SI
SI
discoversearch

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.
Strategies & Market Trends : Systems, Strategies and Resources for Trading Futures -- Ignore unavailable to you. Want to Upgrade?


To: Judy who wrote (21704)4/25/1999 9:25:00 PM
From: Jorj X Mckie  Respond to of 44573
 
Judy,
I will dig a little on the RMBS thing. It gets over my head pretty quickly, but I do have those engineer friends to consult with.

Saw on Lastshadow's thread that you brought up that ORCL is at resistance. It broke through a P&F double top on Friday after coming within a couple of boxes from the BSL. Looks to be trending with the BSL in an upward direction now. Support at $23. (I have a small position in ORCL shares).



To: Judy who wrote (21704)4/25/1999 10:42:00 PM
From: Juli  Respond to of 44573
 
Judy, there is an alternate to Ranbus and it is being used now because Rambus has been late getting their product out. I will try to get the details and name of the company and post it here. My daughter, who works at Dell, sent me an article awhile back. Also, I have a good friend who works for Viking memory and she is up to date on Rambus and competition.

Jewel



To: Judy who wrote (21704)4/26/1999 9:52:00 AM
From: Michael Rich  Read Replies (1) | Respond to of 44573
 
Good morning, Judy. I'm back to the rainy Seattle. Cannot
believe it was still raining yesterday!

Holding 5 internet stocks, eh? Me too! Turned most calls
into free shares at about the right time. Sold some puts
on AOL/ATHM/CMGI...a bit too early but looking decent now.
Didn't get all the rush Jer is experiencing tho.

Talk to you later.



To: Judy who wrote (21704)4/26/1999 11:19:00 AM
From: Juli  Respond to of 44573
 
Judy, here is a press release on the alternative to Rambus.

