[m-rev.] for review: proper fix for deep profiling bug

[Zoltan Somogyi]

I forgot to mention, I also got some benchmark results. Here they are.
It seems that --should-pretest-equality is a win for speed, though it
causes a size increase of about 1%.

Zoltan.

EXTRA_MCFLAGS = --should-pretest-equality
EXTRA_CFLAGS = -DMR_DISABLE_CHECK_DU_EQ
GRADE = asm_fast.gc
mercury_compile.01 average of 6 with ignore=1     19.25

EXTRA_MCFLAGS = --no-should-pretest-equality
EXTRA_CFLAGS = -DMR_DISABLE_CHECK_DU_EQ
GRADE = asm_fast.gc
mercury_compile.02 average of 6 with ignore=1     20.39

EXTRA_MCFLAGS = --should-pretest-equality
EXTRA_CFLAGS = 
GRADE = asm_fast.gc
mercury_compile.03 average of 6 with ignore=1     19.72

EXTRA_MCFLAGS = --no-should-pretest-equality
EXTRA_CFLAGS = 
GRADE = asm_fast.gc
mercury_compile.04 average of 6 with ignore=1     19.83

EXTRA_MCFLAGS = --should-pretest-equality
EXTRA_CFLAGS = 
GRADE = hlc.gc
mercury_compile.07 average of 6 with ignore=1     20.17

EXTRA_MCFLAGS = --no-should-pretest-equality
EXTRA_CFLAGS = 
GRADE = hlc.gc
mercury_compile.08 average of 6 with ignore=1     21.61

--------------------------------------------------------------------------
mercury-reviews mailing list
Post messages to:       mercury-reviews at csse.unimelb.edu.au
Administrative Queries: owner-mercury-reviews at csse.unimelb.edu.au
Subscriptions:          mercury-reviews-request at csse.unimelb.edu.au
--------------------------------------------------------------------------