Slashdot Log In
IBM to Unveil Major Tech Advances
Posted by
Hemos
on Thu Dec 02, 1999 01:55 PM
from the all-sorts-of-new-toys dept.
from the all-sorts-of-new-toys dept.
mr wrote to us to point out an article on IBM in today's SF Chronicle. IBM, starting on Monday at the Internation Electron Device Meeting, will be disclosing eighteen new inventions coming out of their labs. IBM goes to so far to say that it will keep Moore's Law [?] around for at least another decade. The article also talks about some of IBM's recent advancements as well as describing some of the new stuff to be unveiled.
This discussion has been archived.
No new comments can be posted.
IBM to Unveil Major Tech Advances
|
Log In/Create an Account
| Top
| 165 comments
(Spill at 50!) | Index Only
| Search Discussion
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
(1)
|
2
(1)
|
2
Broken link (Score:4)
-----
Advances Today (Score:4)
Motorola today said that they had found a way to make transistors 4 times smaller and be implementable in a short time-span. It reduces power consuption and allows for speed increases. This was on your favourite site [theregister.co.uk].
The trench technology looks cool, burying the DRAM under the processor so it doesn't have to be next to it. That should increase yield whilst not compromising on capabilities of the processor. How much DRAM can you fit in 100mm^2? 400mm^2? That would be the amount of 2nd or 1st level cache your Athlon/Alpha processor could have built in, running at full speed!
Faster CPUs aren't what we need (Score:3)
Mike Eckardt [geocities.com] meckardt@yahoo.spam.com
Moore's Law Predictions (Score:4)
Go Big Blue! (Score:4)
Looks like all the money IBM has historically put into general R&D is paying off once again. That's one thing Big Blue has usually gotten right....and something other large tech firms can learn from. Fund your scientists, and don't necessarily expect products immediately from them. Let them do basic research, and the products will follow.
-- JackCat
OT: Use of moderation descriptors (Score:3)
This is similar to calling the "first post" messages "Flamebait" (well, maybe hot grits and petrified are flamebait). They're not. Trolls at best.
The above is NOT a ulterior plea to be moderated up by folks who want to prove how "fair" they are. I moderate every week or two, just like a lot of you, and I am NOT looking for extra karma points -- each of my posts stands on its own merits (good, bad, or otherwise
Re:Broken link yet again (Score:5)
All the standard benefits of Open Source (bug cleaning, extra features like URL checking, &c.)
The warm fuzzy of knowing that /. is putting its money where its mouth is. IOW, as perhaps the discussion site for Open Source, /. would do well to directly and concretely support the ideals we all support as a community.
So how about it, CmdrTaco? How about letting us take a crack at it?
Ya think in 20 years Microsoft will do the same? (Score:5)
Now what I'm really wondering is this : at least one theory suggests that the government is in the process of doing to MS what it did to IBM back in the 80's. If that's true, and the DOJ keeps MS so tangled up over the next decade that competitors emerge, does anybody think that Microsoft will reinvent itself in a similar way? Sure, we can all hate MS as the big bad corporate enemy now, but we all did that 20 years ago, too, when it was IBM. Now we love them.
IBM and ..... (Score:3)
Probably wrong, but fun to speculate nonetheless.
Re:Faster CPUs aren't what we need (Score:4)
intersting work is being done in this direction under the Processor-in-Memory (PIM) project. [nd.edu]
Another mechanism to decrease the effect of this memory latency is to use large numbers of low-level threads (often automatically generated by the compiler) to mask latency. By decreasing the context switch penalty to a single cycle (or less with interwoven threads) and then switching on every cache miss substantial benifits can be made. One example of this is Tera computing MTA [tera.com] architechture. For certain common simulation tasks the 4-processor TERA machine blew away a multi-node Origin and Cray computer according to This NASA report [esgeroth.org].
Also, Sun's new MAJC [arstechnica.com] architechture uses threads to mask latency.
Interwoven threads (where the processor switches thread every clock cycle) has the benifits of removing branch and data dependancies from a processor pipleine, thus removing the need for processor complexity like data forwarding, speculative execution, and the like. An example of this technique can be found a the TIPSI Project [nd.edu].
Way to go IBM (Score:4)
These are all nice new technologies, but let' s hope IBM knows how to use them. Historically IBM has created quite a few technologies; unfortunately the management has, in the past, simply thrown the innovations away. Here are a couple examples:
On the other hand, many IBM innovations did make it, such as the magnetic hard drive. IBM still makes great hard drives.
