed as the Ryzen 7 3700U, despite shipping a 6C/6T configuration as opposed to 4C/8T. All else equal, I’d expect a 6C/6T CPU to draw more power than a 4C/8T chip, so the fact that AMD is holding the same base clock while adding two cores is a positive sign.

This is strictly back of the envelope math, and AMD is absolutely the company making up ground in this comparison. Intel’s top-end 10nm CPUs are demonstrably more efficient in terms of battery life than AMD’s 12nm equivalents. Since TDP ratings aren’t actually measures of power consumption, we can’t draw firm conclusions about how the two companies will compare, but AMD is set to make some significant gains here.

When I covered Intel’s press conference at CES, I remarked that AMD and Intel were talking about their next-gen products in very different ways, with Intel talking up advanced partnerships in AI and showing off multiple foldable PCs, while AMD focused much more on their own APUs. The difference is more than skin deep. When asked at several points how AMD would respond to specific initiatives Intel has launched in mobile around features like Thunderbolt 3, 1W displays, and Integrated Connectivity (CNVi), Dr. Su gave a broadly similar answer: AMD has been building momentum with OEMs since before Ryzen launched, they’re seeing increased engagement and sales, and the Ryzen 4000 Mobile family should broaden the appeal of the company’s products and allow them to compete in more premium SKUs.

Intel’s massive market share and frequent use of Marketing Development Funds have given the company far more sway over the laptop designs its customers bring to market than AMD can ever hope to wield. There’s a chicken-and-egg problem here, with AMD slowly building both mind share and a customer base. Customers have to associate AMD with premium laptops enough for OEMs to want to spend the time and money to create premium experiences around AMD products. Dr. Su reiterates that AMD is going after volume areas of the market rather than trying to target smaller niches. I’m interpreting this to mean we’ll see AMD trying to build into the true premium market over several years and that it’ll be a similar process to the way the company has won space in the server market over the last three years. A lot will come down to battery life, but the mobile clock speeds AMD has set imply the company has done good work on that front.

Continue reading

AMD’s New Radeon RX 6000 Series Is Optimized to Battle Ampere

AMD unveiled its RX 6000 series today. For the first time since it bought ATI in 2006, there will be some specific advantages to running AMD GPUs in AMD platforms.

Nvidia Will Mimic AMD’s Smart Access Memory on Ampere: Report

AMD's Smart Access Memory hasn't even shipped yet, but Nvidia claims it can duplicate the feature.

AMD’s Retail Market Share Surges Based on European Reseller Data

AMD's European channel sales figures continue to be excellent, with a particular surge in November and December 2020.

AMD’s Reliance on TSMC Isn’t Harming the Company’s Growth Prospects

It has been difficult to buy high-end PC components for nearly six months. There are a number of reasons for this, including pandemic-related impacts, the related surge in demand for all computing hardware, and supply shortages. A lot of eyeballs have been trained on foundries like TSMC, to the point that national governments have put…