Solar flares form and release energy across a large number of magnetic loops. The global parameters of flares, such as the total energy released, duration, physical size, etc., are routinely measured, and the hydrodynamics of a coronal loop subjected to intense heating have been extensively studied. It is not clear, however, how many loops comprise a flare, nor how the total energy is partitioned between them. In this work, we employ a hydrodynamic model to better understand the energy partition by synthesizing Si iv and Fe xxi line emission and comparing to observations of these lines with the Interface Region Imaging Spectrograph (IRIS). We find that the observed temporal evolution of the Doppler shifts holds important information on the heating duration. To demonstrate this, we first examine a single loop model, and find that the properties of chromospheric evaporation seen in Fe xxi can be reproduced by loops heated for long durations, while persistent redshifts seen in Si iv cannot be reproduced by any single loop model. We then examine a multithreaded model, assuming both a fixed heating duration on all loops and a distribution of heating durations. For a fixed heating duration, we find that durations of 100–200 s do a fair job of reproducing both the red- and blueshifts, while a distribution of durations, with a median of about 50–100 s, does a better job. Finally, we compare our simulations directly to observations of an M-class flare seen by IRIS, and find good agreement between the modeled and observed values given these constraints.