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 : AMD:News, Press Releases and Information Only! -- Ignore unavailable to you. Want to Upgrade?

To: StockMan who wrote (4766)	3/5/1998 5:46:00 PM
From: Kevin K. Spurway	Read Replies (1) \| Respond to of 6843

I think YOU need to learn to read. If AMD has 1000 wafer starts TODAY, it doesn't mean they had 1000 wafer starts MONTHS AGO when the wafers that are NOW yielding 50% were started. Understand? That is not to say that the report is accurate. It may be incorrect, for all I know. I'm taking it with a big grain of salt until there's better evidence than the report of one analyst. But I don't think the report is providing information inconsistent with itself. So, to answer your question, what 266's there are are in the channel. That means IBM and Compaq at this point. Kevin

To: StockMan who wrote (4766)	3/5/1998 7:02:00 PM
From: Petz	Read Replies (1) \| Respond to of 6843

RE: confusion in Kumar interview article news.com Here are the two paragraphs about which everyone is confused: Earlier in the quarter, the company was able to isolate, and fix, one of the major manufacturing flaws with the K6. Last fall, yields on some K6 processors were as low as 20 percent, but now yields per wafer have jumped up to 160 chips out of a possible 340 per silicon wafer, or almost 50 percent. While improvement is still needed, the yield ratio is far higher than it was a few months ago. "50 percent is definitely a good starting point," Kumar said. AMD is also showing progress in producing chips according to the more efficient and profitable 0.25 micron manufacturing process. Currently, about 1,000 of the 3,000 wafers started each week at its "Fab 25" plant in Austin, Texas, are made under the 0.25 micron process, he said. 1. Obviously the first paragraph is talking about the 0.25 process. The interviewer can certainly make mistakes in writing but I'm darn sure he/she wrote down the numbers properly. 2. My interpretation of the second paragraph is that what Kumar really said was, "AMD is also making progress converting production to the more efficient and profitable 0.25 micron process... This is consistent with the rest of the paragraph which talks about the number of wafer starts for each process type. Also, why would Kumar refer to the 0.25 process as "more profitable", if its yields were still 1%? It is not a trivial process to convert an assembly line from 0.35 to 0.25 process and there is machine downtime involved in the conversion. At the end of Q4 there were ZERO wafer starts for the 0.25 micron process in Fab 25. Now there are 1000 wafers a week. By the end of 2Q, according to AMD's timeline, there will be 3000 per week. In short, the first paragraph is talking about improved yields and the second about assembly line conversion. Note that, because of a 2 to 3 month wafer processing/packaging/testing cycle, we will see very little, if any, of the 0.25 output from Fab 25 this quarter. But we will see at least 131000160 0.25 micron chips in the second quarter. <Also ask yourselves this, where are the 300Mhz, I haven't read Tom's hardware page lately, but there is no mention about 300Mhz.> The K6-3D is supposed to run at 300 MHz and 350 MHz. But to run these, you need a motherboard that supports 100 MHz bus, and its not quite there yet. AMD is probably witholding samples until the boards work at 100 MHz with a least a few brands of SDRAM. Otherwise, people will publish benchmarks at 75*4, which will be lower performance. Also, I suspect they'll wait until Windows '98 instruction set support. BTW, a Euro magazine did publish some benchmarks of a K6-3D running at 300 MHz. Petz