 |
» |
|
|
 |
 |
 |
Aside from the huge performance improvements over the last few decades, computer memory is more or less the same as it ever was. Most of your computer memory is random-access memory (RAM), the read-and-write medium that performs the lion’s share of what we generally consider to be computing. Its poor cousin is read-only memory (ROM), which contains unchangeable instructions for specialized activities like starting up the computer or running programs and sub-routines.
But there's something interesting happening on the microscopic surface of today’s memory chips. Components are so small, so sensitive, and requiring such small amounts of electrical power to perform their on/off switching functions, that the whole operation has become disrupted by particles from outer space rocketing through your PC at the speed of light. In response, engineers have had to devise error-correcting code (ECC) memory ---a type of memory that includes special circuitry for testing the accuracy of data as it passes in and out of your PC's memory.
|

Cosmic mistakes |
 |
 |
To be specific, the culprits are high-energy subatomic particles, commonly known as cosmic rays. (We'll delay the particle-vs.-wave debate for now, if you don’t mind.) These splinters of shattered nuclei have been bombarding the planet from all directions for as long as there's been a planet to get in their way. But computer scientists didn't really pay them any heed until the late 1970s, when memory bits got small enough to be affected by the occasional supersonic alpha particle.
These particles can literally flip a bit from a one to zero, or vice versa, leaving you with a piece of data that’s exactly opposite of what you want.
|
|
 |
Why errors matter |
 |
 |
Non-corrected errors can result in system crashes, or worse yet, errors that are propagated into other parts of the system. A customer with 50 systems, each with 1 GB of memory, might experience about 15 significant events a year, even with relatively conservative assumptions about error rates. Elevated temperatures or altitudes can drive an even higher failure rate.
|
 |
ECC: the solution |
 |
 |
ECC memory chips add an additional hunk of DRAM (dynamic RAM) for every 8 existing DRAM parts to accommodate data for error correction. Every time 64 bits of data of standard data are transferred to the memory controlled, 8 bits of this ECC data tag along. The ECC circuit looks for errors in the 8-bit block; when it finds them, it performs error correction on the 64-bit chunk as well.
This checking incurs no performance penalty in the latest chipsets. In addition, any 2-bit errors and multi-bit errors (for most failure mechanisms) are detected, and system shutdown is initiated.
|
|
 |
The payoff |
 |
 |
For ECC-based workstations, the probability of a significant error event is very close to zero. That's why HP makes sure that ECC memory is built into its most popular personal workstations including the HP xw4300, xw6200, xw8200 and xw9300 series products.
For customers whose success depends heavily on data integrity, ECC-based workstations provide dramatic improvements. Even if we can't shield ourselves completely from interloping subatomic particles from outer space, ECC memory minimizes the possibility of damage.
|
|
 |
|
|