April 26, 1999, Issue: 1058
> Section: News
>
> Taiwanese rally around PC133 SDRAM alternative
> Mark Carroll
> Taipei, Taiwan - Computer and motherboard manufacturers converged here to
> rally behind the PC133 SDRAM specification, which seems poised to
> giveRambus a run for its money. Nearly 400 engineers attended PC133/266
> Technology Forum '99 to see working computers with PC133 SDRAM on
> reference boards produced by core-logic vendor Via Technologies Inc.
> Representatives of more than a dozen DRAM vendors were on hand to promote
> the specification. In moving to more-advanced process technology, partly
> to shrink die sizes, many of the DRAM vendors are yielding SDRAMs that
> meet the 133-MHz speed grade, resulting in a small premium of 5 percent or
> less for the PC133 parts compared with the PC100 SDRAMs. In addition,
> several DRAM vendors already are shipping commercial samples of the
> 256-Mbit SDRAMs on modules that meet the PC133 specification for the
> server and workstation markets, said Dave Pulling, vice president of
> marketing at core-logic vendor Reliance Computer Corp. (Santa Clara,
> Calif.).
> Eric Chang, senior marketing manager at Via, said that in the current
> quarter, Via will introduce desktop core logic-named "PC133 Enhanced
> Apollo Pro Plus"-which supports the PC133 SDRAMs. Soon after that, Via
> plans to combine graphics functionality on a notebook chip set, and by the
> second half will introduce core logic that supports double-data-rate
> SDRAMs, called PC266, as well as 4X AGP (advanced graphics port), for both
> desktops and notebooks.
> Meanwhile, Smart Modular Technologies (Fremont, Calif.) announced a line
> of buffered and unbuffered PC133 modules, in densities up to 1 Gbyte.
> Smart Modular, which worked closely with Intel Corp. on the validation of
> the PC100 modules, is cooperating with Via to develop the validation and
> characterization test suites needed to prove out the devices, modules and
> PC133-based systems.
> While Smart Modular is unlikely to fill the role Intel played in the
> difficult shift to the PC100 specification, the overall industry is more
> familiar with what is needed now to move to faster modules.
> Keith McDonald, who recently joined Smart Modular as president, said he
> believes both PC133 and Rambus-based solutions will enter the marketplace,
> but at different times and price points. "From Smart's perspective on
> Rambus, it is not a question of if, but when." He sees PC133 systems by
> next month and DDR in the third quarter. The computer industry needs "to
> market something new, and there is nothing to fill that hole right now,"
> McDonald said. "PC133 fills that hole. Rambus will come in later at the
> high end, and then move down to lower price points."
> PC133 SDRAMs will be introduced in both servers and desktops at about the
> same time, McDonald said.
> Bert McComas, an analyst with InQuest Market Research who spoke at the
> forum, said the needed element is a central body that can validate the
> reference boards and initial motherboard designs. "Intel will stick with
> Rambus. They won't support PC133 or DDR publicly this year. Next year,
> Intel might have to flinch, however, if the market acknowledges the value
> of these memory solutions," he said.
> Dong Kim, assistant marketing manager for Hyundai Electronics Co., said
> "the key hurdle is the spacing of the core logic and the memory modules.
> The Via boards place the core logic from 1 to 1.5 cm from the memory
> modules, almost flush with the CPU. Timing and spacing is still an issue.
> By the end of the second quarter, though, it will be worked out."
> Kim added that once testing costs decline, by year's end there may be no
> price premium at all for PC133-based DRAMs.
> Gil Russell, senior technical marketing manager for Infineon Technologies
> Inc. (formerly Seimens Microelectronics Inc.), said, "DRAM vendors will
> have to provide new levels of support for PC133."
> Most of the work, however, will fall on Taiwan's core-logic vendors, who
> are providing the reference boards, McComas said. "Via is concentrating on
> the unbuffered, low-end PC133 solution, while Acer Laboratories Inc. (ALi)
> is focusing on the buffered solutions for the server segment. This
> movement toward PC133 and DDR is a return to an old technology paradigm of
> non-monolithic market pushes but individual creativity."
> "In cost-conscious Taiwan, the difference between a PC133 memory solution
> and a Rambus solution is the key sticking point here. We asked Intel to
> not give up on SDRAM yet," said Jeb Kuo of Acer Open Inc. "The higher cost
> of Rambus DRAM is still its major flaw. It's just too expensive."
> An engineer working at First International Computer Co., who asked not to
> be named, agreed. "We are not in a hurry to develop Rambus DRAM boards.
> First, it's a matter of cost. A Direct RDRAM module could be about $200
> more expensive. Also, we have no working samples of Camino," Intel's core
> logic that will support a 133-MHz front-side bus, the Pentium III
> processor and Direct Rambus.
> McComas said "the total cost premium for a system using Rambus DRAM is
> almost $300. The premium for a raw D-RDRAM is about 35 percent higher than
> for a 133-MHz SDRAM. When packaging, testing, ECC compatibility and
> ramp-up costs are added, the premium is like 50 percent. Further, now that
> OEM orders for DRDRAM are plummeting, the price must rise to pay for
> production costs ,so the final price premium is around 70 percent."
> McDonald said, "based on the data that Smart Modular is familiar with," he
> expects an initial price premium-comparing a 128-Mbyte SDRAM module with
> an RDRAM module or RIMM-to be in the 70 percent range. As volumes ramp,
> RIMM costs will decline.
> But McDonald said he believes an overall DRAM shortage will exist by the
> fourth quarter. That will raise the question of "whether DRAM prices will
> come down as fast as costs."
> Arthur Sainio, product line manager at Smart Modular, said engineers from
> Rambus and Intel are "doing a good job of teaching the industry how to
> build a RIMM module, going from A to Z. Intel's support structure is
> definitely helping the industry work out the snags."
> Meanwhile, South Korea's Samsung Electronics Co. Ltd. announced its 72-
> and 144-Mbit RDRAMs have passed the Rambus DRAM Validation Test conducted
> by Intel. The company said the tests prove Samsung's devices meet Intel's
> specs.
> Samsung said it plans to increase its 128- and 144-Mbit RDRAMs to 5
> million per month by the second half of the year. Samsung said it expects
> the RDRAM market to account for 50 percent, or $13.5 billion, of DRAM
> sales by 2001.
> Also last week, Fujitsu Ltd. said it has started selling samples of its
> 128-Mbit SDRAMs conforming to the PC-133 standard. Using a 0.2-micron
> process technology, the company said it was able to achieve a die size of
> only 85 mm2. The devices target servers and workstations. Fujitsu will
> sell x4, x8 and x16 devices and 168- and 144-pin DIMMs. The company plans
> to sell 1.3 million 133-MHz SDRAMs a month in volume starting in
> October.-Additional reporting by David Lammers and Anthony Cataldo
>
> Rob Schwartz
> 303/221-4144 phone
> 888/794-6854 pager




To: Judy who wrote (21704)4/26/1999 11:24:00 AM
From: Juli  Respond to of 44573
 
Another slightly different article on Rambus alternative.

