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: Tenchusatsu who wrote (55683)	9/17/2001 10:08:05 PM
From: Ronald Ashkenazy	Read Replies (1) \| Respond to of 275872

"...plans may be further advanced than hitherto suspected." theinquirer.net AMD-Intel battle takes techie twist SMP and SMT the grounds By Mike Magee, 17/09/2001 09:50:58 BST THE TANTALISING WORDS Intel used about its Jackson simultaneous multi-threading (SMT) at its Developer Forum last month seem to hint at a Holy Grail which may or may not be realised in the near future. But the significance of SMT, and for that matter symmetric multiprocessing (SMP), to Intel cannot be understated, particularly since the chip giant has acquired Compaq (DEC) engineers who already have implemented some aspects of the technology in the Alpha chip. Because it appears that the architects at AMD, headed by Dirk Meyer who had a leading role in designing the Alpha, could well be ahead of Chipzilla in implementing some advanced technology in the future Sledgehammer microprocessor. One little noticed announcement by Intel at the end of August appointed four new fellows at the chip company - senior engineers and managers who came from the Alpha background. AMD is expected to unveil some elements of its hitherto hush-hush architecture at the Microprocessor Forum later this autumn, but existing documents by chip architects indicate that the plans may be further advanced than hitherto suspected. At last year's Forum, Big Blue said it would incorporate four CPU cores into one package, with each of them communicating with each other, and there are some hints that AMD may do a similar thing with Hammer technology. There are several ways of implementing parallelism on a computer and, theoretically at least, SMT appears to be the favoured route. Nevertheless, putting CPU cores together, according to one paper by distinguished architects, can achieve quite a large performance hit - perhaps as much as 70 per cent. AMD therefore may be able to produce a Hammer chip that can deliver three times the performance of Athlon technology - something that CEO Jerry Hammers claimed was possible as long ago as spring last year. At the same keynote, Sanders also indicated that AMD first and foremost wants the Hammer family to be the fastest possible core for 32-bit code - adding X64 as additional fuel. We now expect Clawhammer to offer serious performance competition to Intel processors while SledgeHammer - which according to some will sample in November, may well include some of the features architects appear to think will give microprocessors an extra edge. How far away Intel is from its SMT implementation is currently unknown, but having both Alpha and for that matter PA-RISC architects on board mean serious technology jumps ahead on its side. To understand SMT and SMP technology, it is worth looking at these publicly available documents here and here. The first, with illustrations, is an IEEE document.