The article didn't say it, but allowing RAM transistors to exist below other circuitry effectively doubles the data density of existing DRAM chips. How does tens of gigabytes of RAM sound? The problem: how do you cool double the amount of heat coming from those RAM chips?
I am looking forward to the faster and better computers and devices that will come from these innovations.
--------------------------------------
OT: Use of moderation descriptors (Score:3)
I agree. I think that the "overrated" and "underrated" should be used to simply mark a post up or down. The others should be used only if they're descriptive. Also, I wish that the term "Troll" was not a negative score. I'm not saying it should be positive, but not all Trolls are bad.
This is similar to calling the "first post" messages "Flamebait" (well, maybe hot grits and petrified are flamebait). They're not. Trolls at best.
Maybe there should be a "First Post Moron" descriptor.
---
Re:Wonderful IBM Commercials (Score:5)
:) True. I'd probably be more worried about it if I wasn't surrounded by people who are very interested in that sort of thing. It's a fact of life that industry wants profiling information.
At least IBM is taking the right approach -- "We want profiling information so that we can help streamline the information we're providing to you." If I know you play tennis, there's two ways you can look at it. One is, "Oooo good, now I can sell him more tennis balls." Everybody hates this, of course, because nobody likes to feel like a target. But the second is, "Hey, you know what? Maybe I really am interested in knowing whose got a deal on tennis balls." Sometimes targeted messaging does actually work. It's really the same thing that the demographics have always been, only with better profiling they really know. They're not assuming "Oh, because you're in group X, there's a Y% likelihood that you play tennis."
There's a new movement in this area. That's to get away from the use of the word "targeting" and to start making use of expressions like "1:1" and "relationship". People are happier having a relationship with the businesses they use. The whole point of the IBM focusgroup commercial is a bunch of people being pissed off because the ad people don't know them.
And just in case anybody is prepared to argue that "1:1 relationship" is just new marketing hype for the same old spam, let me put it this way. When my grandfather walked into the local hardware store, the shopkeeper could say "Hello, Dan! Getting ready to send the kids off to college pretty soon, aren't ya? Got a good sale on bookcases down in aisle 3." And he would never, ever say "Have ya seen our sale on house paint?" if he knew that my dad had aluminum siding. And service like that was *appreciated*. People go on to the internet today and they ask where all the service went. The optimist in me says that all this 1:1 relationship stuff is actually a way to try and bring that *back*. If I really thought that I was just coming up with a better mousetrap (or in this case, spamtrap), I don't think I'd be working where I work.
d
Re:OT: Use of moderation descriptors (Score:3)
then everyone who replies to it should automatically be moderated
down...
IBM's Cunning Plan Revealed... (Score:5)
"I'm worried. We have those press announcements on Monday, and you -know- Slashdot'll cover them. Our servers will never cope!"
"It's ok. I've just put up some web pages, pre-announcing the announcements. If the servers melt down this week, we'll still have the weekend to replace them, and Slashdot readers don't care about repeat announcements."
"That's cunning! Do you think they'll fall for it?"
"I think so. The system load was showing 490% CPU usage, and rising fast, the last time I looked."
* In the distance, the sound of a hard disk spinning out of the drive bay and colliding with a UPS unit. An IL&M techie is on-hand to supply the effects *
* Outside the building, the T1 link is glowing red, then blue, before finally exploding as the energy from the packets causes the fibre to undergo nuclear fusion. *
Re:heat dissipation (Score:3)
In fact reducing the number of off-die memory accesses may reduce the power (no need to source/sink to those external bus signal's caps)
The more complex anything gets - the lower the yield (basic rule of nature - I suspect it applies to life too :-)
Re:How many other companies can do... (Score:3)
How about the transistor? Or the laser? Or Information Theory? The solar cell, and perhaps the communications satellite? Cellular phones? How about the first photonic computer? How about Radio Astronomy including the Big Bang remnant radiation? The application of statistics to the social sciences? Or waveguide optical fibre (making transoceaninc optical cables possible).
AT&T had the best industrial labs in the world before they spun them off as Lucent. No other lab in the world has come close to contributing as much. Eleven workers at Bell Labs have been awarded Nobel Prizes.
