January 3, 2011
Intel today officially announced the first products based on the much-discussed Sandy Bridge CPU architecture, and first impressions are highly favorable, with my take being that Sandy Bridge represents the first step in a very aggressive product road map for Intel in 2011.
Sandy Bridge is the next architectural spin after Intel’s Westmere shrink of the predecessor Nehalem architecture (the “tick” in Intel’s famous “tick-tock” progression of architectural changes followed by process shrink) and incorporates some major innovations compared to the previous architecture:
- Minor but in toto significant changes to many aspects of the low-level microarchitecture – more registers, better prefetch, changes to the way instructions and operands are decode, cached and written back to registers and cache.
- Major changes in integration of functions on the CPU die – Almost all major subsystems, including CPU, memory controller, graphics controller and PCIe controller, are now integrated onto the same die, along with the ability to share data with much lower latency than in previous generations. In addition to more efficient data sharing, this level of integration allows for better power efficiency.
- Improvements to media processing – A dedicated video transcoding engine and an extended vector instruction set for media and floating point calculations improves Sandy Bridge capabilities in several major application domains.
- Major improvements in embedded graphics – Intel’s embedded graphics have never satisfied dedicated gamers, and the new graphics pipeline in Sandy Bridge probably won’t either. But for the rest of us, there are major improvements for desktops and laptops. In addition, power to the graphics processor can be managed separately from that of the main CPU, improving power efficiency. Do we sense a theme here? It’s not clear yet what changes Intel will make to the embedded graphics on the server variants that will roll out during the year, but Intel has made a point of mentioning that the graphics processing architecture is composed of a number of dedicated-function models that, at an architectural level, can be easily modified or repurposed, leading us to speculate that on a server variant they could be either removed or dedicated to specialized functions of interest to server users.
So what is the net in terms of performance? For servers, performance remains a work in progress, but third-party reviews of the performance of the first available product, the Core i7 2600K (I for one long for the good old days of simpler and more mnemonic product names) indicate significant performance gains on a per-core basis. My best guess for any early server chips, think “Sandy Bridge EP” is that a six core Sandy Bridge should offer 50%–100% better performance than the equivalent Westmere-based product.
Combine the prospects of a major performance boost with somewhat better power efficiency and we have the basis for a major cycle of new product innovation in the server space, along with increased heat on AMD. All in all, Sandy Bridge looks like great news for infrastructire and operations groups across all segments.