April 26, 1999, Issue: 1157
> Section: News
>
> PC133 now 'PC' at PC OEMs
> Andrew MacLellan
> Silicon Valley- OEM interest in PC133 SDRAM has increased substantially in
> the weeks following Intel Corp.'s delay of Direct Rambus DRAM,
> particularly among desktop-PC makers, according to a number of industry
> executives.
> With demand prompting Smart Modular Technologies Inc., Unigen Corp., and
> other module makers to ramp production of PC133 DIMMs, momentum for what
> until recently had been the industry's fall-back architecture is clearly
> on the rise.
> "In general, you're getting a couple of segments looking at it, especially
> consumer, because of the delay of Rambus and a perceived price increase
> associated with that," said Dan Pleshko, director of memory and
> microprocessor corporate procurement at Compaq Computer Corp., Houston.
> Kingston Technology Co., a module maker based in Fountain Valley, Calif.,
> is producing PC133 DIMMs on a contractual basis for Toshiba Corp., which
> has also seen OEM interest grow in the weeks since Intel delayed Direct
> RDRAM.
> "We've got two of the top three PC OEMs planning to use PC133," said Kevin
> Kilbuck, manager of memory engineering at Toshiba America Electronic
> Components Inc., Irvine, Calif. "I wouldn't say they're 100% confirmed to
> be using it, but they are at least qualifying the chipset and Toshiba
> PC133 modules."
> Keith McDonald, newly appointed president of Smart Modular in Fremont,
> Calif., said calls for the architecture have been pouring in from PC OEMs
> considering PC133 for use in new desktop platforms.
> "What I know from my past is that Rambus was on the front burner and PC133
> was a back-up strategy," said McDonald, who until last month was senior
> vice president of sales and marketing for DRAM maker Samsung Semiconductor
> Inc. "Now, it's been moved to the front burner, and over the last four
> weeks, we've seen many requests for PC133 modules."
> The bulk of the demand is coming from desktop and workstation
> manufacturers, according to Scott Marx, director of sales at Unigen, a
> module maker also based in Fremont. "Even the embedded computers that are
> currently using SDRAM are moving to PC133 in the next generation."
> Though it's stolen the spotlight for now, PC133 is still viewed largely as
> an interim solution that will fill market demand only until the advent of
> Direct RDRAM, executives agreed.
> "Rambus is going to happen, it's just a question of whether the timing
> fits everyone's cycles," Pleshko said. "Then we have the millennium and
> issues associated with it, and the [Rambus] infrastructure costs. There's
> a lot to look at."
> Rambus chips offer a dramatic 1.6-Gbyte/s peak bandwidth, twice that
> available through today's PC100 SDRAM. However, technical issues have
> caused Intel to push out its Camino chipset launch to September, leaving
> the Rambus chip without a means of communicating with the host processor.
> In the meantime, many component suppliers are rushing to see how much of a
> market they can establish with PC133, a natural follow-on to the PC100
> architecture. PC133 requires little new technology compared with Rambus,
> but increases peak throughput to only 1 Gbyte/s.
> Still, because it's relatively easy to manufacture, suppliers are hoping
> to build up the market, a tactic that could undercut initial demand for
> Direct RDRAM, industry executives said.
> "I previously didn't think PC133 was very serious for the future,"
> McDonald said. "How it looks to me now is that PC133 could be a big speed
> bump for Rambus."
> Unimpressed with PC133's modest performance increase-and concerned that
> support for it could steer the industry away from the Rambus memory
> path-Intel has steadfastly refused to endorse the alternative interface.
> This is significant because Intel owned 80% of the chipset market in 1998,
> according to Mercury Research Corp., Scottsdale, Ariz. While other chipset
> manufacturers-most notably Via Technologies Inc. and Reliance Computer-are
> rolling out their own PC133-compatible chipsets, the absence of an
> Intel-branded part will certainly restrict supply, and sends a cautionary
> message to conservative PC vendors.
> Calling PC133 SDRAM "a distraction," Avo Kanadjian, vice president of
> memory marketing at Samsung Semiconductor, San Jose, said the lack of
> chipset support will limit the memory to server applications. "I don't
> expect any excitement to be created with another synchronous DRAM [PC]
> machine," he said. "I do expect excitement with a Rambus machine."
> Kanadjian said Samsung has run its Direct RDRAM through an Intel-sponsored
> validation program and is beginning customer qualification of volume
> production wafers. Samsung expects to begin manufacturing chips 10 to 12
> weeks following the qualification process, during which initial production
> output levels are determined.
> Despite Direct RDRAM's schedule slip, Samsung sees a market for about 40
> million 128-Mbit Rambus units in 1999, and said it has the capacity to
> meet that entire demand. The last several weeks have resulted in all
> facets of the Rambus supply chain coming together, Kanadjian noted, a
> departure from the industry's previous, more insular development strategy.
> "Clearly, we have a full understanding of everybody's schedule, and we're
> confident that we will keep to schedule, provided that everyone executes
> according to plan," he said. "The last time, we were in the dark and were
> only concerned with our own performance. ... This time around we're very
> much synchronized with the other suppliers, and we're kept very much in
> the process of how our OEMs are doing."
> Most industry analysts agree that Rambus is an inevitability because of
> its sheer performance and narrow, 16-bit bus. And companies such as
> Samsung, which have committed to an early and sustained Rambus ramp, have
> little choice but to follow through with their manufacturing plans.
> However, other DRAM makers, reciting the mantra of "evolution not
> revolution," believe the industry may be better off supporting more
> incremental performance advances. Appearing at a recent PC133 forum that
> Via Technologies sponsored in Taiwan, Micron Technology, Infineon
> Technology (formerly Siemens Semiconductors), and Hyundai Electronics
> Industries outlined their SDRAM roadmaps, which include double-data-rate
> SDRAM and PC133.
> By the end of the year, the groundswell of support for PC133 may prove to
> be overstated. But in the quest for product differentiation, PC makers
> will continue casting about for a 1999 market performer, hedging their
> bets in the event Rambus misses its revised third-quarter launch date.
> "It's all about the race for benchmarks," McDonald said
>
> Rob Schwartz
> 303/221-4144 phone
> 888/794-6854 pager




