The LES/FDFapproachfor turbulent combustionoffers the benefitsof both large eddysimulation(LES) to treattheturbulentflow, andthePDFapproachto treatturbulence-chemistryinteractions(in termsof the filtereddensityfunction,FDF).Theapproachis implementedasaparticlemeshmethodandcomputationally themostexpensi veaspectis determiningthechangein particlecompositionoveratimestepdueto reaction. Thiscostcanbesignificantlyreducedby usingin situ adapti ve tabulation(ISAT). In thiswork weinvestigate the computationalperformanceof several strategies for the parallel implementationof ISAT in LES/FDF calculations.Thecapabilityof performingLES/FDF/ISAT computationsof turbulentflamesis developedby incorporatingthe ISAT algorithmin theStanfordstructuredlarge eddysimulation(LES) andcomposition “filtered densityfunction” (FDF) code. The LES/FDF/ISAT simulationof a spatiallydevelopingmixing layer is usedasthe testcaseto studythe performanceandload balancingof different ISAT strategiesfor idealizedturbulent flamesof bothhydrogenandmethane.Detailed9-speciesand35-speciesmechanisms areemployed for the hydrogenflameandthe methaneflame,respecti vely. The resultsshow that whenit is almostalwayspossibleto retrieve from the ISAT table,thenusingpurely local processing(without any messagepassing)is optimal. But whena significantnumberof direct integrationsof thechemicalkinetic equationis required,thenparallelstrategies,suchastheuniform randomdistribution (URAN) strategy, are advantageous. Finally, a simplemodelis developedto explain theobservedcomputationalperformanceof thedifferentparallelstrategiesin differentsimulations.