Stream formats is the standard output format for GDSII data. The following picture shows the data architecture of GDSII。Stream records are always an even number of bytes. If a record contains ASCII string data and the ASCII string is an odd number of bytes, the last character is a null character.
Stream format is upward compatible, which means newer releases of the Stream translators can read libraries produced with an older release.


Common sense: 1 word=2 byte
1byte=8bit

X(HEX)=XXXX(BIN)

There are two parts in this report which one describe the data architecture, another one describe the syntax of GDSII.



Figure 1. Data
Architecture
of
GDSII

The table below shows the data type:


DATA TYPE

VALUE

No data present	0
Bit Array	1
Two-Byte Signed Integer	2
Four-Byte Signed Integer	3
Four-Byte Real	4 (not used)
Eight-Byte Real	5
ASCII String	6

Bit Array (1)
A bit array is a word which contains bits or group of bits that
represent data. A bit array allows one word to contain more than
one piece of information.

Two-Byte Signed Integer (2)
2-byte integer = 1 word 2s-complement representation
The range of two-byte signed integers is -32,768 to 32,767.
The following is a representation of a two-byte integer, where S
is the sign and M is the magnitude.

SMMMMMMM MMMMMMMM

00000000 00000001 = 1

00000000 00000010 = 2

00000000 10001001 = 137

11111111 11111111 = 1

11111111 11111110 = -2

11111111 01110111 = -137

Four-Byte Signed Integer (3)
4-byte integer = 2 word 2s-complement representation
The range of four-byte signed integers is -2,147,483,648 to
2,147,483,647.
The following is a representation of a four-byte integer, where S
is the sign and M is the magnitude.

SMMMMMMM MMMMMMMM MMMMMMMM MMMMMMMM

Because of the scalar of record type ,I can not enumerate one by one. In my opinion, a specify record type looks like tied to a data type changeless. For instance, the record type HEADER , it was represent by 0002 in GDSII, and the first two numbers 00 are the record type of HEADER, the last two numbers 02 are the data type. The four numbers is always together for representing HEADER. The others record types are the same.
Now I show you two examples here to explain what a record means.

000C 1206 6578 616D 706C 6531
This record is 12 (C hex) bytes. It is the SNAME (12 hex) record. The data type is an ASCII string (06). This record contains the name of referenced structure EXAMPLE1.

0024 1003 0000 1388 0000 6D60 0000 2EE0 0000 6D60
0000 1F40 0000 84D0 0000 1388 0000 6D60
This is XY record. The coordinates are (5, 28), (12, 28),
(8, 34), and (5, 28).




Stream Syntax:

Symbol	Symbol	Meaning
Double colon	: :	“Is composed of”
Square brackets	[ ]	An element that can be absent or occur one time.
Braces	{ }	One of the elements within the braces can occur.
Braces with an	{ }*	The elements within the braces
asterisk		can be absent or occur one or
		more times.
Braces with a plus	{ }+	The elements within braces must occur one or more times.
Angle brackets	< >	These elements are further defined in the Stream syntax list.
Vertical bar	|	“Or”

这是我自己总结的点关于LAYOUT的知识，好久没发点东西了，有心打算在IC方面发展的可以看下。