AArch64 mixed-size model

Experiments mostly consist in comparing model output with hardware output, see here.

Tested systems are identified by CPU commercial names. We provide our test base of 2669 mixed-sized tests, as well as model and consolidated hardware logs. Notice that logs also contain output for non mixed-sized tests. Simulating those tests in mixed-size and non mixed-size modes yields no difference. As a consequence, those non mixed-size tests do not appear in tables below.

• Comparison with hardware
  • Invalid behaviours
  • Observed Read-after-Read hazards
  • New this round
• Additional comparison
  • Comparison with the operational (FLAT) model
    • More behaviours
    • Less behaviours

This document was translated from L^AT_EX by H^EV^EA.