Unconfined compression and tension tests have been performed on cylindrical samples prepared from a newly developed concrete. A 3-inch Hopkinson bar setup has been employed to experimentally extract the stress-strain relation of the concrete at different strain rates. A novel procedure is introduced to conduct tension tests. Initiation and propagation of cracks in concrete samples are captured by high-speed photography. The experimental data will be used to improve the existing concrete material models. Extracting stress-strain relations from the Hopkinson-bar experimentally obtained data requires assumptions about the sample size and interface friction between the sample and the bars. The effect of violating these assumptions on the validity of experimentally acquired stress-strain results is explored by performing a full-scale finite element simulation of the entire process. A new method of accounting for the dispersion in the Hokinson cylindrical bars using the finite-element results is introduced and verified by comparing these results with the generally used analytical method. Numerical results reveal that the strain and stress history are not uniform within the sample. The usual method of analyzing the Hopkinson bar experimental results that generally overlook this fact can be corrected by minor calibration on the data.