Assembly language, commonly called assembly or asm, is a human-readable notation for the machine language that a specific computer architecture uses. Machine language, a pattern of bits encoding machine operations, is made readable by replacing the raw values with symbols called mnemonics.
For example, a computer with the appropriate processor will understand this x86/IA-32 machine language:
For programmers, however, it is easier to remember the equivalent assembly language representation:
mov al, 061h
which means to move the hexadecimal value 61 (97 decimal) into the processor register with the name “al”. The mnemonic “mov” is short for “move”, and a comma-separated list of arguments or parameters follows it; this is a typical assembly language statement.
Transforming assembly into machine language is accomplished by an assembler, and the reverse by a disassembler. Unlike in high-level languages, there is usually a 1-to-1 correspondence between simple assembly statements and machine language instructions. However, in some cases an assembler may provide pseudoinstructions which expand into several machine language instructions to provide commonly needed functionality. For example, for a machine that lacks a “branch if greater or equal” instruction, an assembler may provide a pseudoinstruction that expands to the machine’s “set if less than” and “branch if zero (on the result of the set instruction)”.
Every computer architecture has its own machine language, and therefore its own assembly language. Computers differ by the number and type of operations they support. They may also have different sizes and numbers of registers, and different representations of data types in storage. While all general-purpose computers are able to carry out essentially the same functionality, the way they do it differs, and the corresponding assembly language must reflect these differences.
In addition, multiple sets of mnemonics or assembly-language syntax may exist for a single instruction set. In these cases, the most popular one is usually that used by the manufacturer in their documentation.
Instructions in assembly language are generally very simple, unlike in a high-level language. Any instruction that references memory (for data or as a jump target) will also have an addressing mode to determine how to calculate the required memory address. More complex operations must be built up out of these simple operations. Some operations available in most instruction sets include:
- set a register (a temporary “scratchpad” location in the CPU itself) to a fixed constant value
- move data from a memory location to a register, or vice versa. This is done to obtain the data to perform a computation on it later, or to store the result of a computation.
- read and write data from hardware devices
- add, subtract, multiply, or divide the values of two registers, placing the result in a register
- perform bitwise operations, taking the conjunction/disjunction (and/or) of corresponding bits in a pair of registers, or the negation (not) of each bit in a register
- compare two values in registers (for example, to see if one is less, or if they are equal)
affecting program flow
- jump to another location in the program and execute instructions there
- jump to another location if a certain condition holds
- jump to another location, but save the location of the next instruction as a point to return to (a call)
Some computers include one or more “complex” instructions in their instruction set. A single “complex” instruction does something that may take many instructions on other computers. Such instructions are typified by instructions that take multiple steps, may issue to multiple functional units, or otherwise appear to be a design exception to the simplest instructions which are implemented for the given processor. Some examples of such instruction include:
- saving many registers on the stack at once
- moving large blocks of memory
- complex and/or floating-point arithmetic (sine, cosine, square root, etc.)
- performing an atomic test-and-set instruction
- instructions that combine ALU with an operand from memory rather than a register
A form of complex instructions that has become particularly popular recently are SIMD operations that perform the same arithmetic operation to multiple pieces of data at the same time, which have appeared under various trade names beginning with MMX and AltiVec.
The design of instruction sets is a complex issue, with a simpler instruction set (generally grouped under the concept RISC) perhaps offering the potential for higher speeds, while a more complex one (traditionally called CISC) may offer particularly fast implementations of common performance-demanding tasks, may use memory (and thus cache) more efficiently, and be somewhat easier to program directly in assembly.
Assembly language directives
In addition to codes for machine instructions, assembly languages have extra directives for assembling blocks of data, and assigning address locations for instructions or code.
They usually have a simple symbolic capability for defining values as symbolic expressions which are evaluated at assembly time, making it possible to write code that is easier to read and understand.
Like most computer languages, comments can be added to the source code; these often provide useful additional information to human readers of the code but are ignored by the assembler and so may be used freely.
They also usually have an embedded macro language to make it easier to generate complex pieces of code or data.
In practice, the absence of comments and the replacement of symbols with actual numbers makes the human interpretation of disassembled code considerably more difficult than the original (high level) source would be.
