The reaction-rate ratio C8/Ftot, neutron captures in 238U to total fissions, has been measured in 80 out of 96 fuel rods of a Westinghouse SVEA-96+ boiling water reactor fuel assembly. High-resolution gamma spectroscopy was performed on individual fuel rods, withdrawn from the SVEA-96+ assembly after irradiation at low power in the center of the LWR-PROTEUS reactor core. Absolute experimental errors of 1.7% and relative errors of 0.6% (for rod-to-rod ratios) were achieved. The experimental results were used as a database for validation of four different calculational tools: CASMO-4 and HELIOS as commercial assembly codes, the Paul Scherrer Institute in-house code BOXER, and the Monte Carlo transport code MCNPX. In general, on the level of a few percent, there is good agreement between experiment and calculations, the use of a recently proposed 239Np gamma-ray emission probability improving even further the agreement. However, the highly heterogeneous design of the SVEA-96+ assembly (both in terms of material compositions and neutron moderation conditions) causes some problems. Clear deviations from assembly mean values are found among the burnable absorber fuel rods that are grouped in clusters (direct neighbors), a unique feature of this assembly design. For these rods the codes overpredict C8/Ftot by several percent, including MCNPX. Additional trends, not present in the results from the Monte Carlo calculation which generally shows the best overall agreement with experiment, are identified for the deterministic codes.