|Developer Diary · You Heard It Here First · Wednesday 18 February 2004|
|Intel's Very Fast Memory Architecture|
In the last couple of days Intel has been on the move. At the Intel Developer Forum Craig Barrett announced plans for 64-bit extensions to the x86 line and at some architecture conference Intel engineers announced a new memory architecture. My favorite thing about these kinds of announcements is the engineer-speak. For example, engineers seem to favor one-sided comparatives using the word "very". In particular I might point out how "integrated circuits" became "large-scale integrated circuits" and then they became "very large-scale integrated circuits (VLSI)". Since then integrated circuits have gotten larger and larger, but we didn't have "Very Very Very Large Scale Integrated Circuits". I guess you could just keep adding Vs: VVVVLSI. Another favorite of mine is "high frequency" and "very high frequency" and "ultra-high frequency". I guess under this logic cell phones are "ultra ultra very high frequency" and radar is "ultra ultra ultra ultra very very high frequency" and light is "ultra ultra ultra ultra ultra ultra very very very very mammothly extremely high frequency".
Anyway according to the current crop of Intel engineers the new memory architecture will be "very fast". (Until a better one comes along which I guess will be "very very fast"). I wonder do they consider DDR DRAM "slow"? In that case DRAM must be very slow and plain old RAM very very ultra slow. Maybe I should just wait until we have very very very fast memory, but if I wait until very very very very fast memory comes along I might be retired.
Also funny is the way Barrett and Mike Fister are putting the quantum spin on Itanium. They claimed there are 100,000 Itanium systems which is ridiculous. According to analysts there are only a few thousand systems out there most of which are experimental. Reminds me of Nokia's claim of having sold 400,000 N-Gages. Right.
After giving us the lowdown on their Itanium sales they announced 64-bit x86 extensions which basically drive a stake through the heart of the Itanic (finally) and then they went on to say that Itanium will be positioned as an architecture for "big databases". I think they meant to say "very big databases" or perhaps "very very big databases". The question is just how big does the database have to be to qualify for the Itanium? My guess is that it has to be very very very ultra ultra ultra big.
|return to John Chamberlain's home · diary index|
|Developer Diary · about · email@example.com · bio · Revised 18 February 2004 · Pure Content|