Same change as Revision r1395 for C++ port: Small speedup, per issue 422

[zxing.git] / cpp / core / src / zxing / common / reedsolomon / GF256.cpp

diff --git a/cpp/core/src/zxing/common/reedsolomon/GF256.cpp b/cpp/core/src/zxing/common/reedsolomon/GF256.cpp
index 35fc692..51b621e 100644 (file)
--- a/cpp/core/src/zxing/common/reedsolomon/GF256.cpp
+++ b/cpp/core/src/zxing/common/reedsolomon/GF256.cpp
@@ -109,13 +109,10 @@ int GF256::multiply(int a, int b) {
   if (a == 0 || b == 0) {
     return 0;
   }
-  if (a == 1) {
-    return b;
-  }
-  if (b == 1) {
-    return a;
-  }
-  return exp_[(log_[a] + log_[b]) % 255];
+  int logSum = log_[a] + log_[b];
+  // index is a sped-up alternative to logSum % 255 since sum
+  // is in [0,510]. Thanks to jmsachs for the idea
+  return exp_[(logSum & 0xFF) + (logSum >> 8)];
 }
 
 GF256 GF256::QR_CODE_FIELD(0x011D); // x^8 + x^4 + x^3 + x^2 + 1