@inproceedingsIru2018, Author = {Irune Yarza and Mikel Azkarate-Askasua and Kim Grüttner and Wolfgang Nebel}, Title = {Real-Time Capable Retargeting of Xilinx MicroBlaze Binaries using QEMU - A Feasibility Study}, Year = {2018}, Month = {1}, Booktitle = {10th Workshop on Rapid Simulation and Performance Evaluation: Methods and Tools (RAPIDO)}, type = {inproceedings}, Abstract = {The great expansion and fast evolution of embedded systems market and the known advantages of the use of multi-core processors in this area are generating interest on improved embedded systems technologies (e.g., shrinking transistor size, new on-chip architectures), which at the same time is shortening the obsolescence periods of the underlying hardware. As a consequence, software designed for those platforms (a.k.a legacy code), that might be functionally correct and validated code, will be lost when changing the underlying Instruction Set Architecture (ISA) and peripherals. Given that many embedded systems execute Real-Time (RT) applications, the legacy code migration problem directly affects RT systems. Dynamic Binary Translation (DBT) techniques have been widely used for the migration of legacy code to a new hardware platform. However, there are no works which consider their use for RT legacy code migration. Therefore, this paper analyzes the suitability of a DBT based emulator, Quick EMUlator (QEMU), as a mean for RT legacy code migration, mainly focusing on its temporal capacities. To this end, a bare-metal execution on a Xilinx MicroBlaze processor has been compared against a QEMU emulated MicroBlaze running on an ARM Cortex-A9 processor. Results show that the proposed approach could provide hard RT performance in 55% and soft RT performance in 74% of our considered benchmarks.} @COMMENTBibtex file generated on