T783 regresses severely in allocations with new pattern match checker
The new pattern match checker (Phab:D1535) allocates 115% more than the previous checker on the T783 testcase,
bytes allocated value is too high: Expected T783(normal) bytes allocated: 526230456 +/-10% Lower bound T783(normal) bytes allocated: 473607410 Upper bound T783(normal) bytes allocated: 578853502 Actual T783(normal) bytes allocated: 1134085384 Deviation T783(normal) bytes allocated: 115.5 % *** unexpected stat test failure for T783(normal)
I suspect this isn't avoidable as this testcase consists of nothing more than 500 guarded equations, so exercises the checker quite thoroughly. That being said, perhaps it's worth a closer look.