My work and vacation

A long long term of running toward deadlines, which came one after another while I was running, was over.

I write books von Beruf.
In fact, that brings not so much income that I could live on.
Because I am a housewife and it’s a part time job.
In fact, the above reason is not correct. I have no children and my husband doesn’t want me a complete householding, So I can a pretty amount of time for writing.
I should say, therefore, that I have not yet enough talent to live on it.
Perhaps it is quite a japanese way to making this kind of explanation, which no one would care…

I write for only one publisher, which runs a monthly magazine and books.
For a book, the editor gives me a topic and I make a plan of what to write.
Every book has 200-300 pages.
It is required to write one in 2-3 months. The exact schedule is on my decision. But, once I told it to the editor, I have to meet the deadline, for he make his own schedule after I send him the first draft.
I usually set a period for one book within two months, for my work is targeted to beginners with a simple and friendly explanation, rather than deep contemplation or detailed inspection.
On the other hand I write an article or more for the monthly magazine of the same publisher. It takes a day or two (for three pages), sometimes a week (for more), to write one.

Total amount of work comes to a good amount for me.

In addition, I go to university once a week for an adult education. It is not a special education system taylored for adults, which had long been apart from lectures, but the very lectures which the students take. So sometimes it takes me a lot of time to prepare for it.

Since last autumn I wrote four books – Ruby on Rails, iMovie, Android programming and the one which is not to be announced before it is published. Articles for the magazine, which is simply uncountable. I took 6 lectures, all of which were so precious.

And now, I’m going to be in vacation, that is, I will not get a new topic from my editor for a while, because he will take his own vacation.

What I’m going to do then?
At first, I will learn English more.


Change in LSI design with growing integration (2)

In the year 1990s, requirements of transistors for a single LSI chip increased to 100 thousands – a million. For one enginner, who could draw 40 transistors in a day,it would have take 250 years to build up such a chip. It was unrealistic.

So LSI design came to be drawn based on logic gates. A symbol of 2-input NAND represented 4 transisistors. Then logic gates with 3 or more inputs were used to reduce components on drawings.

Arround the year 2000 transistors used amounted over 10 millions.
Then computer programming were introduced to draw such complicated patterns.
This mothod is called “RT-level design.” R for register and T for transfer.

“This word, RT-level, is a common sense in this field of industry” said our professor, “If one uses the word RT-level, everyone related to LSI design can imagine what is it. He must do”

Change in LSI design with growing integration(1)

Scince Moore’s law has been a goal that people must reach to, ways of LSI desing have been developed according to number of transistors to be integrated.

In 1980 – Full cutsom.
A whole LSI chip was built up from scratch by a factory.

This is a very simple schematic figure of a pattern for a single transistor.

Those days the designers wrote circuits by hand on drawing desks.
One designer could write 40 of the above pattern in one day.
Those days an LSI was composed of 10,000 transistors, so one would have needed 250 days to complete the design. Of course, many people shared the work to shorten the time needed.

Moore’s law

For evolution of LSIs there is the famous Moore’s law, which tells that integration of LSI (number of transistors installed on one chip) grows to be quadruple every three years.
This Moore’s law is not a natural law, e.g. the integration grows not spontaneously. It is a requirement, which every venders should meet to keep his product always successful. Or it is a standard to evaluate the quality of a technology. One could say for example, “Vendor A is enough productive, because his products always meet Moore’s law,” or “Your design for the new LSI is useless, because it can’t satisfy Moore’s law!”

It could be called rather “Moore’s wish” or “Moore’s psychic power” – so our professor says.

Aspects of LSI design

Now, I have almost finished to write my new book and regained tome to do various things. I would like to try again to express in English what I have learned in the University.

There are three types of aspects for how to design a system LSI chip.

(1)Low cost: Such things as DRAMs or flash memories need to be so cheaper as possible. Costs of semiconductor chips mainly depend on size. Usually a 5mmx5mm chip costs 1000 yen.
It would take a year with 100 workers to design a chip with 10% smaller than a previous, but it would gain more 10 billion yen per year.

(2)Processing power: For Microprocessor “spec” would mean prosessing power, e.g. clock frequency, in Japan or possibly in America. For example, it took a year with 1000 workers to increase a frequency from 1.9GHz to 2GHz.

(3)Flexibility: a typical case for it is that of Japanese mobile phones, in which it is required at first to release new models every three months. It often occurs that an LSI factory has to start to design without any information of required specs for the new model.

In addition to those above, this year a new aspect has appeared – power saving. Although some investigators has worked for it, this aspect had not been payed attention.
It is said to be enable to reduce power consumption to 1/100 with hardware and to 1/2 with software improvement. Though software improvement has less effect, it can be started immediately.

Non-periodic flow(1)

Yesterday I spend all day long to read an article written in English. The very title of it I cannot remember, because I have already put it in a bag , not to forget to take to the University on Thursday. It was an article of “The Journal of Atmospheric Science”. It explained that behaviors of hydrodynamic flow, which are seen in atmospheric phenomena, are so unstable that one cannot predict them during a long period. This article is said to be the first to describe a chaotic behavior in a mathematical way.
Precisely, hydrodynamic flow in atmosphere were in the article classified as non-periodic flow with loss of its energy.