Usage of assembly language
Historically, a large number of programs have been written entirely in assembly language. A classic example was the early IBM PC spreadsheet program Lotus 123. Even into the 1990s, the majority of console video games were written in assembly language, including most games written for the Sega Genesis and the Super Nintendo Entertainment System. The popular arcade game NBA Jam (1993) was also coded entirely using assembly language.
There is some debate over the continued usefulness of assembly language. It is often said that modern compilers can render higher-level languages into codes that run as fast as hand-written assembly, but counter-examples can be made, and there is no clear consensus on this topic. It is reasonably certain that, given the increase in complexity of modern processors, effective hand-optimization is increasingly difficult and requires a great deal of knowledge.
However, some discrete calculations can still be rendered into faster running code with assembly, and some low-level programming is actually easier to do with assembly. Some system-dependent tasks performed by operating systems simply cannot be expressed in high-level languages. In particular, assembly is often used in writing the low level interaction between the operating system and the hardware, for instance in device drivers. Many compilers also render high-level languages into assembly first before fully compiling, allowing the assembly code to be viewed for debugging and optimization purposes.
It’s also common, especially in relatively low-level languages such as C, to be able to embed assembly language into the source code with special syntax. Programs using such facilities, such as the Linux kernel, often construct abstractions where different assembly language is used on each platform the program supports, but it is called by portable code through a uniform interface.
Many embedded systems are also programmed in assembly to obtain the absolute maximum functionality out of what is often very limited computational resources, though this is gradually changing in some areas as more powerful chips become available for the same minimal cost.
Another common area of assembly language use is in the system BIOS of a computer. This low-level code is used to initialize and test the system hardware prior to booting the OS and is stored in ROM. Once a certain level of hardware initialization has taken place, code written in higher level languages can be used, but almost always the code running immediately after power is applied is written in assembly language.
Assembly language is also valuable in reverse engineering, since many programs are distributed only in machine code form, and machine code is usually easy to translate into assembly language and carefully examine in this form, but very difficult to translate into a higher-level language. Tools such as the Interactive Disassembler make extensive use of disassembly for such a purpose.
MenuetOS, a floppy-based operating system with a fully functional GUI is written entirely in assembly. A 64bit version is also available. The author claims that only through assembly language could he produce his system in less than 1.4 megabytes.
A utility program called an assembler is used to translate assembly language statements into the target computer’s machine code. The assembler performs a more or less isomorphic translation (a one-to-one mapping) from mnemonic statements into machine instructions and data. This is in contrast with high-level languages, in which a single statement generally results in many machine instructions.
Many sophisticated assemblers offer additional mechanisms to facilitate program development, control the assembly process, and aid debugging. In particular, most modern assemblers include a macro facility (described below), and are called macro assemblers.
Typically a modern assembler creates object code by translating assembly instruction mnemonics into opcodes, and by resolving symbolic names for memory locations and other entities. The use of symbolic references is a key feature of assemblers, saving tedious calculations and manual address updates after program modifications. Most assemblers also include macro facilities for performing textual substitution—e.g., to generate common short sequences of instructions as inline, instead of called subroutines.
Assemblers are generally simpler to write than compilers for high-level languages, and have been available since the 1950s. Modern assemblers, especially for RISC architectures, such as SPARC or POWER, as well as x86 and x86-64, optimize Instruction scheduling to exploit the CPU pipeline efficiently.
Number of passes
There are two types of assemblers based on how many passes through the source are needed to produce the executable program.
* One-pass assemblers go through the source code once and assume that all symbols will be defined before any instruction that references them.
* Two-pass assemblers create a table with all symbols and their values in the first pass, then use the table in a second pass to generate code. The assembler must at least be able to determine the length of each instruction on the first pass so that the addresses of symbols can be calculated.
The advantage of a one-pass assembler is speed, which is not as important as it once was with advances in computer speed and abilities. The advantage of the two-pass assembler is that symbols can be defined anywhere in program source code. This lets programs be defined in more logical and meaningful ways, making two-pass assembler programs easier to read and maintain.
More sophisticated high-level assemblers provide language abstractions such as:
* Advanced control structures
* High-level procedure/function declarations and invocations
* High-level abstract data types, including structures/records, unions, classes, and sets
* Sophisticated macro processing (although available on ordinary assemblers since late 1950s for IBM 700 series and since 1960’s for IBM/360, amongst other machines)
* Object-oriented programming features such as classes, objects, abstraction, polymorphism, and inheritance.