Skip to content
MN

//Now · updated 24 August 2026

What I'm doing now.

A now page, in the spirit of nownow.com: the short list of things actually in flight, kept current or deleted.

Updated
24 August 2026
In flight
7
Deliberately not
3

7 active

  1. 01

    Twelve manuscripts out at once

    Four at IEEE Embedded Systems Letters, three at ACM TECS, two at IEEE Computer Architecture Letters, and one each at IEEE Internet of Things Journal, IEEE Transactions on Computers and IEEE Design & Test. Not one decision back yet. The next move on each is whatever the reviewers say.

    In review
  2. 02

    Two waiting on a portal to open

    The quantization-format paper is finished and waiting on HotMobile, which does not reopen until roughly October. The GPU profiling methodology paper is waiting on the EuroMLSys call, expected somewhere between November and January, and its own number audit says fix two things first.

    Queued
  3. 03

    Server-class measurement

    Bringing the Cisco UCS bench up for NUMA and CPU-inference energy work: the same memory-wall thesis, one scale up, on hardware old enough to be honest about.

    Building
  4. 04

    Open-source contributions

    Three pull requests merged into NVIDIA garak, seven still open there, and four open at vLLM, in the same areas the research already touches: inference performance and LLM evaluation.

    In motion
  5. 05

    Artifact evaluation for ATC '26 (ACM SIGOPS, formerly USENIX ATC)

    On the Artifact Evaluation Committee. The review window is 22 September to 14 October 2026, and it is the fixed point the rest of the autumn has to work around.

    Committed
  6. 06

    JOSS reviewing

    First review delivered on 19 August: Optiland, an optical design package, 31 checklist items and three issues that the authors resolved. A second review is open now for a library on non-stationary extreme value distributions, and a third assignment is waiting on its editor.

    Reviewing
  7. 07

    One talk still out

    A talk submitted to Embedded Vision Summit 2027, a vision-framed treatment of the measurement work. Not accepted yet. The PyTorch Conference poster was declined on capacity in August.

    Submitted

Deliberately not doing

  • × Adding a newsletter
  • × Chasing model-of-the-week benchmarks
  • × Writing about tools I have not measured

If this page looks stale, it is. The date at the top is the contract.