516 Computer Architecture and Organization
7. Assuming that only two memory addressing modes are felt to be essential for above problem
namely Immediate and Register Indirect, ensure that necessary data paths and control signals are
provided. Present the set of micro-operations necessary to implement these two addressing modes.
Assume that the processor would offer memory mapped I/O scheme.
8. Assume that for above problem, the maximum size of external memory was allowed to be 4 kB
offering 12 address lines adopting Princeton architecture. The 12-bit Program Counter is to be gen-
erally incremented by one. For branching operation, either it is to be reloaded with an immediate
value or to be calculated by adding the offset available within a register with the present value of
the Program Counter. Design the micro-architecture necessary to implement this Program Counter
operations including all control signals and related data paths.
9. Assume that the processor would have an internal stack area of eight levels capable of storing 12-bit
return address for the same problem. Design the system stack and stack pointer with necessary data
paths and related control signals.
10. A small 8-column numeric printer has eight disc type wheels with 10 decimal digits, the decimal
point and a dash sign (-) embossed around each wheel. After completion of one-line printing opera-
tion or after reset, the printer uses the following mechanism to print the next line.
For the left-most wheel, 12 clock pulses are generated by the printer and at each clock pulse the
wheel is rotated by 30° angle. The wheel’s rotation is stopped if the printer receives a blocking signal
(from the host). The rotation and signal transmission for the next wheel starts and angular position of all
eight wheels are nalized in this manner. Then the printer’s hammer has to be activated by an external
signal from the host for printing operation. After activation of the printing-hammer, the printer auto-
matically rolls the paper one line up.
(a) Draw the timing diagram indicating the printer operation.
(b) Prepare a schematic of the hardware that are necessary for the printer, assuming a small proces-
sor is available within it.
11. A pipelined processor with two stages for instruction fetch and instruction execution is to be designed.
Instruction fetch is expected to be a completely independent operation. The instruction execution would
be completed by fetching the operands from internal registers to ALU-registers. Instruction decoding
is to be performed just after instruction fetch by placing the fetched opcode within the decoding buffer.
Assuming all instructions are of 8-bit uniform length, design the data paths and control signals for
R The instruction fetch and instruction decoding unit
R Instruction execution unit with ALU, two ALU-operand registers, one ALU result register and
eight general purpose registers.
12. A super-scalar processor has a oating point unit for oating point ALU operations and an integer
unit with integer ALU for all other routine operations. The oating point unit and the integer unit
are self suf cient and independent of each other and can function simultaneously. The pre-decoding
unit selects two instructions from the instruction queue so that one is a oating point type and the
other is normal integer type and places in two separate queues for instruction decoding and execu-
tion, as shown in the schematic on following page.
Design necessary data paths and control signals for the pre-decoding unit including the instruction
queue and oating point and integer queues. Assume that if the most signi cant bit of the opcode byte
is 1, then it is a oating point instruction otherwise it is a normal integer instruction.
Z06_GHOS1557_01_SE_C21_App_F.indd 516Z06_GHOS1557_01_SE_C21_App_F.indd 516 4/29/11 5:47 PM4/29/11 5:47 PM