Astropulse is a volunteer computing project that is run by the same people as the SETI@Home project. Its purpose is to search for short duration (as brief as 400 nanoseconds) broadband radio pulses. It is not currently known what might produce such phenomena because there has been relatively little research into this subject. Two interesting theories are E.T. and evaporating black holes.
Radio SETI traditionally looks for narrow-band signals as an indication of intelligent life. One theory is that an alien civilisation might not choose to concentrate power in a continuous narrow-band signal but instead concentrate power in a broad-band but narrow time frame pulse. This method of communication has a couple advantages, especially if the extra-terrestrial civilisation were trying to broadcast to other potential civilisations (like our METI experiments). A broad-band signal has a greater chance of being detected than a narrow-band one.
There is also the phenomenon of radio dispersion. Higher frequencies travel slightly faster through interstellar matter than lower radio frequencies, so a radio pulse transmitted from another star on a broad range of frequencies simultaneously would arrive at earth dispersed with the higher frequencies coming first. This dispersion would make it easier to distinguish between signals coming from other stars or from our own planet. Or if an alien civilisation were really trying to broadcast their existence they could transmit a pulse with a great negative dispersion (instead of broadcasting on a broad band of frequencies at the same instant they could start on the lower frequencies and move up at a constant rate). A signal of this sort would almost certainly have to be created by an intelligent being.
A problem about looking for radio pulses is that if you don't know how far away they are coming from then you do not know how much they will be dispersed. So in order to do a thorough search, one has to analyse the radio telescope data for many dispersions lengths. This gets to be very computationally expensive. That is why the Astropulse project is so well suited the the volunteer computing methodology that the SETI@Home project uses. The data from the Arecibo radio telescope is broken down and analysed by volunteers' computers. The Astropulse application checks for multiple dispersion rates including negative dispersion.