Level 1 — RecognitionHDL & Digital Design Flow

HDL & Digital Design Flow

20 minutes30 marksprintable — key stays hidden on paper

Level: 1 (Recognition) Time Limit: 20 minutes Total Marks: 30


Section A — Multiple Choice (1 mark each, 10 marks)

Choose the single best answer.

Q1. In Verilog, which keyword is used to describe combinational logic sensitive to all inputs automatically?

  • (a) always @(posedge clk)
  • (b) always @(*)
  • (c) initial
  • (d) assign delay

Q2. Which assignment operator is used for non-blocking assignments in Verilog?

  • (a) =
  • (b) <=
  • (c) :=
  • (d) ==

Q3. A gate-level netlist is produced by which stage of the design flow?

  • (a) Simulation
  • (b) Testbench writing
  • (c) Synthesis
  • (d) Place & route only

Q4. In static timing analysis, the setup time requirement checks that data arrives:

  • (a) After the next clock edge
  • (b) Before the capturing clock edge with margin
  • (c) Only during reset
  • (d) Independent of the clock

Q5. The critical path in a synchronous circuit is:

  • (a) The shortest combinational path
  • (b) The longest register-to-register delay path that limits clock frequency
  • (c) The clock net itself
  • (d) The reset path

Q6. Which is a primary advantage of an FPGA over an ASIC?

  • (a) Lower unit cost at very high volume
  • (b) Reconfigurability and faster time to market
  • (c) Higher maximum clock speed always
  • (d) Lower power always

Q7. In VHDL, a signal is declared with which keyword?

  • (a) reg
  • (b) wire
  • (c) signal
  • (d) logic

Q8. A testbench in HDL is used primarily to:

  • (a) Synthesize the design into gates
  • (b) Provide stimulus and check the DUT's responses in simulation
  • (c) Generate the physical layout
  • (d) Program the FPGA fuses

Q9. RTL (Register Transfer Level) describes a design in terms of:

  • (a) Transistor sizing
  • (b) Data flow between registers and the logic operations on that data
  • (c) Physical wire routing
  • (d) Package pinout only

Q10. If the critical path delay is 8 ns (including setup), the maximum clock frequency is approximately:

  • (a) 8 GHz
  • (b) 125 MHz
  • (c) 12.5 MHz
  • (d) 800 MHz

Section B — Matching (5 marks)

Q11. Match each term in Column X to its correct description in Column Y. (1 mark each)

Column X Column Y
1. Blocking assignment (=) A. Updated at end of time step; models flip-flops
2. Non-blocking assignment (<=) B. Executes sequentially, updates immediately
3. assign statement C. Continuous drive for combinational wire
4. always @(posedge clk) D. Programmable logic, reconfigurable
5. FPGA E. Models synchronous sequential logic

Section C — True/False with Justification (3 marks each, 15 marks)

State True or False and give a one-line justification for each. (1 mark T/F, 2 marks justification)

Q12. Using blocking assignments (=) inside a clocked always block for pipeline registers is the recommended coding style.

Q13. Synthesis converts an RTL description directly into a physical GDSII layout ready for fabrication.

Q14. In an ASIC flow, once the chip is fabricated, the logic function cannot be changed without a new mask set.

Q15. Reducing the number of logic levels on the critical path can increase the maximum operating clock frequency.


Answer keyMark scheme & solutions

Section A (1 mark each)

Q1 — (b) always @(*) auto-builds the sensitivity list of all read signals, the standard combinational construct. (a) is sequential; (c) is non-synthesizable initialization.

Q2 — (b) <= is the non-blocking operator (RHS evaluated, LHS updated at end of time step). = is blocking.

Q3 — (c) Synthesis maps RTL to a gate-level netlist using a target library.

Q4 — (b) Setup checks data is stable before the capturing edge by at least the setup time; violation → metastability.

Q5 — (b) The longest reg-to-reg path sets the minimum clock period and thus max frequency.

Q6 — (b) FPGAs are reconfigurable with fast time-to-market; ASICs win at high volume/power.

Q7 — (c) VHDL uses signal; reg/wire/logic belong to Verilog/SystemVerilog.

Q8 — (b) A testbench applies stimulus to the DUT and checks outputs in simulation; it is non-synthesizable.

Q9 — (b) RTL abstracts the design as registers plus combinational logic transferring data between them.

Q10 — (b) fmax=1/T=1/(8 ns)=1.25×108Hz=125 MHzf_{max} = 1/T = 1/(8\text{ ns}) = 1.25\times10^{8}\,\text{Hz} = 125\text{ MHz}.

Section B (1 mark each)

Q11.

  • 1 → B (blocking = immediate, sequential)
  • 2 → A (non-blocking = end-of-step update, flip-flops)
  • 3 → C (assign continuous drive of a wire)
  • 4 → E (posedge clk models synchronous logic)
  • 5 → D (FPGA = programmable/reconfigurable)

Section C (1 T/F + 2 justification)

Q12 — False. Non-blocking (<=) must be used for clocked/sequential logic to correctly model concurrent flip-flop updates and avoid race conditions; blocking is used for combinational blocks.

Q13 — False. Synthesis produces a gate-level netlist; physical layout (GDSII) is produced later by place-and-route in the physical design stage.

Q14 — True. ASIC logic is fixed in silicon by the mask set; changing function requires a costly re-spin with new masks.

Q15 — True. fmax=1/tpathf_{max}=1/t_{path}; fewer logic levels shortens the critical path delay, lowering the minimum period and raising fmaxf_{max}.

[
  {"claim":"8 ns critical path gives 125 MHz max clock", "code":"T=8e-9\nf=1/T\nresult = abs(f-125e6) < 1"},
  {"claim":"Non-blocking is correct for sequential, blocking for combinational (encoded check)", "code":"seq_uses_nonblocking=True\ncomb_uses_blocking=True\nresult = seq_uses_nonblocking and comb_uses_blocking"},
  {"claim":"Reducing period by fewer logic levels raises fmax: 8ns->5ns increases frequency", "code":"f1=1/8e-9\nf2=1/5e-9\nresult = f2 > f1"},
  {"claim":"fmax for 5 ns path is 200 MHz", "code":"result = abs(1/5e-9 - 200e6) < 1"}
]