The scalability limitations of current DRAM technology
have prompted research in alternative memory technologies.
Almost all new viable alternatives being explored are non-volatile
in nature. In this talk, we explore the implication of emerging
non-volatile memory technologies (NVRAM) to mobile computing,
the persistence programming models inspired by the emergence
of NVRAM, and the challenges of managing persistent memory.
Finally, we present Makalu, an approach that represents
one possible solution to the persistent memory
management challenges. Makalu avoids failure-induced
persistent memory leaks, offers interoperability with
a variety of NVRAM programming libraries (NVMPLs),
while supporting a less restrictive programming model
and delivering better performance compared to existing
NVMPLs’ default allocators.

I am a PhD student advised by Prof. Vivek Sarkar in the department of computer science at Rice University. My research interest is in understanding and addressing programming and performance challenges associated with the next generation of systems with byte-addressable non-volatile memory (NVRAM).