We won't burn up too much of your bandwidth setting the stage here. AMD's second-generation Bulldozer core processor microarchitecture, codenamed Piledriver, has made headlines at HotHardware many times in the past few months, including our CES sneak peek of the chip that AMD is launching today for the mobile market, codenamed Trinity. What this launch is all about is AMD's answer to Intel's Ivy Bridge-based Core series processors for notebooks. It's that straightforward, though we'll start by level-setting expectations based on how both companies and their respective architectures approach computing workloads.
There's little debate that Intel's strength lies in their base x86 CPU architecture, which has held a significant advantage over AMD for several generations of product now, both in desktop and mobile flavors. AMD has been left to compete on cost more than anything else (and in some recent products, low power) and when you consider Intel's clear advantage in leading-edge silicon manufacturing process, that's a thin line to walk.
Conversely, AMD enjoys a distinct advantage over Intel with their Radeon graphics engine, in terms of the integrated solutions they've brought to market in the mobile and desktop spaces and in its software support. The dividing lines are fairly plain to see, with a quick examination of how each company partitions silicon resources. Intel spends a disproportionate about of silicon area on x86 CPU and memory resources, whereas, even within Trinity, AMD carves out more area for the GPU. Regardless, Ivy Bridge brought Intel's graphics game up significantly, as you're aware if you've been keeping up on our coverage here, with a fully DX11 capable graphics engine, along with a demonstrated up to 2X performance boost over Sandy Bridge's graphics engine, not to mention the traditional IPC kicker Intel engineered into their 3rd generation Core CPU cores.
For AMD, Trinity has also been reported as offering much-needed performance enhancements in IPC (Instructions Per Cycle) but also more of the same strength in gaming and multimedia horsepower, with an enhanced second generation Radeon HD graphics engine. In the pages ahead, we'll dive into AMD's new Trinity architecture and AMD's new A10 series APU, along with lots of benchmark data looking at the results of AMD's recent efforts from a number of angles.
AMD Trinity A10 APU Die - 1.303B Transistors, 32nm - 246mm2
- Die size: 246mm2
- 1.303B transistors
- Process: 32nm SOI
- 2.09W MobileMark 07 power consumption
- 1.08W idle power
- Power reduction during HD media playback
- Unified Northbridge (UNB)
- Quad-core and Dual-core configurations
- Updated AMD Radeon™ DirectX®11 GPU
- 3 dedicated display outputs
- 4 independent display controllers
- DisplayPort 1.2 with symbol rates of 1.62, 2.7 and 5.4 Gbit/s
- UVD and AMD Accelerated Video Converter
- IOMMU v2 (Input/Output Memory Management)
AMD A-Series Trinity Processors (click for high res)
As you can see, Trinity sips barely over 1W of power at idle and the current top-end A10 chip has a 35W TDP with a max clock speed of 3.2GHz, versus the Core i7-3720QM we tested not long ago at 45W. Beyond that, you can see that AMD is bringing out several versions of their Trinity architecture, with both dual-core and quad-core variants, along with various Radeon HD 7000 series integrated GPUs on board, clocked at different speeds.
AMD's lowest power chip is the A6-4455M, which is a dual-core CPU with a Radeon HD 7500G graphics engine on board consisting of 256 Radeon cores. Today we'll be showing performance of the A10-4600M Trinity processor--a quad-core design with on-board Radeon HD 7660G graphics comprised of 384 Radeon cores, running at a peak clock speed of 686MHz for the GPU and 3.2GHz on the CPU. First, let's look under the hood a bit with Trinity.