AMD Claims Zen Processor Cores Will Compete With Intel, Announces Next Gen Graphics With HBM Stacked Memory
AMD this afternoon kicked off is first Financial Analyst Day in three years. During the hiatus, AMD has had a change of leadership (former CEO Rory Reed parted ways and was replaced by Dr. Lisa Su) and the company has seen its market share erode in the processor and GPU markets — the downturn in the PC market has been especially tough for AMD.
On the x86 side, AMD is pushing forward with the 2016 launch of processors based on its new Zen core. We briefly touched on Zen last week when slides leaked to the Internet, but AMD’s Mark Papermaster was on hand to give us some more juicy details on what to expect. For starters, Zen-based processors will bring with them a significant boost in absolute performance and a roughly 40 percent instructions per clock (IPC) increase. We will also see a shift from Clustered Multithreading (CMT) — which is currently employed in Bulldozer architecture — to Simultaneous Multithreading (SMT).
Although AMD isn’t getting down and dirty with specifics, it touts Zen as having a “high-bandwidth, low-latency cache system” and the fact that it will be built on an “energy-efficient” FinFET process. And the first processor out the door will be geared for gaming enthusiasts — that’s right, Zen will debut with fresh FX processors that support DDR4 memory. But whereas current FX processors are based on 32nm process technology, their Zen-based successor will take advantage of 14nm process technology which helps explain AMD’s confidence in performance as well as thermal and power efficiency improvements.
In addition to the 14nm FinFET process, all of AMD’s next generation desktop-class processors (including those that are Zen-based) will move to the new AM4 platform. That means that all of AMD processors will be able to use the same socket, which is a welcome departure from AMD’s past efforts in consolidating desktop processor platforms.
Moving past your garden variety Zen, there are also plans for a Zen+ which looks to be headed our way in the 2017 to 2018 timeframe. The goal with Zen is to introduce a high-performance, efficient core that can scale top-top-bottom, and then incrementally improve upon the design moving forward.
After playing second fiddle for so long to Intel, which has long had the upper hand in terms of performance and process technology, AMD is confident that its Zen cores will put it on equal ground with its toughest rival. We’d like to believe that AMD has what it takes this time around, but we’ve been disappointed before. With that being said, we're confident this time that AMD will make a valiant effort this time around with Zen and we can’t wait until we can get processors based on the core on our test benches.
Moving on to the graphics side of the equation, AMD is pushing its next generation Radeon graphics architecture during Q2. If you recall, HP spilled the beans earlier this week on the R9 380 with its Back-to-School desktop lineup. The next generation high performance GPUs will of course support DirectX 12, hardware acceleration for LiquidVR, FreeSync support, and smooth 4K gaming.
But perhaps the biggest addition with AMD’s new high-end graphics core comes by way of the world’s first implementation of High Bandwidth Memory (HBM). AMD’s 3D-stacked HBM gives a 1-2 punch of greatly improved memory bandwidth while at the same time reducing power consumption and footprint. AMD is claiming its HBM gives it a 3x performance-per-what advantage over existing GDDR5 memory solutions and an impressive 50 percent power savings. If AMD can truly deliver in real-world testing, it’ll have a winner on its hands.
Interestingly enough, according to AMD, HBM on GPUs is just the start — the company seems to be implying that it will extend the technology to other products in its portfolio. Could this mean a future APU with HBM? Your guess is as good as ours, but it absolutely makes sense in AMD’s traditionally bandwidth starved APUs.