Chapter 4
ARM Architecture
4.1) Introduction
The most important part in our project is the microprocessor design. The
design passes through many steps until it reaches the final form. The
project is an ARM processor that is constructed from the following
components:
Arithmetic Logic Unit
Booth multiplier
Barrel shifter
Control unit
Register file
These components will be covered later on this chapter. The processor also
has others components:
Program status register: It contains the processor flags (Z, S, V and C).
The modes bits also exist in the program status register in addition to the
interrupt and fast interrupt disable bits
Some special registers: Some registers are used like the instruction
register, memory data read and write register and memory address register
Priority encoder: The encoder is used in the multiple load and store
instruction to indicate which register in the register file to be loaded or
stored
Multiplexers: Many multiplexers are used to control the operation of the
processor buses
Because of the limited project time, we implement these components in a
behavioral model. Each component is represented with an entity. Each entity
has its own architectures that can be optimized for certain requirements
depending on its application. This made the design easier to construct and
maintain. Figure (4-1) shows the processor block diagram.
Figure (4-1): ARM Block Diagram
4.2) Arithmetic Logic Unit (ALU)
The ALU has two 32-bits inputs. The first comes from the register
file while the other comes from the shifter. ALU outputs modify the status
register flags. The V-bit output goes to the V flag as well as the Cout
goes to the C flag. While the most significant bit actually represents the
S flag, the ALU output is NORed to get the Z flag. The entity
representation of the ALU is shown in figure (4-2).
[pic]
Figure (4-2): ALU entity
The ALU has a 4-bit function bus that allows up to 16 opcode to be
implemented. The different opcodes and their corresponding arithmetic and
logic operations are shown in table (2-1). The ALU test bench simulation
results are shown in figure (4-3).
Figure (4-3): ALU simulation results
4.3) Booth multiplier
The multiplier has three 32-bit inputs. All the inputs come from the
register file. The multiplier output is only the 32 least significant bits
of the product. The entity representation of the multiplier is shown in
figure (4-4). The multiplication starts whenever the start input goes
active. The output fin goes high when finishing.
Figure (4-4): Multiplier entity
4.3.1) Booth algorithm
The multiplier is implemented using the modified booth algorithm. The
multiplier is scanned sequentially from right to left. In this case,
however, three adjacent bits are examined during each step of the
procedure. According to the value of the three bits, the multiplicand is
added to or subtracted form the accumulated partial product and the later
is then shifted. Figure (4-5) shows a version of this algorithm to multiply
two 32-bits fractions using the logic circuit of Figure (4-6).
Figure (4-5): The Booth's algorithm
Intermediate results are stored in the accumulator-multiplier register pair
A, Q. An extra cell Q(32) is appended to the right most end of Q to
facilitate the triple-bit inspection process.
Figure (4-6): Multiplier logic circuit
Booth algorithm is an interesting multiplication algorithm for twos-
complement numbers. It treats positive and negative numbers uniformly.
Furthermore, runs of 0s or 1s in the multiplier are skipper over without
any addition or subtraction being performed, thereby making possible faster
multiplication. Figure (4-7) shows the simulation results for the
multiplier test bench. It is clear that the multiplication finishes only in
16 clock cycle.
Figure (4-7): Multiplier simulation results
4.4) Barrel shifter
The barrel shifter has a 32-bit input to be shifted. This input is
coming from the register file or it could be immediate data. The shifter
has other control inputs coming from instruction register. Shift field in
the instruction controls the operation of the barrel shifter. This field
indicates the type of shift to be performed (logical left or right,
arithmetic right or rotate right). The amount by which the register should
be shifted is contained in an immediate field in the instruction or it
could be the lower 6 bits of a register in the register file. Figure (4-8)
shows the entity representation of the barrel shifter.
Figure (4-8): Shifter entity
The shift_val input bus is 6-bits allowing up to 32 bit shifts. The
shift_typ indicates the required shift type for 00, 01, 10, 11 are
corresponding to logical shift left, logical shift right, arithmetic shift
right and rotate right respectively.
The barrel shifter is mainly constructed with multiplexers. Figure (4-
10) shows the logical circuit for the barrel shifter. The test bench
simulation of the shifter is also shown in Figure (4-9).
Figure (4-9): Shifter simulation results
Figure (4-10): Shifter logic circuit
4.5) Control unit
For any microprocessor, control unit is the heart of the system. It
is responsible for the system operation and so the control unit design is
the most important part in the hole design. Control unit is usually a pure
combinational circuit. In our design, the control unit is implemented by
simple state machine. The processor timing is also included in the control
unit. Figure (4-11) shows the state machine that implements the control
unit.
Signals from the control unit are connected to every component in the
processor to supervise its operation. The value of those signals for each
instruction at every state is explained in Appendix A.
Figure (4-11): Control unit state machine
4.6) ARM7 Functional Diagram
The last thing that needs to be explained is how the ARM can be used
and how will the chip appear. Figure (4-12) shows the functional block
diagram of our ARM processor. The different signals that interface with the
processor are input, output or supervisory signals that can be used to
control the ARM operation.
Figure (4-12): ARM Functional Diagram