* [PATCH] optimization docs part 2
@ 2002-08-30 8:50 Jerome L Quinn
2002-09-16 10:39 ` Geoff Keating
0 siblings, 1 reply; 3+ messages in thread
From: Jerome L Quinn @ 2002-08-30 8:50 UTC (permalink / raw)
To: gcc-patches
Here's the flag listing for the -O flags. I scanned the sources to try to
get these lists right, but there may still be errors. Please let me know
and I'll fix up the lists to match reality.
2002-08-30 Jerry Quinn <jlquinn@us.ibm.com>
* gcc/gcc/doc/invoke.texi (Optimization Options): List the options
enabled by each -O flag.
--- gcc/gcc/doc/invoke.texi.safe 2002-08-25 01:03:45.000000000
-0400
+++ gcc/gcc/doc/invoke.texi 2002-08-25 01:05:21.000000000 -0400
@@ -3196,6 +3196,22 @@
time, without performing any optimizations that take a great deal of
compilation time.
+@option{-O} turns on the following optimizations:
+@gccoptlist{-fdefer-pop
+-fmerge-constants
+-fthread-jumps
+-floop-optimize
+-fcrossjumping
+-fif-conversion
+-fif-conversion2
+-fdelayed-branch
+-fguess-branch-probability
+-fcprop-registers}
+
+@option{-O} also turns on @option{-fomit-frame-pointer} on machines
+where doing so does not interfere with debugging.
+
+
@item -O2
@opindex O2
Optimize even more. GCC performs nearly all supported optimizations
@@ -3204,11 +3220,24 @@
As compared to @option{-O}, this option increases both compilation time
and the performance of the generated code.
-@option{-O2} turns on all optional optimizations except for loop
-unrolling, function inlining, and register renaming. It also turns on
-the @option{-fforce-mem} and @option{-fstrict-aliasing} option on all
-machines and frame pointer elimination on machines where doing so does
-not interfere with debugging.
+@option{-O2} turns on all optimizations specified by @option{-O}. It
+also turns on the following optimizations:
+@gccoptlist{-fforce-mem
+-foptimize-sibling-calls
+-fstrength-reduce
+-fcse-follow-jumps -fcse-skip-blocks
+-frerun-cse-after-loop -frerun-loop-opt
+-fgcse -fgcse-lm -fgcse-sm
+-fdelete-null-pointer-checks
+-fexpensive-optimizations
+-fregmove
+-fschedule-insns -fschedule-insns2
+-fcaller-saves
+-fpeephole2
+-freorder-blocks -freorder-functions
+-fstrict-aliasing
+-falign-functions -falign-jumps
+-falign-loops -falign-labels}
Please note the warning under @option{-fgcse} about
invoking @option{-O2} on programs that use computed gotos.
@@ -3221,7 +3250,7 @@
@item -O0
@opindex O0
-Do not optimize.
+Do not optimize. This is equivalent to not using any optimization flags.
@item -Os
@opindex Os
@@ -3229,6 +3258,11 @@
do not typically increase code size. It also performs further
optimizations designed to reduce code size.
+@option{-Os} disables the following optimization flags:
+@gccoptlist{-falign-functions -falign-jumps -falign-loops
+-falign-labels -fprefetch-loop-arrays}
+
+
If you use multiple @option{-O} options, with or without level numbers,
the last such option is the one that is effective.
@end table
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH] optimization docs part 2
2002-08-30 8:50 [PATCH] optimization docs part 2 Jerome L Quinn
@ 2002-09-16 10:39 ` Geoff Keating
0 siblings, 0 replies; 3+ messages in thread
From: Geoff Keating @ 2002-09-16 10:39 UTC (permalink / raw)
To: Jerome L Quinn; +Cc: gcc-patches
Jerome L Quinn <jlquinn@us.ibm.com> writes:
> Here's the flag listing for the -O flags. I scanned the sources to try to
> get these lists right, but there may still be errors. Please let me know
> and I'll fix up the lists to match reality.
This patch also requires an assignment.
> 2002-08-30 Jerry Quinn <jlquinn@us.ibm.com>
> * gcc/gcc/doc/invoke.texi (Optimization Options): List the options
> enabled by each -O flag.
Please make sure that the ChangeLog entry you finally commit is
properly formatted: two spaces after the date and between the name and
the e-mail address, one blank line after the header line, and one tab
before every line of the body.
