diff --git a/flake.nix b/flake.nix
index f3e1324..dfa8251 100644
--- a/flake.nix
+++ b/flake.nix
@@ -11,38 +11,23 @@
 
           verilator = import ./verilator.nix pkgs ;
 
-          rot-sim = pkgs.runCommandCC "rot-sim" {} ''
-            cp "${./rot.v}" rot.v
-            ${verilator}/bin/verilator --cc -O3 --exe --build --trace ${./sim_main.cpp} rot.v
+          vflags = "-Wpedantic -Wwarn-lint -Wwarn-style";
+          verilate-src = cmd: ''
+            cp -r ${./verilog-src} ./verilog-src
+            cp -r ${./simulation} ./simulation
+            find ./verilog-src/ -name '*.v' -exec ${verilator}/bin/verilator ${vflags} ${cmd} {} +
+          '';
+
+          lint = pkgs.runCommand "lint" {} ''
+             ${verilate-src "--lint-only"}
+             echo "compiler didn't get angry :3"
+             : 3 > $out
+          '';
+
+          alu-sim = pkgs.runCommandCC "alu-sim" {} ''
+            ${verilate-src "--cc --build --exe ./simulation/test_alu.cpp"}
             mv obj_dir "$out"
-          '';
-
-          rot-trace = pkgs.runCommand "rot-trace" {} ''
-            mkdir "$out"
-            ${rot-sim}/Vrot $out/trace.vcd
-          '';
-
-          rot-trace-view = pkgs.writeScriptBin "rot-view" ''
-            ${pkgs.gtkwave}/bin/gtkwave ${rot-trace}/trace.vcd
-          '';
-
-          rot-synth = pkgs.runCommand "rot-synth" {} ''
-            mkdir "$out"
-            ${pkgs.yosys}/bin/yosys -p "read -vlog95 ${./rot.v} ; synth_ice40 -json $out/synth.json"
-          '';
-
-          rot-route = pkgs.runCommand "rot-route" {} ''
-            mkdir "$out"
-            ${pkgs.nextpnr}/bin/nextpnr-ice40 --pcf ${./rot.pcf} --json ${rot-synth}/synth.json --asc $out/routed.asc --up5k --top rot --freq 12
-          '';
-
-          rot-pack = pkgs.runCommand "rot-pack" {} ''
-            mkdir "$out"
-            ${pkgs.icestorm}/bin/icepack ${rot-route}/routed.asc $out/packed.bin
-          '';
-
-          rot-upload = pkgs.writeScriptBin "rot-upload" ''
-            ${pkgs.icestorm}/bin/iceprog ${rot-pack}/packed.bin
+            mkdir "$out/bin" && cp "$out/Valu" "$out/bin/alu-sim"
           '';
 
           deps = with pkgs; [
@@ -51,15 +36,9 @@
       in rec {
         packages.verilator = verilator;
 
-        packages.rot-sim = rot-sim;
-        packages.rot-trace = rot-trace;
-        packages.rot-trace-view = rot-trace-view;
-
-        packages.rot-synth = rot-synth;
-        packages.rot-route = rot-route;
-        packages.rot-pack = rot-pack;
-        packages.rot-upload = rot-upload;
+        packages.lint = lint;
 
