You might not think the adage "less is more" applies to much in the world of computing. After all, isn't more always...
better—more memory, more processing power, more storage?
Yes, but there are some kinds of "less" that all of us can live with: less waste heat, less energy consumption, and, of course, less cost.
Intel's Core and Core 2 processors are working exercises in how less really can be more. Instead of increasing computing power by placing two discrete processors on a motherboard, why not take the guts of two processors, put them side by side on the same die, and make them work together?
That's essentially what Intel has done with the Core lineups, and the performance figures seem to bear out this move. A single Core 2 Duo Extreme running at 3.2 GHz can outperform a single Pentium D at the same clock speed. . .and not by a little, either. (The same goes for single Core chips versus multiple older-school Intel processors, and the Athlon 64 lineup, too.)
What specifically makes the Core appealing versus a discrete dual-processor system? There are several things that bear looking at in detail.
Speed. A single Core 2 processor running at a given clock speed can crank through many of the same multithreaded real-world applications twice as fast as its Pentium D or even Pentium EE predecessors. (The key word here is multithreaded; see the next section for caveats about single-threaded applications.)
A dual-processor computer dealing with multithreaded applications is also faster than a single processor, but there's also a certain amount of overhead involved in dispatching threads, managing communication between processors across the system bus, and so on. The AMD Athlon chipset has dedicated data channels for interprocessor communication, but they're still not as fast as having two cores on the same die.
Lower power consumption. The highest power consumption rating for any of the Core processors is 65 watts, compared to the 100+ watts of earlier Intel offerings. Less power also means less waste heat, and less energy required to cool each processor. A dual-processor setup burns at least twice as much energy.
Shared cache. Since both cores exist on the same die, they can use the same local cache. This means fewer round trips across the bus to the system's main memory, and it speeds things up all the more regardless of which core is executing what thread.
With dual processors, each processor has its own discrete cache; if each processor requests the same memory, it'll be cached separately and redundantly. Once again, some of this is offset in AMD's architecture, but not as completely as it would be in a dual-core setup.
Smaller form factor than a dual-processor setup. A motherboard with one processor socket requires less space and less wiring than a board with two sockets.
One possible drawback to the multiple cores as implemented by Intel is memory latency. Although this problem is more theoretical than practical, it's still worth mentioning. Unlike the AMD chipsets, the Intel Core processors don't have a dedicated data channel to the system's memory. All transactions to memory have to pass through the Northbridge chipset, which is also shared by traffic to the PCI bus and other devices. In theory, this slows down memory access, but this latency tends to be concealed from the end user—and the CPU itself—due to a large on-die L2 cache.
About the author: Serdar Yegulalp is editor of the Windows Power Users Newsletter, which is devoted to hints, tips, tricks, news and goodies for Windows NT, Windows 2000 and Windows XP users and administrators. He has more than 10 years of Windows experience under his belt, and contributes regularly to SearchWinComputing.com and SearchSQLServer.com.
More information on this topic:
- Article: Vista weighs heavily on minds at Intel
- Topics: Processors
- RSS: Sign up for our RSS feed to receive expert advice every day.