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.

Politics : Formerly About Advanced Micro Devices -- Ignore unavailable to you. Want to Upgrade?

To: Scumbria who wrote (51660)	3/4/1999 6:19:00 PM
From: RDM	Respond to of 1572946

I would not be surprised that the VLIW machine will be three years in reality before the first silicon of the K7 variety. It is hard to be specific because they began some before the official start which was the investment of the venture capital funding and they may be a little late. I had no idea that this process was a close to typical as it seems. It as times seemed a comedy of surprises and delays, but all non-trivial designs have that. Scumbria, Thank you for the insight.

To: Scumbria who wrote (51660)	3/4/1999 8:47:00 PM
From: wily	Read Replies (1) \| Respond to of 1572946

Matt, the author of the K7 Preview article replied to you through me via email. He asked me to post his response here: I can't reply on SI or figure out the posters' email addresses, but I can say this: An 18-month figure _was_ quoted. I may have been incorrect in saying that it will have taken 18 months from initial concept to final shipping product. What is more likely is 18 months from the start of chip (that is, the circuits that make up the chip) design. In any case, it was one of the MAIN POINTS of the lecture that the chip was finished MUCH more quickly than most microprocessors are. There were MIT EE professors in the room who considered what he said to be unusually fast, but there were no comments to the nature of "that's impossible!" I will be posting an update soon, and I will touch on this in some detail. If you could post something to the thread saying that I was planning on making another post soon, I would appreciate it. -- Matt

To: Scumbria who wrote (51660)	3/5/1999 2:57:00 AM
From: Paul Engel	Respond to of 1572946

SCUMbria - Compaq AlphaServer KICKS SOME SERIOUS BUTT newsalert.com March 04, 1999 16:45 Compaq ProLiant 7000 Posts Industry's Best Performance for Microsoft Exchange; Highest Ever MAPI Results for Two- and Four-Processor Systems Jump to first matched term HOUSTON--(BUSINESS WIRE)--March 4, 1999--Compaq Computer Corporation (NYSE:CPQ), the world's leading PC server vendor, today announced that it has set world-record Exchange Server performance numbers on the Compaq ProLiant 7000 server by achieving the highest-ever Microsoft Messaging Application Program Interface (MAPI) Messaging Benchmark (MMB) on any two-processor or four-processor platform: -- The highest MMB using four 450 MHz Intel Pentium II Xeon processors (21,500 MMB) -- The highest MMB using two 450 MHz Intel Pentium II Xeon processors (14,600 MMB) MMB measures throughput in terms of a specific profile of user actions, executed over an 8-hour day. Compaq leads the industry in deploying and managing Microsoft Exchange on its products by conducting extensive integration engineering and capacity planning. Not only does Compaq provide world-class, highly scalable, highly available server platforms, but also the experience and services necessary for successful deployments of messaging and collaborative applications. For more information about the benchmark results and their relationship to real-world deployment, please visit the following Compaq Web site: compaq.com Company Background: Founded in 1982, Compaq Computer Corporation is a Fortune Global 100 company. Compaq is the second largest computer company in the world and the largest global supplier of personal computers. For more corporate information, see: compaq.com. Compaq, Registered U.S. Patent and Trademark Office. Product names mentioned herein may be trademarks and/or registered trademarks of their respective companies. For further editorial information, contact: CONTACT: Compaq Computer Corporation Chris Woodward, 508-467-7977 chris.woodward@compaq.com or Shandwick International Julie Finn, 617-351-4189 jfinn@shandwick.com Paul PS Whops - sorry - I thought this was an AlphaServer when I posted it !

To: Scumbria who wrote (51660)	3/6/1999 3:51:00 AM
From: Ali Chen	Read Replies (1) \| Respond to of 1572946

<Three years seems to be pretty standard for all microprocessors, except for ridiculously complicated ones like ... Merced.> I see some problem with this statement. It seems apparent that if a processor heavily relies on explicit parallelizm that supposed to be dispatched by a special compiler, the design must be much simplier since you do not need to dig deep into the stream of stupid instructions to uncover internal parallelizm on the fly, renaming, speculating and predicting a lot of things. Therefore I would rather consider the K7 as "ridiculously complicated" as compared to a dumb machine to execute bunches of instructions carefully prepared in advance, in already parallel form, with predications and prefetching. From here it follows that the "ridiculously long" development of Merced has some other reasons like serious internal and/or conceptual problems. It may never see the light. - Ali