[llvm-dev] Memory Store/Load Optimization Issue (Emulating stack)
Hal Finkel via llvm-dev
llvm-dev at lists.llvm.org
Wed Feb 10 13:41:43 PST 2016
----- Original Message -----
> From: "Paul Peet" <paulpeet17 at gmail.com>
> To: "Hal Finkel" <hfinkel at anl.gov>
> Cc: "llvm-dev" <llvm-dev at lists.llvm.org>, "Daniel Berlin" <dberlin at dberlin.org>
> Sent: Wednesday, February 10, 2016 3:33:08 PM
> Subject: Re: [llvm-dev] Memory Store/Load Optimization Issue (Emulating stack)
>
>
>
>
> 2016-02-10 22:23 GMT+01:00 Hal Finkel < hfinkel at anl.gov > :
>
>
> ----- Original Message -----
> > From: "Paul Peet via llvm-dev" < llvm-dev at lists.llvm.org >
> > To: "Daniel Berlin" < dberlin at dberlin.org >
> > Cc: "llvm-dev" < llvm-dev at lists.llvm.org >
> > Sent: Wednesday, February 10, 2016 3:13:15 PM
> > Subject: Re: [llvm-dev] Memory Store/Load Optimization Issue
> > (Emulating stack)
> >
> >
> >
> > Thanks for the answers. Although I am not sure if I've understood
> > the
> > docs about how inttoptr/ptrtointr are different when compared to
> > gep.
> > It says: "It’s invalid to take a GEP from one object, address into
> > a
> > different separately allocated object, and dereference it.".
>
> This refers to the underlying allocation that created the memory.
> Where did %sp come from? Is it an alloca instruction, or from some
> other source?
>
>
>
>
> It's allocated via malloc and passed to the function.
>
> define { i32, i32, i8* } @test(i32 %foo, i32 %bar, i8* %sp_x)
>
Then you're fine unless you run off the end of the memory you malloc'd. This is not saying anything tricky here: No buffer overruns. Even if there happens to be another valid object there, you can't run off the end of your allocated buffer (do so is undefined behavior).
>
> > To go back to my intention why I am doing this, I would like to
> > "emulate" some x86 instructions with llvm-ir but as far as I
> > understand that aliasing rule, I am not sure if I am breaking that
> > rule.
> >
> >
> > For example when translating this x86 code to llvm ir:
> >
> >
> > push eax
> > add esp, 2
> > push ecx
> > ...
> >
> >
> >
> > ; push foo (On "stack")
> > %sp_1 = getelementptr i8, i8* %sp, i32 -4
> > %sp_1_ptr = bitcast i8* %sp_1 to i32*
> > store i32 %foo, i32* %sp_1_ptr, align 4
> >
> >
> > %sp_x = getelementptr i8, i8* %sp_1, i32 2
> >
> >
> > ; push bar
> > %sp_2 = getelementptr i8, i8* %sp_x, i32 -4
> > %sp_2_ptr = bitcast i8* %sp_2 to i32*
> > store i32 %bar, i32* %sp_2_ptr, align 4
> >
> >
> > Both objects (eax, ecx) will overlap because of the size difference
> > (eax = i32). What are the consequences when doing this. Will this
> > break alias analysis for the further instructions?
> >
>
> Partially overlapping writes to do not, in themselves, break
> anything. AA should handle that just fine.
>
>
> What do you mean by partially?
>
It looks like the first two bytes of %bar overlap with the second two bytes of %foo. That's a partial overlap of the writes (as the writes are four bytes each).
-Hal
>
> -Hal
>
>
>
> >
> > 2016-02-10 21:24 GMT+01:00 Daniel Berlin < dberlin at dberlin.org > :
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > On Wed, Feb 10, 2016 at 12:18 PM, Paul Peet via llvm-dev <
> > llvm-dev at lists.llvm.org > wrote:
> >
> >
> >
> > Thank you for the hint.