> --- gcc/gcc/doc/invoke.texi.safe 2002-08-25 01:03:45.000000000
> -0400
> +++ gcc/gcc/doc/invoke.texi 2002-08-25 01:05:21.000000000 -0400
> @@ -3196,6 +3196,22 @@
> time, without performing any optimizations that take a great deal of
> compilation time.
>
> +@option{-O} turns on the following optimizations:
I would say "turns on the following optimization flags", instead,
since this isn't a complete list of optimizations that are switched on by
-O1, only those that have separate flags.
> +@gccoptlist{-fdefer-pop
> +-fmerge-constants
> +-fthread-jumps
> +-floop-optimize
> +-fcrossjumping
> +-fif-conversion
> +-fif-conversion2
> +-fdelayed-branch
> +-fguess-branch-probability
> +-fcprop-registers}
> +
> +@option{-O} also turns on @option{-fomit-frame-pointer} on machines
> +where doing so does not interfere with debugging.
> +
> +
Why the two new blank lines here?
> @item -O0
> @opindex O0
> -Do not optimize.
> +Do not optimize. This is equivalent to not using any optimization flags.
I think it would be clearer to write "This is the default.". -O0 is
not the same as -O1 followed by -fno-defer-pop and so on.
> @item -Os
> @opindex Os
> @@ -3229,6 +3258,11 @@
> do not typically increase code size. It also performs further
> optimizations designed to reduce code size.
>
> +@option{-Os} disables the following optimization flags:
> +@gccoptlist{-falign-functions -falign-jumps -falign-loops
> +-falign-labels -fprefetch-loop-arrays}
> +
> +
Should this also say something like "This enables the following
optimisation flags" (plus a list of the flags)?
Thank you for trying to do this, I find the existing structure (where
flags are listed in order according to what -Ox flag enables them)
confusing myself, and I know others also find it confusing because they
keep putting new flags in the wrong place.
--
- Geoffrey Keating <geoffk@geoffk.org>
^ permalink raw reply [flat|nested] 3+ messages in thread
* [PATCH] Optimization docs part 2
@ 2002-12-16 22:09 Jerry Quinn
0 siblings, 0 replies; 3+ messages in thread
From: Jerry Quinn @ 2002-12-16 22:09 UTC (permalink / raw)
To: gcc-patches
This patch adds the list of -O levels that each flag is active
at. Hopefully there aren't any bugs here.
2002-12-17 Jerry Quinn <jlquinn@optonline.net>
* gcc/doc/invoke.texi (Optimization Options): List -O levels
for each optimization flag.
--- gcc/doc/invoke.texi.p2 Tue Dec 17 00:32:16 2002
+++ gcc/doc/invoke.texi Tue Dec 17 01:01:29 2002
@@ -3517,13 +3517,17 @@
the compiler normally lets arguments accumulate on the stack for several
function calls and pops them all at once.
+Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -fforce-mem
@opindex fforce-mem
Force memory operands to be copied into registers before doing
arithmetic on them. This produces better code by making all memory
references potential common subexpressions. When they are not common
subexpressions, instruction combination should eliminate the separate
-register-load. The @option{-O2} option turns on this option.
+register-load.
+
+Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
@item -fforce-addr
@opindex fforce-addr
@@ -3546,10 +3550,14 @@
whether a target machine supports this flag. @xref{Registers,,Register
Usage, gccint, GNU Compiler Collection (GCC) Internals}.
+Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -foptimize-sibling-calls
@opindex foptimize-sibling-calls
Optimize sibling and tail recursive calls.
+Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -ftrapv
@opindex ftrapv
This option generates traps for signed overflow on addition, subtraction,
@@ -3571,6 +3579,8 @@
declared @code{static}, then the function is normally not output as
assembler code in its own right.
+Enabled at level @option{-O3}.
+
@item -finline-limit=@var{n}
@opindex finline-limit
By default, gcc limits the size of functions that can be inlined. This flag
@@ -3614,6 +3624,8 @@
linker support it. Use @option{-fno-merge-constants} to inhibit this
behavior.
+Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -fmerge-all-constants
Attempt to merge identical constants and identical variables.
@@ -3638,6 +3650,9 @@
This option is only meaningful on architectures that support such
instructions, which include x86, PowerPC, IA-64 and S/390.
+The default is @option{-fbranch-count-reg}, enabled when
+@option{-fstrength-reduce} is enabled.
