reorganization

src/bigmath/chain/segment.cpp

@@ -1,12 +1,11 @@
 #include "segment.hpp"
 
+#include <bigmath/chain/segment_ops.hpp>
+
 #include <bigmath/util/describe.hpp>
 
 namespace bigmath {
 
-    u8 add(u8* a, const u8* b, u8* c, u8 carry, u8 length);
-    u8 sub(u8* a, const u8* b, u8* c, u8 borrow, u8 length);
-
     const u8 segment::digit_count = 128;
     const u8 segment::byte_count = segment::digit_count / 2;
 
@@ -85,64 +84,7 @@ namespace bigmath {
         segment low;
         segment high;
 
-        const u32 blen = segment::byte_count;
+        bigmath::mul_long(digits, s.digits, high.digits, low.digits, byte_count);
-        const u32 blen2 = blen * 2;
-        const u32 blen4 = blen * 4;
-
-        // Step 1: Unpack raw BCD bytes into raw digit buffers (0-9)
-        u8 digits_a[blen2];
-        u8 digits_b[blen2];
-        const u8* bytes_a = bytes();
-        const u8* bytes_b = s.bytes();
-
-        for (u32 i = 0; i < blen; ++i) {
-            u8 byte_a = bytes_a[i];
-            u8 byte_b = bytes_b[i];
-
-            digits_a[i * 2] = byte_a & 0x0F;
-            digits_a[i * 2 + 1] = byte_a >> 4;
-
-            digits_b[i * 2] = byte_b & 0x0F;
-            digits_b[i * 2 + 1] = byte_b >> 4;
-        }
-
-        // Step 2: Accumulate digit products into temporary array of size blen4
-        u16 accum[blen4] = {};
-
-        for (u32 i = 0; i < blen2; ++i) {
-            u8 d1 = digits_a[i];
-            if (d1 == 0) continue; // Skip zeros
-
-            for (u32 j = 0; j < blen2; ++j) {
-                u8 d2 = digits_b[j];
-                accum[i + j] += static_cast<u16>(static_cast<u16>(d1) * static_cast<u16>(d2));
-            }
-        }
-
-        // Step 3: Base-10 Carry Propagation and Direct Packed BCD Storage
-        u32 carry = 0;
-        u8 out_digits[blen4];
-
-        for (u32 i = 0; i < blen4; ++i) {
-            u32 sum = accum[i] + carry;
-            out_digits[i] = static_cast<u8>(sum % 10);
-            carry = sum / 10;
-        }
-
-        // Step 4: Pack result digits directly into low and high segment bytes
-        for (u32 i = 0; i < blen; ++i) {
-            u8 low_digit0 = out_digits[i * 2];
-            u8 low_digit1 = out_digits[i * 2 + 1];
-
-            u8 high_digit0 = out_digits[i * 2 + blen2];
-            u8 high_digit1 = out_digits[i * 2 + blen2 + 1];
-
-            low.set(static_cast<u8>(i * 2), low_digit0);
-            low.set(static_cast<u8>(i * 2 + 1), low_digit1);
-
-            high.set(static_cast<u8>(i * 2), high_digit0);
-            high.set(static_cast<u8>(i * 2 + 1), high_digit1);
-        }
         
         return { low, high };
     }

src/bigmath/chain/segment.hpp

@@ -2,6 +2,8 @@
 
 #include <bigmath/BigMath.hpp>
 
+#include <bigmath/util/util.hpp>
+
 namespace bigmath {
 
     /**

src/bigmath/chain/segment_ops.cpp

@@ -69,4 +69,69 @@ namespace bigmath {
         return borrow;
     }
 
+    void mul_long(const u8* a, const u8* b, u8* c_high, u8* c_low, u8 length) {
+        const u32 blen = length;
+        const u32 blen2 = blen * 2;
+        const u32 blen4 = blen * 4;
+
+        // Step 1: Unpack raw BCD bytes into raw digit buffers (0-9)
+        u8 digits_a[blen2];
+        u8 digits_b[blen2];
+
+        for (u32 i = 0; i < blen; ++i) {
+            u8 byte_a = a[i];
+            u8 byte_b = b[i];
+
+            digits_a[i * 2] = byte_a & 0x0F;
+            digits_a[i * 2 + 1] = byte_a >> 4;
+
+            digits_b[i * 2] = byte_b & 0x0F;
+            digits_b[i * 2 + 1] = byte_b >> 4;
+        }
+
+        // Step 2: Accumulate digit products into temporary array of size blen4
+        u16 accum[blen4] = {};
+
+        for (u32 i = 0; i < blen2; ++i) {
+            u8 d1 = digits_a[i];
+            if (d1 == 0) continue; // Skip zeros
+
+            for (u32 j = 0; j < blen2; ++j) {
+                u8 d2 = digits_b[j];
+                accum[i + j] += static_cast<u16>(static_cast<u16>(d1) * static_cast<u16>(d2));
+            }
+        }
+
+        // Step 3: Base-10 Carry Propagation and Direct Packed BCD Storage
+        u32 carry = 0;
+        u8 out_digits[blen4];
+
+        for (u32 i = 0; i < blen4; ++i) {
+            u32 sum = accum[i] + carry;
+            out_digits[i] = static_cast<u8>(sum % 10);
+            carry = sum / 10;
+        }
+
+        // Step 4: Pack result digits directly into low and high segment bytes
+        for (u32 i = 0; i < blen; ++i) {
+            u8 low_digit0 = out_digits[i * 2];
+            u8 low_digit1 = out_digits[i * 2 + 1];
+
+            u8 high_digit0 = out_digits[i * 2 + blen2];
+            u8 high_digit1 = out_digits[i * 2 + blen2 + 1];
+
+            // No longer using the object version,
+            // so we set the bytes ourselves!
+            c_low[i] = low_digit0 | low_digit1 << 4;
+            //low.set(static_cast<u8>(i * 2), low_digit0);
+            //low.set(static_cast<u8>(i * 2 + 1), low_digit1);
+
+            c_high[i] = high_digit0 | high_digit1 << 4;
+            //high.set(static_cast<u8>(i * 2), high_digit0);
+            //high.set(static_cast<u8>(i * 2 + 1), high_digit1);
+        }
+
+        // end!
+    }
+
 }

src/bigmath/chain/segment_ops.hpp

@@ -0,0 +1,19 @@
+#pragma once
+
+#include <bigmath/BigMath.hpp>
+
+namespace bigmath {
+
+    u8 add(u8* a, const u8* b, u8* c, u8 carry, u8 length);
+
+    u8 sub(u8* a, const u8* b, u8* c, u8 borrow, u8 length);
+
+    void mul_long(const u8* a, const u8* b, u8* c_high, u8* c_low, u8 length);
+
+    void mul_karatsuba(const u8* a, const u8* b, u8* c_high, u8* c_low, u8 length);
+
+    void div_long(const u8* a, const u8* b, u8* quot, u8* rem, u8 length);
+
+    void div_newtonraphson(const u8* a, const u8* b, u8* quot, u8 length);
+
+}