Beam driven plasma wakes show great promise for meter scale accelerators with high gradients. Plasma wakefield theory indicates that the achievable gradient is proportional to N/{sigma}{sub z}{sup 2}, and the bunches as short as 12 {micro}m {approx} 40 fsec in RMS length which are now possible at the Stanford Linear Accelerator Center (SLAC) are predicted to allow gradients in the tens to hundreds of GeV/m. We discuss the three stages of compression needed to achieve such short bunches. No technique currently available can measure these longitudinal profiles directly shot by shot, requiring an indirect method. We added a magnetic chicane near the end of SLAC's 3 km main accelerator to measure the energy spread of each bunch in a nondestructive manner. Additionally, we performed a series of detailed simulations of the main accelerator in LiTrack, a code developed at SLAC. By comparing each measured spectrum against the library of spectra from simulations, we can find the best match to determine the input conditions to the accelerator and the total longitudinal phase space of every shot in the machine. We discuss several methods employed to verify that the longitudinal profiles coming from simulations are accurate. We can use this information tomore » understand which particles are accelerated in each bunch, and by how much. Additionally, we use the longitudinal information to choose a subset of shots that always have the same incoming profiles to see the differing acceleration experienced by those shots as we vary the plasma density and length. This allows a more robust calculation of achieved gradient, as well as illuminating the effect of transverse deflections on that acceleration. Finally, we discuss other applications, as the technique for measuring the energy spectra and for matching to simulations is quite general.« less