+        packages.alu-sim = alu-sim;
         devShells.default = pkgs.mkShell {
           packages = deps;
         };
diff --git a/rot.pcf b/fpga-files/rot.pcf
similarity index 100%
rename from rot.pcf
rename to fpga-files/rot.pcf
diff --git a/rot.v b/rot.v
deleted file mode 100644
index efe71bc..0000000
--- a/rot.v
+++ /dev/null
@@ -1,30 +0,0 @@
-module rot (input clk, output D0, output D1, output D2, output D3);
-
-reg ready = 0;
-reg [23:0] divider;
-reg [3:0] rot;
-
-always @(posedge clk) begin
-    if (ready)
-        begin
-        if (divider == 12_000_000)
-            begin
-            divider <= 0;
-            rot <= {rot[2], rot[1], rot[0], rot[3]};
-            end
-        else
-            divider <= divider + 1;
-        end
-    else begin
-        ready <= 1;
-        rot <= 4'b1;
-        divider <= 0;
-    end
-end
-
-assign D0 = rot[0];
-assign D1 = rot[1];
-assign D2 = rot[2];
-assign D3 = rot[3];
-
-endmodule
diff --git a/sim_main.cpp b/simulation/sim_main.cpp
similarity index 100%
rename from sim_main.cpp
rename to simulation/sim_main.cpp
diff --git a/simulation/test_alu.cpp b/simulation/test_alu.cpp
new file mode 100644
index 0000000..ed77946
--- /dev/null
+++ b/simulation/test_alu.cpp
@@ -0,0 +1,18 @@
+#include "Valu.h"
+#include "verilated.h"
+#include <iostream>
+
+int main(int argc, char **argv) {
+    VerilatedContext *vctx = new VerilatedContext;
+
+    Valu *valu = new Valu(vctx);
+    valu->A = 0x2137;
+    valu->B = 0x1234;
+
+    valu->op = 0b000;
+    valu->eval();
+    printf("A = %04x + B = %04x = O = %04x\n", valu->A, valu->B, valu->O);
+    valu->op = 0b001;
+    valu->eval();
+    printf("A = %04x - B = %04x = O = %04x\n", valu->A, valu->B, valu->O);
+}
diff --git a/verilog-src/alu.v b/verilog-src/alu.v
new file mode 100644
index 0000000..3fde343
--- /dev/null
+++ b/verilog-src/alu.v
@@ -0,0 +1,28 @@
+/*
+  A and B are inputs, O is out
+  OP is the operation to perform:
+    000 = add
+    001 = sub
+    010 = mul
+    (011)
+    100 = and
+    101 = or
+    110 = xor
+    111 = not A
+*/
+module alu(
+  input [31:0] A,
+  input [31:0] B,
+  input [2:0] op,
+  output [31:0] O,
+  output Cout,
+  output Fzero
+);
+
+wire subtraction = op == 3'b001;
+wire [31:0] adder_B = subtraction ? ~B : B;
+
+carry_select_adder a(A, adder_B, subtraction, O, Cout);
+assign Fzero = & O;
+
+endmodule
diff --git a/verilog-src/alu/carry_select_adder.v b/verilog-src/alu/carry_select_adder.v
new file mode 100644
index 0000000..4215209
--- /dev/null
+++ b/verilog-src/alu/carry_select_adder.v
@@ -0,0 +1,29 @@
+// 5 blocks of size 6, final block of size 2
+module carry_select_adder(
+  input [31:0] A,
+  input [31:0] B,
+  input Cin,
+  output [31:0] O,
+  output Cout
+);
+
+// TODO: First block should not be carry_select, a normal RCA is better
+// i think?
+wire Cout_5;
+carry_select_block#(6) B0 (A[5:0], B[5:0], Cin, O[5:0], Cout_5);
+
+wire Cout_11;
+carry_select_block#(6) B1 (A[11:6], B[11:6], Cout_5, O[11:6], Cout_11);
+
+wire Cout_17;
+carry_select_block#(6) B2 (A[17:12], B[17:12], Cout_11, O[17:12], Cout_17);
+
+wire Cout_23;
+carry_select_block#(6) B3 (A[23:18], B[23:18], Cout_17, O[23:18], Cout_23);
+
+wire Cout_29;
+carry_select_block#(6) B4 (A[29:24], B[29:24], Cout_23, O[29:24], Cout_29);
+
+carry_select_block#(2) B6 (A[31:30], B[31:30], Cout_29, O[31:30], Cout);
+
+endmodule
diff --git a/verilog-src/alu/carry_select_block.v b/verilog-src/alu/carry_select_block.v
new file mode 100644
index 0000000..ffd7d8b
--- /dev/null
+++ b/verilog-src/alu/carry_select_block.v
@@ -0,0 +1,39 @@
+module carry_select_block#(
+  parameter N = 4
+) (
+  input [N-1:0] A,
+  input [N-1:0] B,
+  input Cin,
+  output [N-1:0] O,
+  output Cout
+);
+
+// Case for Cin = 0
+wire [N-1:0] O_C0;
+wire Cout_C0;
+
+begin
+  wire [N-1:1] carry;
+  fa fa0(A[0], B[0], 0, O_C0[0], carry[1]);
+  genvar i; for (i = 1; i <= N - 1; i = i + 1) begin
+    fa fai(A[i], B[i], carry[i], O_C0[i], carry[i+1]);
+  end
+  assign Cout_C0 = carry[N-1];
+end
+
+// Case for Cin = 1
+wire [N-1:0] O_C1;
+wire Cout_C1;
+begin
+  wire [N-1:1] carry;
+  fa fa0(A[0], B[0], 1, O_C1[0], carry[1]);
+  genvar i; for (i = 1; i <= N - 1; i = i + 1) begin
+    fa fai(A[i], B[i], carry[i], O_C1[i], carry[i+1]);
+  end
+  assign Cout_C1 = carry[N-1];
+end
+
+assign O = Cin == 0 ? O_C0 : O_C1;
+assign Cout = Cin == 0 ? Cout_C0 : Cout_C1;
+
+endmodule
diff --git a/verilog-src/alu/fa.v b/verilog-src/alu/fa.v
new file mode 100644
index 0000000..8581d5b
--- /dev/null
+++ b/verilog-src/alu/fa.v
@@ -0,0 +1,7 @@
+// One bit full-adder
+module fa(input A, input B, input Cin, output O, output Cout);
+
+assign O = A ^ B ^ Cin;
+assign Cout = (A & B) | (B & Cin) | (Cin & A);
+
+endmodule