+
@item -fno-function-cse
@opindex fno-function-cse
Do not put function addresses in registers; make each instruction that
@@ -3647,6 +3662,8 @@
that alter the assembler output may be confused by the optimizations
performed when this option is not used.
+The default is @option{-ffunction-cse}
+
@item -ffast-math
@opindex ffast-math
Sets @option{-fno-math-errno}, @option{-funsafe-math-optimizations}, @*
@@ -3770,6 +3787,8 @@
Perform the optimizations of loop strength reduction and
elimination of iteration variables.
+Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -fthread-jumps
@opindex fthread-jumps
Perform optimizations where we check to see if a jump branches to a
@@ -3778,6 +3797,8 @@
second branch or a point immediately following it, depending on whether
the condition is known to be true or false.
+Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -fcse-follow-jumps
@opindex fcse-follow-jumps
In common subexpression elimination, scan through jump instructions
@@ -3786,6 +3807,8 @@
@code{else} clause, CSE will follow the jump when the condition
tested is false.
+Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -fcse-skip-blocks
@opindex fcse-skip-blocks
This is similar to @option{-fcse-follow-jumps}, but causes CSE to
@@ -3794,15 +3817,21 @@
@option{-fcse-skip-blocks} causes CSE to follow the jump around the
body of the @code{if}.
+Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -frerun-cse-after-loop
@opindex frerun-cse-after-loop
Re-run common subexpression elimination after loop optimizations has been
performed.
+Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -frerun-loop-opt
@opindex frerun-loop-opt
Run the loop optimizer twice.
+Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -fgcse
@opindex fgcse
Perform a global common subexpression elimination pass.
@@ -3813,6 +3842,8 @@
the global common subexpression elimination pass by adding
@option{-fno-gcse} to the command line.
+Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -fgcse-lm
@opindex fgcse-lm
When @option{-fgcse-lm} is enabled, global common subexpression elimination will
@@ -3820,6 +3851,8 @@
allows a loop containing a load/store sequence to be changed to a load outside
the loop, and a copy/store within the loop.
+Enabled by default when gcse is enabled.
+
@item -fgcse-sm
@opindex fgcse-sm
When @option{-fgcse-sm} is enabled, A store motion pass is run after global common
@@ -3827,17 +3860,23 @@
When used in conjunction with @option{-fgcse-lm}, loops containing a load/store sequence
can be changed to a load before the loop and a store after the loop.
+Enabled by default when gcse is enabled.
+
@item -floop-optimize
@opindex floop-optimize
Perform loop optimizations: move constant expressions out of loops, simplify
exit test conditions and optionally do strength-reduction and loop unrolling as
well.
+Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -fcrossjumping
@opindex crossjumping
Perform cross-jumping transformation. This transformation unifies equivalent code and save code size. The
resulting code may or may not perform better than without cross-jumping.
+Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -fif-conversion
@opindex if-conversion
Attempt to transform conditional jumps into branch-less equivalents. This
@@ -3845,11 +3884,15 @@
some tricks doable by standard arithmetics. The use of conditional execution
on chips where it is available is controlled by @code{if-conversion2}.
+Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -fif-conversion2
@opindex if-conversion2
Use conditional execution (where available) to transform conditional jumps into
branch-less equivalents.
+Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -fdelete-null-pointer-checks
@opindex fdelete-null-pointer-checks
Use global dataflow analysis to identify and eliminate useless checks
@@ -3862,10 +3905,14 @@
@option{-fno-delete-null-pointer-checks} to disable this optimization
for programs which depend on that behavior.
+Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -fexpensive-optimizations
@opindex fexpensive-optimizations
Perform a number of minor optimizations that are relatively expensive.
+Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -foptimize-register-move
@itemx -fregmove
@opindex foptimize-register-move
@@ -3873,18 +3920,21 @@
Attempt to reassign register numbers in move instructions and as
operands of other simple instructions in order to maximize the amount of
register tying. This is especially helpful on machines with two-operand
-instructions. GCC enables this optimization by default with @option{-O2}
-or higher.
+instructions.
Note @option{-fregmove} and @option{-foptimize-register-move} are the same
optimization.
+Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -fdelayed-branch
@opindex fdelayed-branch
If supported for the target machine, attempt to reorder instructions
to exploit instruction slots available after delayed branch
instructions.
+Enabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -fschedule-insns
@opindex fschedule-insns
If supported for the target machine, attempt to reorder instructions to
@@ -3893,6 +3943,8 @@
by allowing other instructions to be issued until the result of the load
or floating point instruction is required.
+Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -fschedule-insns2
@opindex fschedule-insns2
Similar to @option{-fschedule-insns}, but requests an additional pass of
@@ -3900,6 +3952,8 @@
especially useful on machines with a relatively small number of
registers and where memory load instructions take more than one cycle.
+Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -fno-sched-interblock
@opindex fno-sched-interblock
Don't schedule instructions across basic blocks. This is normally
@@ -3956,8 +4010,7 @@
This option is always enabled by default on certain machines, usually
those which have no call-preserved registers to use instead.
-For all machines, optimization level 2 and higher enables this flag by
-default.
+Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
@item -ftracer
@opindex ftracer
@@ -3984,6 +4037,8 @@
If supported by the target machine, generate instructions to prefetch
memory to improve the performance of loops that access large arrays.
+Disabled at level @option{-Os}.
+
@item -fmove-all-movables
@opindex fmove-all-movables
Forces all invariant computations in loops to be moved
@@ -4020,6 +4075,9 @@
are implemented in the compiler; some targets use one, some use the
other, a few use both.
+@option{-fpeephole} is enabled by default.
+@option{-fpeephole2} enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -fbranch-probabilities
@opindex fbranch-probabilities
After running a program compiled with @option{-fprofile-arcs}
@@ -4057,11 +4115,16 @@
reduce non-determinism, possibly at the expense of inferior
optimization.
+The default is @option{-fguess-branch-probability} at levels
+@option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -freorder-blocks
@opindex freorder-blocks
Reorder basic blocks in the compiled function in order to reduce number of
taken branches and improve code locality.
+Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -freorder-functions
@opindex freorder-functions
Reorder basic blocks in the compiled function in order to reduce number of
@@ -4074,6 +4137,8 @@
Also profile feedback must be available in to make this option effective. See
@option{-fprofile-arcs} for details.
+Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
+
@item -fstrict-aliasing
@opindex fstrict-aliasing
Allows the compiler to assume the strictest aliasing rules applicable to
@@ -4119,8 +4184,7 @@
allowed to alias. For an example, see the C front-end function
@code{c_get_alias_set}.
-For all machines, optimization level 2 and higher enables this flag by
-default.
+Enabled at levels @option{-O2}, @option{-O3}, @option{-Os}.
@item -falign-functions
@itemx -falign-functions=@var{n}
@@ -4139,6 +4203,8 @@
If @var{n} is not specified, use a machine-dependent default.
+Enabled at levels @option{-O2}, @option{-O3}.
+
@item -falign-labels
@itemx -falign-labels=@var{n}
@opindex falign-labels
@@ -4153,6 +4219,8 @@
If @var{n} is not specified, use a machine-dependent default which is
very likely to be @samp{1}, meaning no alignment.
+Enabled at levels @option{-O2}, @option{-O3}.
+
@item -falign-loops
@itemx -falign-loops=@var{n}
@opindex falign-loops
@@ -4163,6 +4231,8 @@
If @var{n} is not specified, use a machine-dependent default.
+Enabled at levels @option{-O2}, @option{-O3}.
+
@item -falign-jumps
@itemx -falign-jumps=@var{n}
@opindex falign-jumps
@@ -4173,6 +4243,8 @@
If @var{n} is not specified, use a machine-dependent default.
+Enabled at levels @option{-O2}, @option{-O3}.
+
@item -fssa
@opindex fssa
Perform optimizations in static single assignment form. Each function's
@@ -4203,11 +4275,15 @@
make debugging impossible, since variables will no longer stay in
a ``home register''.
+Enabled at levels @option{-O3}.
+
@item -fno-cprop-registers
@opindex fno-cprop-registers
After register allocation and post-register allocation instruction splitting,
we perform a copy-propagation pass to try to reduce scheduling dependencies
and occasionally eliminate the copy.
+
+Disabled at levels @option{-O}, @option{-O2}, @option{-O3}, @option{-Os}.
@item --param @var{name}=@var{value}
@opindex param
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2002-12-17 6:09 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2002-08-30 8:50 [PATCH] optimization docs part 2 Jerome L Quinn
2002-09-16 10:39 ` Geoff Keating
2002-12-16 22:09 [PATCH] Optimization " Jerry Quinn
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).