Schawlow and Townes Invent the Laser
The invention of the laser, which stands for light amplification by stimulated emission of radiation, can be dated to 1958 with the publication of the scientific paper, Infrared and Optical Masers, by Arthur L. Schawlow, then a Bell Labs researcher, and Charles H. Townes, a consultant to Bell Labs. That paper, published in Physical Review, the journal of the American Physical Society, launched a new scientific field and opened the door to a multibillion-dollar industry.
The work of Schawlow and Townes, however, can be traced back to the 1940s and early 50s and their interest in the field of microwave spectroscopy, which had emerged as a powerful tool for puzzling out the characteristics of a wide variety of molecules. Neither man was planning on inventing a device that would revolutionize a number of industries, from communications to medicine. They had something more straightforward in mind, developing a device to help them study molecular structures.
The beginnings at Bell Labs Townes, armed with a Ph.D. degree in physics from the California Institute of Technology, joined Bell Labs in 1939, where he worked on a variety of problems, including microwave generation, vacuum tubes, and magnetics. He then moved on to solid-state physics, studying electron emissions from surfaces. One day, about a year after Townes arrived at Bell Labs, Mervin Kelley, then director of Townes' laboratory, informed the group, "On Monday, I want you to start a radar bombing system." Townes wasn't enthusiastic about the assignment, but realized that World War II had invaded the quiet hallways of Bell Labs. "We worked at it pretty hard, and after about a year we had a system which we put in an airplane, and actually used. It worked.
For those who are interested in the history of the first transistor, here is
an excerpt from "The TRANSISTOR - A Crystal Triode" by Fink and Rockett in
ELECTRONICS 21, 68-71 (Sep 1948), describing the work at the Bell Telephone
Laboratories:
"Although investigation of semiconductors at BTL dates back a number of years,
with the end of the war a concentrated basic research progrm was undertaken."
"The group on semiconductors, led by William Shockley, one of this country's
leading solid-state physicists, was seeking answers to three basic questions:
(1) physically, what is a semiconductor, (2) how doers its physical nature
produce its observed properties, and (3) how does the fabrication and
processing of the material affect its physical nature? Among the
semiconductors studied were silicon, copper oxide, and germanium."
"A great deal of empirical information had been amassed on these substances
during their use, particularly as detectors in microwave equipment ("Crystal
Rectifiers", H.C. Torrey and C.A. Whitmer, Mcgraw-Hill, 1948). In particular
it was known that their resistivities were determined chiefly by impurities,
and furthermore that their resistivities could be varied over wide ranges by
applying various external influences (light in the case of photocells,
electric potential in the case of rectifiers and detectors, or temperature
in the case of Thermistors)."
"Likewise, a high potential applied externally (without making contact) to
a semiconductor should change its resistivity. Using a sheet of germanium
as one plate of a capacitor, Shockley and his colleagues measured the change
in resistance produced by changing the voltage across the capacitor. The
change in resistance was much smaller than anticipated in the light of
prevailing theory. Conclusion: something wrong with theory. So John Bardeen,
a theoretical physicist in the group, devised a theory of surface states that
would account for the measured change as well for older known effects
unexplained by previous theories."
"The new theory suggested new experiments, which, when performed, called for
refinements in the theory. While W.H. Brattain and John Bardeen were following
up the consequences of the refined theory of surface states they invented the
Transistor."
The discovery was made in December 1947, but not announced to the world at
large until July, 1948, after additional devices has been fabricated and
tested.
-----------------------------------------------
Further details are given by J. Bardeen and W.H. Brattain themselves in
their introduction to "Physical Principles Involved in Transistor Action",
published simultaneously in BELL SYSTEM TECHNICAL JOURNAL 28, 239-277
(Apr 1949) and PHYSICAL REVIEW 75, 1208-1225 (1949) as follows:
"The properties of germanium as a semi-conductor and as a rectifier have
been investigated by a group working under the direction of K. Lark-Horovitz
at Purdue University. Work at the Bell Telephone Laboratories
was initiated by R.S. Ohl before the war in connection with the development
of silicon rectifiers for use as detectors at microwave frequencies. Research
and development on both germanium and silicon rectifiers during and since the
war has been done in large part by a group under J.H. Scaff. The background
of information obtained in these various investigations hs been invaluable."
