-----
Comment #6 on issue 422 by [email protected]: suggested speedup in com.google.zxing.common.reedsolomon http://code.google.com/p/zxing/issues/detail?id=422
I'll commit the first changes. It can actually get faster by removing the check for one operand being 1. Saving two branches helps the pipeline more than it hurts by having to carry out the rest of the operation in those cases.
Sure post a patch with the rest of your findings. I am sure it can be sped up.
Credit goes partly to Google's code base, we got permission to pinch and adapt some of the routines used internally.