To: Judy who wrote (21704)4/26/1999 11:27:00 AM
From: Juli  Read Replies (1) | Respond to of 44573
 
Another article from a week ago.

April 19, 1999, Issue: 1156 Section: Business & Finance Rambus' outlook
> marred by Intel delay Andrew MacLellan Rambus Inc. last week said
> earnings for the remainder of 1999 will be flat at best, following word
> earlier this year from development partner Intel Corp. that the
> introduction of its Camino chipset will be delayed until late in the third
> quarter. The Mountain View, Calif., memory-interface designer had hoped
> 1999 would prove a banner year. Having worked with Intel since late 1996
> promoting Direct Rambus DRAM-a high-speed memory technology for use in
> next-generation Pentium-based PCs-Rambus was about to see its ship come
> in. However, the delay of Intel's Camino chipset, which allows the Direct
> RDRAM memory chips to communicate with the processor, has frozen Rambus'
> short-term market opportunity. In a second-quarter earnings report issued
> last week, Rambus said year-to-year revenue increased by 2% for the period
> ended March 31. The company's sales slipped 7%, to $9.9 million, from the
> first quarter, but sales for the six-month period were up 7%, to $20.4
> million, from the same time frame in 1998. With the Camino launch date now
> pushed back to September, Rambus said it expects volume shipments of
> Direct RDRAM to begin in earnest in December, with royalties from
> licensees to follow by March 2000.
>
>
> Rob Schwartz
> 303/221-4144 phone
> 888/794-6854 pager