[A summary of the wartime weapons research can be found in "Development of
Silicon Crystal Rectifiers for Microwavve Radar Receivers" by Scaff and Ohl,
BELL SYSTEM TECHNICAL JOURNAL 26, 1-30 (Jan 1947)]
"The general research program leading to the transistor was initiated and
directed by W. Shockley. Work on germanium and silicon was emphasized because
they are simpler to understand than most other semi-conductors. One of the
investigations undertaken was the study of the modulation of conductance of a
thin film of semi-conductor by an electric field applied by an electrode
insulated from the film. [described in "Modulation of Conductance by Surface
Charges" by Shockley and Pearson, PHYSICAL REVIEW 74, 232 (July 15, 1948)]
If, for example, the film is made one plate of a parallel plate condenser,
a charge is induced on the surface. If the individual charges which make up
the induced charge are mobile, the conductance of the film will depend on
the voltage applied to the condenser. The first experiments performed to
measure this effect indicated that most of the induced charge was not mobile.
This result, taken along with other unexplained phenomena such as the small
contact potential difference between n- and p- type silicon and the
independence of the rectifying properties of the point contact rectifier on
the work function of the metal point, led one of the authors [Bardeen,
"Surface States and rectification at metal semiconductor contact", PHYSICAL
REVIEW 71, 717-727 (1947)] to an explanation in terms of surface states.
This work led to the concept that space charge barrier layers may be present
at the free surfaces of semi-conductors such as germanium and silicon,
independent of a metal contact. Two experiments immediately suggested were
to measure the dependence of contact potential on impurity concentration
and to measure the change of contact potential on illuminating the surface
with light. Both of these experiments were successful and confirmed the
theory. [Brattain & Shockley, PHYSICAL REVIEW 72, 345L (1947)] It was while
studying the latter effect with a silicon surface immersed in a liquid that
it was found that the density of surface charges and the field in the space
charge region could be varied by applying a potential across an electrolyte
in contact with the silicon surface. While studying the effect of field
applied applied by an electrolyte on the current voltage characeristic of
a high-back-voltage germanium rectifier, the authors were led to the concept
that a portion of the current was being carried by holes flowing near the
surface. Upon replacing the electrolyte with a metal contact transistor
action was discovered."
"The germanium used in the transistor is an n-type or excess semi-conductor
with a resistivity of the order of 10 ohm-cm, and is the same as the material
used in high-back-voltage germanium rectifiers." ["Preparation of High Back
Voltage Germanium Rectifiers" by J.H. Scaff and H.C. Theuerer, NATIONAL
DEFENSE RESEARCH COMMITTEE 14-555 (Oct 24, 1945)]
"Our discussion has been confined to the transistor in which two point
contacts are placed in close proximity on one face of a germanium block.
It is apparent that the principles can be applied to other geometrical
designs and to other semi-conductors. Some preliminary work has shown
that transistor action can be obtained with silicon and undoubtedly other
semi-conductors can be used."
-----------------------------------------------
Later that year, William Shockley extended the theory from simple metal-
semiconductor junctions to full semiconductor-semiconductor junctions,
and suggested that the behavior of a transistor made from such junctions
would be easier to predict than that of the point-contact transistors
made so far. From the introduction to "The Theory of p-n Junctions in
Semiconductors and p-n Junction Transistors", BELL SYSTEM TECHNICAL JOURNAL
28, 435-489 (July 1949):
"As is well known, silicon and germanium may be either n-type or p-type
semiconductors, dependig on which of the concentrations Nd of donors or
Na of acceptors, is the larger. If, in a single sample, there is a
transition from one type to the other, a rectifying photosensitive p-n
junction is formed. The theory of such junctions is in contrast to those
of ordinary rectifying junctions because, on both sides of the junction,
both electron flow and hole flow must be considered. In fact, a major
portion of the hole current may persist into the n-type region and vice-
versa. In later sections we show how this feature has a number of
interesting consequences"
"A p-n junction may act as an emitter in the transistor sense, since it
can inject hole current into n-type material."
"The p-n-p transistor has the interesting feature of being calculable
to a high degree. One can consider such questions as the relative ratios
of width to length of the n-region and the effect of altering impurity
contents and scaling the structure to operate in different frequency
ranges."
-----------------------------------------------
With the assistance of several others (Morgan, Sparks, & Teal), Shockley
went on to produce a working p-n junction transistor in 1951, and in 1952
he went on to develop the theory behind the field-effect transistor, which
is the type most commonly used today in computer chips. ["A Unipolar
Field-Effect Transistor", PROCEEDINGS OF THE I.R.E. 40, 1365+ (Nov 1952)]
-George Fergus