gh-157742: Speed up int multiplication by not zeroing the whole result in x_mul()

gh-157742: Speed up int multiplication by not zeroing the whole result in x_mul() #140851