Use larger block sizes for decode; optimize warp and block reduce fully

Use larger block sizes for decode; optimize warp and block reduce fully #37