Shift a signed 12-bit sample right while keeping its sign, and divide it by a power of two the way C rounds.
Hardware context: A cipher round that rotates a word by a key-dependent amount, and a sensor filter that scales signed samples. Shift a signed sample right and divide it rounding toward zero.
A sensor sends 12-bit two's-complement samples on a plain logic port. A filter needs two things for each sample. One is a fast scale-down by 2^n that keeps the sign. The other is an exact quotient that must match the C reference model.
This is an Implement problem. Write sra12. y is x shifted right by n places, filling with copies of the sign bit. From n = 12 up every bit of y is the sign bit. q is x divided by 2^n and rounded toward zero, as integer division in C does.
The check applies three (x, n) pairs, one per nanosecond, and records y and q after each.
Signals in the Timeline
| Signal | Width | Role |
|---|---|---|
script | 48 | input of each check |
Full specification
module sra12 (
input logic [11:0] x, // a two's-complement sample on a plain (unsigned) port
input logic [3:0] n, // 0 to 15 places
output logic [11:0] y, // x shifted right by n, keeping the sign
output logic [11:0] q // x divided by 2^n, rounded toward zero
);
// your design
endmodule
| x, n | y (shift) | q (division) |
|---|---|---|
| 100, 2 | 25 | 25 |
| −7, 1 | −4 | −3 |
| −1, 13 | −1 | 0 |
The result fields are y_stepK and q_stepK for the steps 0 to 2, shown as signed decimal numbers.
Examples
Example 1
Positive Values- Input
x=100 n=2; x=2047 n=4; x=7 n=1
- Output
y_step0 = 25, q_step0 = 25, y_step1 = 127, q_step1 = 127, y_step2 = 3, q_step2 = 3
For positive samples the shift and the division give the same answers, with zeros shifted in.
Example 2
Negative Values That Divide Exactly- Input
x=-100 n=2; x=-2048 n=4; x=-8 n=3
- Output
y_step0 = -25, q_step0 = -25, y_step1 = -128, q_step1 = -128, y_step2 = -1, q_step2 = -1
With no remainder the two outputs agree again. The sign bit is copied into the vacated places.
Example 3
Negative Values with a Remainder- Input
x=-7 n=1; x=-1 n=1; x=-101 n=2
- Output
y_step0 = -4, q_step0 = -3, y_step1 = -1, q_step1 = 0, y_step2 = -26, q_step2 = -25
Now they differ. The shift rounds down: −7 shifted by 1 is −4, and −1 stays −1. The division rounds toward zero: −3 and 0.
Constraints and Contract
- Inputs are known 0/1 values. Each pair is held for one nanosecond and the outputs are sampled at its end.
- Keep the module name and the ports, including the unsigned declaration of x. The / and % operators are refused.
Hints
Hint 1
On an unsigned value, >>> fills with zeros. Read x as signed first, with $signed or a signed copy.
Hint 2
For a positive x the shift and the division agree. For a negative x the shift rounds down, toward minus infinity.
Hint 3
Adding 2^n − 1 to a negative x before the shift makes it round toward zero. That sum needs more than 12 bits when n is large.
What the grader checks: Eight directed and ten seeded cases of three steps each cover every shift amount, both signs and exact and inexact divisions. y and q are compared with a model written in JavaScript after every step.
18 checks on Icarus · browser samples / server checks. Run the examples first, then check the full regression. Full checks use the server judge. Downloads contain public samples only. A passing result establishes the tested cases, not a proof for all possible inputs.
See all 3 steps and the final capstone