About Me

I am a 4th Yr. Ph.D student at the University of California, Santa Cruz in the Computer Science and Engineering department advised by Prof. Heiner Litz. Currently, I am working as Graduate Student Researcher in the Center for Research in Systems and Storage (CRSS).

Currently, I am working on characterizing the performance of Approximate Nearest Neighbors (ANN) algorithms on modern CPUs (Intel Sapphire Rapids / NVIDIA Grace) and GPUs (Tensor Core A100 / H100). Especially, I am studying the scenarios where the size of the vector index is too large for the GPU memory and CPU-GPU memory management techniques such as CUDA Unified Memory (aka. UVM) and System memory (on NVIDIA Grace Hopper) come into play. Although, these techniques allow the ANN searches to not OOM (Out of Memory) by using the CPU memory as a backup, the performance degrades exponentially due to the high overhead of page faults. I am looking for ways to get around this overhead using techniques such as prefetching.

Previously, I have interned with companies such as NVIDIA, InfluxData, and JP Morgan & Co.

When I am not working, I like to go for long drives, explore new places, and try out new cuisines !

Research Interests

• Accelerators / GPUs / xPUs
• Heterogeneous systems
• Databases and data management systems

Contact

E-mail | LinkedIn | GitHub | Google Scholar | Medium | Twitter | CRSS

Taking a nap under a tree once helped a guy solve the universe !