Dmitry Stogov
Dmitry Stogov
```asm mov 0x20(%rsi),%rax add $0x30,%rsi shl $0x4,%rax ; this instruction may be eliminated if offset is predcalculated movsd (%rbx,%rax,1),%xmm0 mov -0x20(%rsi),%rax shl $0x4,%rax ; this instruction may be eliminated if...
- built-in disassembler - ability to debug JIT-ed code with GDB (__jit_debug_register_code) - ability to monitor JIT-ed code with sampling profilers (Linux Perf, Oprofile, VTune)
``movsd `` modifies only a part of %xmm register and therefore slower than ``movaps``. I made the same mistake in PHP JIT, and for years though that AVX code is...
$ bin/perf2bolt -p perf.data -o perf.fdata hhvm PERF2BOLT: Starting data aggregation job for perf.data PERF2BOLT: Spawning perf-script job to read branch events PERF2BOLT: Spawning perf-script job to read mem events...