|
February 7, 2001
|
|
By Jack Robertson
February 2, 2001
Electronic Buyers' News
|
The Federal Trade Commission has launched a preliminary investigation of Rambus Inc. for possible antitrust violations stemming from the company's dealings with the JEDEC standards deliberation body, EBN has learned.
The FTC is looking at whether Rambus violated JEDEC rules by failing to disclose it was pursuing patents for synchronous DRAM, even as the technology was being discussed in an open JEDEC session. Though the FTC does not comment on fact-finding missions, analysts and industry executives involved in the JEDEC deliberations said they have been questioned by FTC investigators.
|
|
By Anthony Cataldo and Rick Merritt
February 5, 2001
TechWeb News
|
Those expecting to get a look under the hood of Intel's next-generation IA-64 processor might be disappointed.
Intel Corp., Santa Clara, Calif., has withdrawn the paper from next week's International Solid-State Circuits Conference
(ISSCC). An Intel spokesman said the company wasn't ready to make a full disclosure. ISSCC requires presenters to provide in-depth detail about their designs, he said.
|
|
By Mark LaPedus
February 6, 2001
Semiconductor Business News
|
Transmeta Corp. here today announced that it is readying a new line of x86-based microprocessors, including a low-power, 0.13-micron MPU that is geared to break the 1-GHz barrier.
The Santa Clara, Calif.-based company will soon roll out the TM5800, an x86-based part that will initially run at speeds of at least 700-MHz or faster, said David
Ditzel, president and chief executive officer of Transmeta.
The TM5800 "will be one of the world's first processors based on a 0.13-micron process," he told the Banc of America Securities technology conference here today.
|
|
By Robert Ristelhueber
February 7, 2001
Electronic Buyers' News
|
IBM Corp., which late last year reversed a decision to use Transmeta Corp.'s Crusoe microprocessor in a notebook computer, will no longer be the exclusive foundry supplier to Transmeta.
Taiwan Semiconductor Manufacturing Co. has been signed by Transmeta as a second source for foundry wafers, according to Transmeta chief executive officer David Ditzel, who declined to say whether IBM would remain its primary foundry.
|
|
|
|
|
February 6, 2001
Semiconductor Business News
|
At the International Solid-State Circuits Conference
(ISSCC) here, Rambus Inc. disclosed details about a new 2.2-gigabyte-per-second interface and other techniques that boost memory performance by more than 25% over today's leading-edge technologies. Los Altos, Calif.-based Rambus today said the new developments have resulted in the world's fastest bus technology.
In one of two ISSCC papers, Rambus engineers described a 2.2-Gbyte/sec. signaling interface for main memory that uses a delay-lock loop (DLL) to enable in-system timing calibration with 1.4-degree resolution and output drivers with limited positive feedback to increase voltage margin. This signaling interface has been used to produce prototype chips operating up to 2.6-Gbyte/sec. in standard chip-scale packages
(CSPs), according to Rambus.
|
|
By Mike Clendenin
February 6, 2001
EE Times
|
The battle over DRAM architectures will amount to little more than a skirmish this year, according to the memory and chip set suppliers gathered Tuesday (Feb. 6) at the DDR Summit Conference. Double-data-rate
(DDR) DRAM will dominate and Rambus will be relegated to a sliver on pie charts that detail market share, they said.
That prediction is to be expected from the conference, sponsored by Via Technologies Inc., a leading proponent of the DDR memory standard. The one-day event attracted more than 40 motherboard, module and memory makers, as well as Advanced Micro Devices Inc. And while Rambus DRAM was initially a de facto taboo topic, it was evident that industry leaders were once again validating the shift toward DDR as the most viable main memory standard in the cutthroat DRAM market.
|
|
By Peter Clarke
February 6, 2001
EE Times
|
The challenge of keeping processors fed with data and maximizing the data rate of network-centric processors and systems were addressed in an International Solid-State Circuits Conference session on serial digital interfaces. The session also showcased efforts by both Intel Corp. and Rambus Inc. to improve serial data specifications.
Presentations at the session varied from an academic paper on an 8-giga-symbol/second transceiver to papers from industry research teams on serial schemes down to 500 Mbit/s per channel.
|
The Register Files
|
|
By Andrew Orlowski
February 5, 2001
The Register
|
McKinley, the second IA-64 processor due next year may have been delayed again, suggests chip analyst Linley Gwenapp. He doesn't offer anything to back up his speculation, but the EE Times reports that Intel has pulled a McKinley technical paper it was scheduled to give to a semiconductor conference that starts in San Francisco today.
Indeed, the IEEE website confirms this. The paper was due to give full disclosure on Itanic II. But we wouldn't jump to the same conclusion as Gwenapp - or at least, not in such haste. The prime sponsor of Itanic: The Next Voyage is Hewlett-Packard, and it decided some time ago that it wouldn't waste its time on the first iteration of Itanic, Merced. As we reported here, in August 1999.
|
|
By Andrew Orlowski
February 2, 2001
The Register
|
Transmeta founder Dave Ditzel says we can expect servers with hundreds of Crusoe CPUs later this year.
But the dedicated, small form factor boxen won't look anything like today's SMP kit - even though Ditzel and partner in crime Gary Stimac (who pioneered Compaq's server business and has launched a start-up to OEM Transmeta servers) were SMP pioneers.
"I don't believe in SMP anymore," says
Ditzel, who helped create the first RISC chips and design the SPARC chip for Sun Microsystems. "I just don't see any reason for people to use it."
|
|
By Tony Smith
February 2, 2001
The Register
|
Infineon has been making up and being friendly with Rambus and Hyundai, settling patent infringement legal action with both semiconductor companies.
So, the German semiconductor operation yesterday withdrew its allegations that Rambus' RDRAM violates its patents. That claim was made in direct response to allegations made by Rambus against Infineon that the latter's SDRAM products rip off Rambus intellectual property, assertions Infineon denied.
|