Usually, x86 tutorials don’t spend much time explaining the historical perspective of design and naming decisions. When learning x86 assembly, you’re usually told something along the lines: Here’s
EAX . It’s a register. Use it.
So, what exactly do those letters stand for? E–A–X.
I’m afraid there’s no short answer! We’ll have to go back to 1972…
In 1972, after an odd sequence of events, Intel introduced the world’s first 8-bit microprocessor, the 8008 . Back then, Intel was primarily a vendor of memory chips. The 8008 was commissioned by the Computer Terminal Corporation (CTC) for their new Datapoint 2200 programmable terminal. But the chip was delayed and did not meet CTC expectations. So Intel added a few general-purpose instructions to it and marketed the chip to other customers.
8080 had seven 8-bit registers:
A stood for accumulator , which was an implicit operand and return value of the arithmetic and logical operations.
You might think— gee, seven is a very odd number of registers —and would be right! The registers were encoded as three bits of the instruction, so it allowed for eight combinations. The last one was for a pseudo-register called
M . It stood for memory .
M referred to the memory location pointed by the combination of registers
H stood for high byte, while
L stood for low byte of the memory address. That was the only available way to reference memory in 8008.
A was an accumulator,
L were also used for addressing memory. However,
E were completely generic and interchangeable.
In 1979, Intel was already a microprocessor company, and their flagship processor iAPX 432 is delayed. So as a stop-gap measure, they introduce 8086, a 16-bit microprocessor derived from 8080, which was itself derived from 8008.
To leverage its existing customer base, Intel made 8086 software-compatible down to 8008. A simple translator program would translate from 8008 assembly to 8086 assembly. For that to work well, 8086 instruction set architecture had to map well to 8008, inheriting many design decisions.
8086 had eight 16-bit registers and eight 8-bit registers, and they overlapped as follows:
8086 instructions had a bit flag that specified whether the three-bit encoding of a register referred to one of eight 8-bit registers, or to one of eight 16-bit registers.
As you can see from the figure above, data in the first four 16-bit registers could also be accessed by one of the eight 8-bit registers.
AX was an e X tended 16-bit accumulator, while
AL could be thought of as 8-bit registers on their own or as a way to access the H ight and the L ow bytes of
Since 8008 had seven 8-bit registers, they could be mapped well to the eight 8086 registers, with one to spare.
M pseudo-register was not needed anymore since 8086 allowed for many memory addressing modes. Hence, it freed an encoding for an additional register.
In the following figure you can see how exactly the 8008 registers were mapped to the 8086 ones:
Even though many arithmetic and logical operations could work on any of these registers, none of the registers were truly generic at this point. Each had some instructions introduced that worked for one of the registers but didn’t work for others. The mnemonics are:
BX is base register ,
CX is count register ,
DX is data register , and
AX is still the accumulator .
SP is stack pointer ,
BP is base pointer ,
SI is source index ,
DI is destination index . But we won’t go into details about them here.
8086 also introduced the segment registers, but they were very much a separate beast. Segmented architecture deserves a story on its own, as it is the result of maintaining backward-compatibility with 8080.
It’s 1979, and by all accounts, 8086 instruction set architecture is already a mess.
In 1985 Intel introduced 80386, the first 32-bit processor in the x86 line. An early batch of processors had a defect in one of the 32-bit operations. They were marked as 16-BIT S/W ONLY and sold anyway.
Many new features were introduced, but 80386 continued to be (mostly) binary-compatible down to 8086.
The main registers were E xtended to 32 bits by adding an
EAX is now the E xtended e X tended A ccumulator.
AX now refers to the lower half of
EAX , while
AL continue to refer to the two
And that’s how
EAX got its name.
But wait, there’s more to the story!
In 2003 AMD effectively takes over the architectural leadership and introduces the first 64-bit processor in the x86 lineage. In legacy mode , it is backward-compatible down to 8086.
The eight main registers are extended to 64 bits.
The extended registers get an
R prefix that replaces the
E prefix. So the accumulator is now referred to as
Well, AMD wanted to streamline the register handling. They introduced eight new registers called
R15 . They even discussed calling the extensions of the existing eight registers as
R7 . But they recognized that many instructions have mnemonics that refer to one of the register letters like
B . So they kept the original names, replacing
R . That also provided at lease some consistency with the new
The new registers also got their “narrow” versions. Take
R15 , for example:
And that, folks, was a quick history of x86 accumulator! From an 8-bit
A of 8008, to 16-bit
AX of 8086, to 32-bit
EAX of 80386, to 64-bit