An attack on the last two rounds of MD4

In [Rive90] the MD4 message digest algorithm was introduced taking an input message of arbitrary length and producing an output 128-bit message digest. It is conjectured that it is computationally infeasible to produce two messages having the same message digest, or to produce any message having a given prespecified target message. In this paper it is shown that if the three round MD4 algorithm is stripped of its first round, it is possible to find for a given (initial) input value two different messages hashing to the same output. A computer program implementing this attack takes about 1 millisecond on a 16 Mhs IBM PS/2 to find such a collision.status: publishe