Résumé
The need of non volatility along with the added flexibility of un limited reprogramming like SRAM has lead to the concept of universal memories. MRAM (magnetoresistive random access memory) is one prominent member of them. At present only Flash is providing a limited bridge for that. Flash based FPGAs have several benefits being non volatile but unfortunately also loose many of the features which are only possible with SRAM based FPGAs. MRAMs have potential to bridge this gap. This paper will present a brief survey of our work in this regard for creating the entire eco system of software and hardware tool flows, MRAM layout work at 120 nm, exploration environments to conduct complex experiments especially dynamic reconfiguration and multi context FPGAs. MRAM opens new opportunities for them compared to SRAM and Flash. It will discuss the current status of MRAM in industry and our current and future test chips road maps. Provide several references to industry and our published work for details about MRAMs and eFPGAs, to show why we think MRAM can be very interesting element for FPGAs.