Simulation technology plays an important role in propulsion test facility design and development by assessing risks, identifying failure modes, and predicting anomalous behavior of critical systems. Integrated analyses of facility designs including supersonic diffusers, steam ejectors, valves, cooling spray nozzles, and turning ducts will be carried out for both steady state operation and shutdown/startup transients. The proposed innovation expands on the multi-element unstructured Computational Fluid Dynamics (CFD) which has been validated for complex valve/feed systems and high pressure propellant delivery systems used in engine and component test stands. The objective is to speed-up analyses time of heat transfer and thermal effects through the utilization of Graphic Processing Unit (GPU) enabled computing. The focus is on porting the thermal solver used in CFD multi-physics simulation framework to GPU architectures, so that GPU computing technology can permit modeling tasks that were restricted to high-end computational facilities to be achieved on local workstations. This will significantly enhance current analysis framework by facilitating analyses of complex heat transfer problems on desktop workstations with rapid turnaround. This capability will be expanded to include advanced models for analysis of non-equilibrium two-phase flow dynamics and heat transfer in water injection systems in the flame deflector, steam loaded plume entrainment, chemical steam generator performance, and operation of steam ejectors. The objective of this effort was to identify a novel approach to accelerate the time required for computing high-fidelity heat transfer problems using graphics hardware (GPUs). Traditionally, these types of analyses have been carried out on multiple processors or PC clusters that require significant costs for establishing and maintaining the computing environment. This project aggressively investigated the method of using of GPU enabled computing, to expedite analyses time of heat transfer and thermal effects. To accomplish this, the algorithms used were innovatively re-engineered for use on desktop computers with GPU enabled architectures which resulted in a substantial reduction in the cost of carrying out analyses as well as shortened turnaround time. Preliminary studies demonstrated that when using the GPU-ported heat transfer package, runtime acceleration factors exceeded 800%.. Enabling the capability to gain large levels of computational acceleration is significant due to the fact that it permits desktop workstation computing platforms to perform high-fidelity flow modeling of transient heat loads in suitable and usable timeframes. The findings from this study are additionally significant because it demonstrated a cost-efficient and effective path for exploiting an untapped level of parallelism can successfully be used for high fidelity simulation. The transitioning to GPU computing will give NASA personnel an opportunity to assess the potential applicability of these platforms to other software support systems.