Tanvir's pic

Tanvir Ahmed Khan

I am a 5th-year Ph.D. candidate in the Electrical Engineering and Computer Science Department at the University of Michigan. My research interest lies in the intersection of compilers, operating systems, and computer architecture. Specifically, I am interested in designing novel profile-guided optimization mechanisms to improve instruction and data locality of modern data center applications. A number of such optimization mechanisms that I proposed are being adopted by real-world data center processor designers like Intel. My performance analysis methodology made real-world workloads twice as faster and consequently was adopted in the Arm Neoverse N1 Core, the leading data center CPU powering 49% of all Amazon Web Service machines.

My Ph.D. advisor is Baris Kasikci. I also collaborate with Barzan Mozafari (University of Michigan), Gilles Pokam (Intel Labs), Joseph Devietti (University of Pennsylvania), Daniel A. Jiménez (Texas A&M University), and Heiner Litz (University of California, Santa Cruz). Before starting my Ph.D. at the University of Michigan, I was a Lecturer in the Computer Science and Engineering Department at the Bangladesh University of Engineering and Technology. I also earned my M.Sc. and B.Sc. in Computer Science and Engineering from the Bangladesh University of Engineering and Technology. I also worked as a Research Intern at Microsoft and ARM and as a Software Engineer Intern at Facebook.

News

  • August 2022: I will serve on the CGO'23 Program Committee.
  • July 2022: OCOLOS (Online COde Layout OptimizationS to perform profile-guided optimizations on a running process) and Whisper (profile-guided branch misprediction elimination for data center applications) are accepted to appear at MICRO'22.
  • June 2022: One profile fits all (profile-guided Linux Kernel optimizations for data center applications) is invited to appear at SIGOPS Operating Systems Review. Congrats to Muhammed Ugur, Cheng Jiang, and Alex Erf who performed most of the work as part of their course project under my mentorship.
  • May 2022: Our proposal on data-driven processor design for data center applications is awarded TSA grant. Thanks Intel.
  • April 2022: I will serve on the MICRO'22 External Review Committee.
  • March 2022: I am awarded a Rackham Predoctoral Fellowship.
  • March 2022: Thermometer (profile-guided BTB replacement for data center applications) is accepted to appear at ISCA'22.
  • January 2022: APT-GET (profile-guided timely software prefetching) is accepted to appear at EuroSys'22.
  • January 2022: I am teaching the Advanced Operating Systems (EECS 582) course. If you are a University of Michigan student, consider attending the course.
  • October 2021: Shixin Song, an undergraduate student mentored by me, won first place at MICRO 2021 Student Research Competition (SRC).
  • October 2021: Our performance analysis methodology for data center applications is adopted for Arm Neoverse N1 Core, the leading data center CPU with a 49% AWS EC2 footprint. Driving profile-guided optimizations, this methodology made real-world workloads 2X faster. For more details, check out the full methodology and the blogpost.
  • September 2021: My summer 2020 Facebook internship work on Linux kernel binary layout optimization is presented at the Linux plumbers conference.
  • August 2021: I will serve on the ASPLOS'22 External Review Committee.
  • July 2021: Twig (profile-guided BTB prefetching for data center applications) and PDede (a new BTB design for data center applications) are accepted to appear at MICRO'21.
  • March 2021: DMon (efficient detection and correction of data locality problems using selective profiling) is accepted to appear at OSDI'21.
  • March 2021: Ripple (profile-guided instruction cache replacement for data center applications) is accepted to appear at ISCA'21.
  • December 2020: Persistent memory file mapping work (led by Ian Neal) is accepted to appear at FAST'21.
  • September 2020: I will serve on the EuroSys'21 Shadow PC.
  • August 2020: Video transcoding characterization work is accepted to appear at IISWC'20. Congrats to Yuhan Chen and Jingyuan Zhu who performed most of the work as part of their course project under my mentorship.
  • August 2020: Our proposal on improving instruction locality for data center applications is awarded FoMR grant. Thanks NSF and Intel.
  • July 2020: I will serve on the ASPLOS'21 External Review Committee.
  • July 2020: I-SPY is accepted to appear at MICRO'20.
  • June 2020: Started Internship with Facebook BOLT team.
  • February 2020: Nathan Brown, an undergraduate student mentored by me, won first place at CGO 2020 Student Research Competition (SRC).
  • January 2020: Facebook Fellowship Finalist.
  • January 2020: I will serve on the MLSys'20 Artifact Evaluation Committee.
  • December 2019: I will serve on the PLDI'20 Artifact Evaluation Committee.
  • December 2019: I will serve on the ASPLOS'20 Artifact Evaluation Committee.
  • August 2019: I will serve on the SOSP'20 Artifact Evaluation Committee.
  • June 2019: Presented Huron at PLDI'19. Thanks PLMW Student Scholarship for covering conference registration, accommodation, and travel.
  • May 2019: Started internship with Microsoft Azure.
  • April 2019: Huron artifact is awarded "Artifacts Available" and "Artifacts Evaluated Functional" badge.
  • February 2019: Huron is accepted to appear at PLDI'19.
  • October 2018: Attended OSDI. Thanks OSDI student grant for covering conference registration, accommodation, and travel.
  • August 2017: Started Ph.D. at UMich with Rollin M. Gerstacker Foundation Fellowships.