URL
https://opencores.org/ocsvn/integer_square_root/integer_square_root/trunk
Subversion Repositories integer_square_root
Compare Revisions
- This comparison shows the changes necessary to convert path
/integer_square_root/tags
- from Rev 3 to Rev 5
- ↔ Reverse comparison
Rev 3 → Rev 5
/v1.1/src/ISR.sv
0,0 → 1,87
`timescale 1ns / 1ps |
////////////////////////////////////////////////////////////////////////////////// |
// Company: |
// Engineer: Yihua Liu |
// |
// Create Date: 2022/06/08 16:50:36 |
// Design Name: |
// Module Name: ISR |
// Project Name: lab_3_b |
// Target Devices: xczu7eg-ffvf1517-2-i |
// Tool Versions: |
// Description: |
// |
// Dependencies: |
// |
// Revision: |
// Revision 0.01 - File Created |
// Additional Comments: |
// |
////////////////////////////////////////////////////////////////////////////////// |
|
|
module ISR( |
input reset, |
input [63:0] value, |
input clock, |
output logic [31:0] result, |
output logic done |
); |
logic [63:0] new_value, proposed_solution_square; |
logic [31:0] proposed_solution; |
logic [4:0] i; |
logic start, it_done, flush; |
logic reset_sync; |
|
mult Multiplier ( |
.clock(clock), |
.reset(reset), |
.mcand({32'h00000000, proposed_solution}), |
.mplier({32'h00000000, proposed_solution}), |
.start(start), |
.product(proposed_solution_square), |
.done(it_done) |
); |
|
always_comb begin |
// if (reset_async) begin |
// done = 0; |
// result = 0; |
// end |
// else begin |
// Reduction operator |
// see http://www.asic-world.com/verilog/operators2.html |
// done = ~|i & it_done & ~flush; |
// result[i] = (proposed_solution_square <= new_value) & it_done; |
done = ~|i & it_done & ~flush & ~reset_sync; |
result[i] = (proposed_solution_square <= new_value) & it_done & ~reset_sync; |
// end |
end |
|
always_ff @(posedge clock or posedge reset) begin |
if (reset) begin |
// done <= 0; |
// result <= 0; |
reset_sync <= 1; |
start <= 0; |
flush <= 0; |
i <= 5'b11111; |
proposed_solution <= 32'h80000000; |
new_value <= value; |
end |
else begin |
reset_sync <= 0; |
start <= !it_done || !flush; |
flush <= it_done; |
// if (!it_done && flush) begin |
// flush <= 0; |
// end |
if (i && it_done && !flush) begin |
// flush <= 1; |
i <= i - 1; |
proposed_solution[i-1] <= 1; |
proposed_solution[i] <= result[i]; |
end |
end |
end |
endmodule |
/v1.1/src/mult_stage.sv
0,0 → 1,68
`timescale 1ns / 100ps |
////////////////////////////////////////////////////////////////////////////////// |
// Company: |
// Engineer: Yihua Liu |
// |
// Create Date: 2022/06/06 12:58:15 |
// Design Name: |
// Module Name: mult_stage |
// Project Name: lab_3_a |
// Target Devices: xczu7eg-ffvf1517-2-i |
// Tool Versions: |
// Description: |
// |
// Dependencies: |
// |
// Revision: |
// Revision 0.01 - File Created |
// Additional Comments: |
// |
////////////////////////////////////////////////////////////////////////////////// |
|
|
// This is one stage of an 8 stage (9 depending on how you look at it) |
// pipelined multiplier that multiplies 2 64-bit integers and returns |
// the low 64 bits of the result. This is not an ideal multiplier but |
// is sufficient to allow a faster clock period than straight * |
module mult_stage( |
input clock, reset, start, |
input [63:0] product_in, mplier_in, mcand_in, |
|
output logic done, |
output logic [63:0] product_out, mplier_out, mcand_out |
); |
|
// This parameter is used to change the number of stages. |
// For example, if N_STAGE = 8, we are using an 8-stage pipelined multiplier. |
parameter N_STAGE = 8; |
|
logic [63:0] prod_in_reg, partial_prod_reg; |
logic [63:0] partial_product, next_mplier, next_mcand; |
|
assign product_out = prod_in_reg + partial_prod_reg; |
|
assign partial_product = mplier_in[64/N_STAGE-1:0] * mcand_in; |
|
// assign next_mplier = {{(64/N_STAGE){1'b0}},mplier_in[63:64/N_STAGE]}; |
