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 : Adaptec (ADPT) -- Ignore unavailable to you. Want to Upgrade?

To: Starowl who wrote (2683)	6/15/1998 4:55:00 PM
From: The Philosopher	Read Replies (1) \| Respond to of 5944

The company may not believe in the death of SCSI, but the market was singularly unimpressed with the three announcements. We've all been asking for announcements to buoy up the stock, we get three, and the stock drops again.

To: Starowl who wrote (2683)	6/15/1998 4:55:00 PM
From: Investor2	Respond to of 5944

Looks like quite a bit of ADPT selling in the last 30 minutes of trading. Was there some news out? or a mutual fund dumping? Best wishes, I2

To: Starowl who wrote (2683)	6/15/1998 7:09:00 PM
From: Mark	Read Replies (2) \| Respond to of 5944

Mark Brophy: That's a nice surprise. It means the fab is more capable than I imagined it would be. This raises more questions in my thinking about its utility for the "new Adaptec" (after Symbios comes on board). Although Adaptec hasn't run a fab, as such, Symbios has, so Adaptec would be acquiring the requisite fab expertise with the deal. We can guess what Adaptec would do with the fab until hell freezes over; or we could wait and find out. Starowl - I have assumed it is unlikely that ADPT will be able to put their current product line through this fab....... - (a) they will probably have existing term supply agreements with punitive termination clauses, (b) their designs will be suited to specific fab processes and there is only a slim chance that Symbios will have these same processes. Some arcane data: Forbearance from you tech wizards please, but I recently learned that the 400Mbps (bits) figures we hear for some 1394 devices is actually slower than the 80Mb/sec (bytes) offered by the Ultra2 SCSI card. (To exceed that rate, 1394 would have to get to 800Mbps--which is the rate TI just announced for its 1394 product projected for the end of this year.) Clarifications would be welcome. Some clarification is as follows - SCSI is a "parallel" bus and 1394 is a "serial" bus. SCSI uses separate wires to send the 8 data bits (or 16 or 32 according to which flavour SCSI !) at the same time. SCSI is also "bus" based. This means that many devices can be connected in parallel to the same basic wiring (usually up to 8). This makes the topology nice and easy, but slows down the basic "clock rate". SCSI bus systems run at basic clock rates of 10, 20 or 40 MHz. This means that in an "8bit" implementation, where each clock transfers a whole data byte, the transfer rate is 10, 20 or 40 MB/s. (The "16bit" and "32bit" versions transfer 2x and 4x as much data). In practice, every byte that is sent (in either system) needs some help to guarantee the transfer. In the SCSI (parallel) system additional wires are used to carry this information. This has the effect of increasing the wiring requirement needed to carry the data. 1394 systems are "point to point" systems, and don't have to compromise clock rates to cope with loading issues. They therefore run at higher (than SCSI) clock rates of 200MHz, 400MHz or 800MHz. However, some of the bits which are sent, need to be used for some of the helping signals, which means that some of the clock rate is not used for carrying data. The bottom lines here are that a SCSI system is a parallel system, and has a "data rate" specified in MByte/s which is real; the 1394 is a serial system and has a "bit rate" specified in Mbit/s, which is higher than the rate at which real data is sent. (For Ethernet, which is a serial system, it is usual to assume the peak data rate is about 75% of the bit rate - not sure if a similar ratio is valid for 1394). FYI - the wiring in the SCSI system is not very efficient. SCSI needs 9 helper signals, a few power signals, and every 8 bits of data has a 9-th signal for data integrity - i.e. that's around 20 signals for an 8bit SCSI. Furthermore, each signal uses two wires, which makes a basic SCSI 40 wires ! (16bit SCSI uses 68 wires !!) I don't think 1394 is as simple as a "wire pair", but I doubt it is much more than 3 pairs...... (i.e. 6 wires). Mark