Antimatter is often depicted as the ultimate rocket fuel, while it promises great performance it has many limitations, this thread focuses mainly on the pion antimatter rocket and some variations of itself and how it works as a fuel for (fast) relativistic interstellar rockets. 
Pion rockets work by annihilating hydrogen and antihydrogen, quarks and antiquarks annihilate and create pions, which are either neutral or charged. The charged pions can be redirected out of the magnetic nozzle for thrust, energy is lost due to neutral pions and inefficiencies.
The charged pions travel multiple meters before decaying muons and muon neutrinos, the neutral pions decay almost instantly into gamma-ray pairs. The number of pions varies per annihilation and must, therefore, be averaged.
Robert Frisbee's antimatter ships is a proposal for a 700km long interstellar starship using antimatter, it's delta-V would be limited to 0.25c due to the isp only being 0.33c, this is caused by the inefficient magnetic nozzle and the losses to neutral pions.
Back in 2012, the pion rocket was redesigned, with increased nozzle efficiency, doubling the exhaust velocity, it was also shown that the multiplicity for neutral and charged pions was on average 6.5 and 3 respectively, limiting the isp to only 0.265c and with lots of gamma-rays.
With this data other methods to improve fuel efficiency must be considered, using a dragless scoop we can extract hydrogen from the interstellar medium to annihilate with the onboard antihydrogen, this doubles the effective isp to impressive 0.53c
Producing antimatter is extremely inefficient, Robert l. Forward estimated that a dedicated antimatter factory could produce antimatter at 0.01% efficiency. This means creating one-kilogram antiprotons will require 9*10^20 joules, or 2.5*10^14 kWh!
Even if we covered the entire moon, or mercury, or some equivalent solar satellite, with solar panels it would still take centuries to get enough antimatter for a single manned relativistic rocket, which requires hundreds to thousands of tons of antimatter.
While there are variations on the antimatter rocket, some of which even achieving near-perfect efficiency (Pinch discharge antimatter laser concept by friedwardt winterberg) it is still extremely uneconomical, impractical and dangerous to use.
But antimatter can still make a good fuel for other purposes, using positrons (which can be much more efficiently generated) in small amounts powering thermal engines, using antiprotons can also catalyze both fission and fusion reactions rather efficiently, especially the former.
When it comes to (fast) relativistic interstellar travel external propulsion seems to be the right answer, lasers sails, and especially sail beam, are well equipped when it comes to relativistic travel, as efficiency rises high velocity, and long distance acceleration is possible