// assign next_mcand = {mcand_in[64-64/N_STAGE-1:0],{(64/N_STAGE){1'b0}}}; |
assign next_mplier = mplier_in >> 64/N_STAGE; |
assign next_mcand = mcand_in << 64/N_STAGE; |
|
//synopsys sync_set_reset "reset" |
always_ff @(posedge clock) begin |
prod_in_reg <= #1 product_in; |
partial_prod_reg <= #1 partial_product; |
mplier_out <= #1 next_mplier; |
mcand_out <= #1 next_mcand; |
end |
|
// synopsys sync_set_reset "reset" |
always_ff @(posedge clock) begin |
if(reset) |
done <= #1 1'b0; |
else |
done <= #1 start; |
end |
|
endmodule |
|
/v1.1/src/pipe_mult.sv
0,0 → 1,57
`timescale 1ns / 100ps |
////////////////////////////////////////////////////////////////////////////////// |
// Company: |
// Engineer: Yihua Liu |
// |
// Create Date: 2022/06/06 13:54:21 |
// Design Name: |
// Module Name: mult |
// Project Name: lab_3_a |
// Target Devices: xczu7eg-ffvf1517-2-i |
// Tool Versions: |
// Description: |
// |
// Dependencies: |
// |
// Revision: |
// Revision 0.01 - File Created |
// Additional Comments: |
// |
////////////////////////////////////////////////////////////////////////////////// |
|
|
// This is an 8 stage (9 depending on how you look at it) pipelined |
// multiplier that multiplies 2 64-bit integers and returns the low 64 bits |
// of the result. This is not an ideal multiplier but is sufficient to |
// allow a faster clock period than straight * |
// This module instantiates 8 pipeline stages as an array of submodules. |
module mult( |
input clock, reset, |
input [63:0] mcand, mplier, |
input start, |
|
output [63:0] product, |
output done |
); |
|
// This parameter is used to change the number of stages. |
// For example, if N_STAGE = 8, we are using an 8-stage pipelined multiplier. |
parameter N_STAGE = 8; |
logic [63:0] mcand_out, mplier_out; |
logic [(N_STAGE-1)*64-1:0] internal_products, internal_mcands, internal_mpliers; |
logic [N_STAGE-2:0] internal_dones; |
|
mult_stage #(.N_STAGE(N_STAGE)) mstage [N_STAGE-1:0] ( |
.clock(clock), |
.reset(reset), |
.product_in({internal_products,64'h0}), |
.mplier_in({internal_mpliers,mplier}), |
.mcand_in({internal_mcands,mcand}), |
.start({internal_dones,start}), |
.product_out({product,internal_products}), |
.mplier_out({mplier_out,internal_mpliers}), |
.mcand_out({mcand_out,internal_mcands}), |
.done({done,internal_dones}) |
); |
|
endmodule |
/v1.1/src/test_ISR.sv
0,0 → 1,180
`define HALF_CYCLE 1000 |
`timescale 1ns / 1ps |
////////////////////////////////////////////////////////////////////////////////// |
// Company: |
// Engineer: Yihua Liu |
// |
// Create Date: 2022/06/08 16:51:12 |
// Design Name: |
// Module Name: testbench |
// Project Name: lab_3_b |
// Target Devices: xczu7eg-ffvf1517-2-i |
// Tool Versions: |
// Description: |
// |
// Dependencies: |
// |
// Revision: |
// Revision 0.01 - File Created |
// Additional Comments: |
// |
////////////////////////////////////////////////////////////////////////////////// |
|
|
module testbench(); |
|
logic [63:0] value; |
logic clock, reset; |
|
logic [31:0] result; |
logic done; |
|
logic [31:0] cres; |
|
wire correct = (cres===result)|~done|reset; |
|
integer value_test, j; |
time start_time = $time, finish_time; |
|
task ISR_test; |
input [63:0] value; |
output [31:0] cres; |
cres = 0; |
for (integer i = 31; i >= 0; i--) begin |
cres[i] = 1; |
if (cres * cres > value) |
cres[i] = 0; |
end |
endtask |
|
ISR isr( |
.reset(reset), |
.value(value), |
.clock(clock), |
.result(result), |
.done(done) |
); |
|
always @(posedge clock) |
#2 if(!correct) begin |
$display("Incorrect at time %4.0f",$time); |
$display("cres = %h result = %h",cres,result); |
$display("@@@Failed"); |
$finish; |
end |
|
always begin |
#`HALF_CYCLE; |
clock=~clock; |
end |
|
task wait_until_done; |
forever begin : wait_loop |
@(posedge done); |
