EVAX paper proposes on-demand defenses for microarchitectural attacks
Authors report cutting InvisiSpec overhead from 27% to 1.26% and Fencing from 74% to 3.45%.
4 stories, newest first · one card per story
Authors report cutting InvisiSpec overhead from 27% to 1.26% and Fencing from 74% to 3.45%.
Converts latency differences into architectural values via parallel threads and shared memory.
TranScope detects LLM training membership from TLB and accelerator timing without a surrogate model.

Lemire's analysis shows AMD Ryzen X3D chips gained ~50% Geekbench 6 performance in two years through wider cores, larger caches, and 512-bit SIMD units.
Today 1.5K readers · 2.4K views · 753 reads · all time 189K views