Technical notes
Short write-ups of things I measured or argued. Published as pages rather than PDFs, so they can be read rather than claimed. Each one states what it does not establish as clearly as what it does.
Measured against your own flow: a VEK280 latency note
End-to-end latency on hardware came in 1.25 % from the Vitis simulation estimate, but only because both latency probes sat inside the programmable-logic domain. Also: the honest crossover condition where an AI Engine array stops beating a plain DSP implementation.
CERN openlab · CMS Level-1 TriggerReconfigurable compute under orbital thermal and radiation constraints
In an orbital data centre the radiator, not the die, sets sustained throughput. That makes duty cycle a control problem, argues for a fabric with a genuine degraded mode, and decides where fault-recovery authority has to live.
Written September 2026