@(negedge clock); |
if(done) begin |
finish_time = $time; |
$display("Done one calculation of ISR"); |
// Calculate the total number of cycles you need to do one calculation of ISR |
// Note that in this way the additional cycles added during simulation will be counted |
// So the displayed value may be 1 to 3 cycles more than 600 cycles |
$display("Number of cycles needed: %d", (finish_time - start_time) / (2 * `HALF_CYCLE)); |
start_time = finish_time; |
disable wait_until_done; |
end |
end |
endtask |
|
initial begin |
$monitor("Time:%4.0f done:%b cres:%h result:%h reset:%h",$time,done,cres,result,reset); |
// Test 1 |
value = 0; // square number |
reset = 1; |
clock = 0; |
ISR_test(value, cres); |
#2000; |
@(negedge clock); |
reset = 0; |
wait_until_done(); |
@(negedge clock); |
// Test 2 |
reset = 1; |
value = 4; // square number |
ISR_test(value, cres); |
@(negedge clock); |
reset = 0; |
wait_until_done(); |
@(negedge clock); |
// Test 3 |
reset = 1; |
value = 121; // square number |
ISR_test(value, cres); |
@(negedge clock); |
reset = 0; |
wait_until_done(); |
@(negedge clock); |
// Test 4 |
reset = 1; |
value = 1000000; // square number |
ISR_test(value, cres); |
@(negedge clock); |
reset = 0; |
wait_until_done(); |
// Test 5 |
reset = 1; |
value = 130; // non square number |
ISR_test(value, cres); |
@(negedge clock); |
reset = 0; |
wait_until_done(); |
// Test 6 |
@(negedge clock); |
reset = 1; |
value = 2; // non square number (value to be discarded) |
ISR_test(value, cres); |
@(negedge clock); |
reset = 0; |
@(negedge clock); |
@(negedge clock); |
@(negedge clock); |
reset = 1; // reset is asserted part way through a computation |
value = 3; // non square number (new value to be latched into) |
ISR_test(value, cres); |
@(negedge clock); |
reset = 0; |
wait_until_done(); |
@(negedge clock); |
@(negedge clock); |
// Test 7 |
reset = 1; |
value = 64'h7FFF_FFFF_FFFF_FFFF; |
ISR_test(value, cres); |
@(negedge clock); |
reset = 0; |
wait_until_done(); |
@(negedge clock); |
value_test = 64'hFFFF_FFFF_FFFF_FFFF; |
// Short loops |
for (j = 0; j < 100; j++) begin |
reset = 1; |
value = value_test; |
ISR_test(value, cres); |
@(negedge clock); |
reset = 0; |
wait_until_done(); |
value_test = value_test - 64'h0000_0000_7FFF_FFFF; |
end |
// Random testing |
for (j = 0; j < 1000; j++) begin |
reset = 1; |
value = {$random, $random}; |
ISR_test(value, cres); |
@(negedge clock); |
reset = 0; |
wait_until_done(); |
end |
$display("@@@Passed"); |
$finish; |
end |
|
endmodule |
/v1.1/README.md
0,0 → 1,48
# Integer Square Root |
|
## Algorithm |
|
``` |
procedure ISR(value) |
for i<-31 to 0 do |
proposed_solution[i]<-1 |
if proposed_solution^2 > value then |
proposed_solution[i]<-0 |
end if |
end for |
end procedure |
``` |
|
## Specification |
|
- If reset is asserted during a rising clock edge (synchronous reset), the value signal is to be stored. |
- If reset is asserted part way through a computation, the result of that computation is discarded and a new value is latched into the module. |
- When the module has finished computing the answer, the output is placed on the result line and done line is raised on the same cycle. |
- It must not take more than 600 clock cycles to compute a result (from the last |
clock that reset is asserted to the first clock that done is asserted.) |
|
## ISR State Machine |
|
Computing: $\sqrt{\mathtt{value}}$ |
|
- On a reset |
- guess initialized to `32'h8000_0000` |
- `value` is clocked into a register |
|
- guess gets the next bit set each time we cycle through the FSM again |
|
- Square `guess` (multiply it with itself) |
|
- Wait until the multiplier raises its done |
|
- if `guess` <= `value` |
|
- Keep the current bit |
|
- else |
|
- Clear the current bit |
|
- Move to the next bit |
|
- After the last bit, raise `done` |