The BXD strains are an important reference population of recombinant inbred lines which have been phenotyped extensively. To facilitate interactive use of genotype-phenotype relationships in this population, we sought to speed up eQTL scans where we perform a univariate genome scan for every trait in a collection of omic traits. By using easily parallelizable operations such as matrix multiplication, vectorized operations, and elementwise operations, we are able to decrease runtimes approaching real-time computation. We used parallelization using different CPU threads as well as GPUs. We found that the speed advantage of GPUs is dependent on problem size and shape. These results indicate a pathway for speeding up eQTL scans using LMMs. Our implementation is in the Julia programming language.