Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Advanced Micro Devices - Moderated (AMD) -- Ignore unavailable to you. Want to Upgrade?

To: TGPTNDR who wrote (81219)	5/31/2002 7:33:16 PM
From: Win Smith	Read Replies (1) \| Respond to of 275872

Historically, it's not quite that simple. Since the '60s, there have been 16, 32, 36, and 60 bit architectures that were all relatively successful. There have been 64 bit microprocessors available since I think the MIPS R4000 circa '90 or so. The 4-8-16-32 bit progression has mostly to do with what's feasible to build into a 1-chip microprocessor. For microprocessors, by the time 16 bit processorr where feasible, 16 bits for addressing was already too small; Gordon Bell acknowledged the 16 bit limitation as the primary problem of the PDP-11 right from the start, well before microprocessors came on the scene. Machines that can actually use >4gb of main memory are still pretty rare, though at current memory prices it's feasible to cross that barrier if you need it. Except you got to get out of the PC class to do it, in terms of physical addressability. Going past 64 bits, I just can't see it being worthwhile without some kind of totally new computing paradigm.

To: TGPTNDR who wrote (81219)	6/1/2002 6:19:06 AM
From: pgerassi	Read Replies (1) \| Respond to of 275872

Dear Win: Using just single die MPUs for simplicity and off the top of my head: Intel made the 8008 while they designed the 4004 so 4 to 8 bit MPU transition was none at all unless you use "shipped" time. I forget who made the first 16 bit MPU but, TI with the 990/4, Motorola with the 68K and DEC with the LSI-11 had 16 bits before Intel did 8086 in 1978-9. National Semi had 32 bits by then with the 32016. Intel shipped the 80386 by 1985-6. 64 bits occurred by 1990 by MIPs followed quickly by many others. AMD AXP64 will ship this year. Actually the transitions are going slower over time as MPUs advance to the forefront of the cutting edge CPU designs. As to when the transition to 128 bit happens, it will be about 20 years from now when memory is 3D instead of the 2D it is now. 60 bits needs 20 bits on a side (current is 32K on a side for 1Gb 2D memory chips) or 1M bits per edge. Then 256 such chips will be greater than 2^64 bytes. This is feasible with today's cutting edge geometries but, the technology will be some bulk organic or simulated organic type using nanotech. At 1000 atoms per bit (and that is probably high), 1 gram of matter would have 10^19 bits of more than 64 bits. /* fantasy on / To need more than 128 bits for addressing, even at 1 atom per bit, you would need at least 10^17 grams or 10^11 metric tons. Only a planetary computer would have that much. Beyond 256 bits, even the universe does not have that many atoms. A multiuniverse computer network? / fantasy off */ Pete