> >
> >
> > I adjusted the code and it works:
> >
> >
> > The code after replacing inttoptr with getelementptr:
> >
> >
> >
> > define { i32, i32, i8* } @test(i32 %foo, i32 %bar, i8* %sp) {
> > entry:
> > ; push foo (On "stack")
> > %sp_1 = getelementptr i8, i8* %sp, i32 -4
> > %sp_1_ptr = bitcast i8* %sp_1 to i32*
> > store i32 %foo, i32* %sp_1_ptr, align 4
> >
> >
> > ; push bar
> > %sp_2 = getelementptr i8, i8* %sp_1, i32 -4
> > %sp_2_ptr = bitcast i8* %sp_2 to i32*
> > store i32 %bar, i32* %sp_2_ptr, align 4
> >
> >
> > ; val1 = pop (val1 = bar)
> > %sp_3_ptr = bitcast i8* %sp_2 to i32*
> > %val1 = load i32, i32* %sp_3_ptr, align 4
> > %sp_3 = getelementptr i8, i8* %sp_2, i32 4
> >
> >
> > ; val2 = pop (val2 = foo)
> > %sp_4_ptr = bitcast i8* %sp_3 to i32*
> > %val2 = load i32, i32* %sp_4_ptr, align 4
> > %sp_4 = getelementptr i8, i8* %sp_3, i32 4
> >
> >
> > %ret_1 = insertvalue { i32, i32, i8* } undef, i32 %val1, 0
> > %ret_2 = insertvalue { i32, i32, i8* } %ret_1, i32 %val2, 1
> > %ret_3 = insertvalue { i32, i32, i8* } %ret_2, i8* %sp_4, 2
> >
> >
> > ret { i32, i32, i8* } %ret_3
> > }
> >
> >
> > After optimization ("opt -instcombine ./code.ll -S")
> >
> >
> >
> > define { i32, i32, i8* } @test(i32 %foo, i32 %bar, i8* %sp) {
> > entry:
> > %sp_1 = getelementptr i8, i8* %sp, i64 -4
> > %sp_1_ptr = bitcast i8* %sp_1 to i32*
> > store i32 %foo, i32* %sp_1_ptr, align 4
> > %sp_2 = getelementptr i8, i8* %sp, i64 -8
> > %sp_2_ptr = bitcast i8* %sp_2 to i32*
> > store i32 %bar, i32* %sp_2_ptr, align 4
> > %ret_1 = insertvalue { i32, i32, i8* } undef, i32 %bar, 0
> > %ret_2 = insertvalue { i32, i32, i8* } %ret_1, i32 %foo, 1
> > %ret_3 = insertvalue { i32, i32, i8* } %ret_2, i8* %sp, 2
> > ret { i32, i32, i8* } %ret_3
> > }
> >
> >
> > My only questions are now:
> > - How is it that inttoptr cannot provide that specific alias
> > information so it can optimize that store/load away ?
> > Because nothing tracks what happens to the ints, and what happens
> > when they are converted back to pointers and whether it's sane :)
> > http://llvm.org/docs/GetElementPtr.html#how-is-gep-different-from-ptrtoint-arithmetic-and-inttoptr
> >
> >
> >
> >
> >
> > - Might it be possible to get inttoptr providing such alias
> > analysis
> > ?
> > It doesn't make a lot of sense to try in most cases.
> > Most of the cases ptrtoint/inttoptr is useful are those where you
> > want to do crazy things to the pointer.
> >
> >
> >
> >
> >
> >
> > - I came across MemorySSA while browsing though the llvm source. Is
> > it possible that one can use MemorySSA to do such optimization
> > without alias analysis ?
> >
> >
> > MemorySSA relies on alias analysis to generate the SSA form.
> >
> >
> >
> >
> > - Where do I have to look in the source which is doing this kind of
> > optimization (Is it instcombine which uses lib/Analysis/Loads.cpp
> > ?)
> >
> >
> > It's probably a combination of opts. The most likely candidate is
> > -gvn, but I would look at the pass dumps after each opt
> >
> >
> >
> >
> >
> >
> > Regards,
> > Paul
> >
> >
> >
> >
> >
> >
> > 2016-02-10 0:26 GMT+01:00 Philip Reames < listmail at philipreames.com
> > >
> > :
> >
> >
> >
> > Two points:
> > - Using inttoptr is a mistake here. GEPs are strongly preferred and
> > provide strictly more aliasing information to the optimizer.
> > - The zext is a bit weird. I'm not sure where that came from, but
> > I'd
> > not bother looking into until the preceding point is addressed.
> >
> > In general, you may find these docs useful:
> > http://llvm.org/docs/Frontend/PerformanceTips.html
> >
> > Philip
> >
> >
> >
> >
> >
> > On 02/08/2016 06:54 AM, Paul Peet via llvm-dev wrote:
> >
> >
> >
> >
> >
> > Hello,
> >
> >
> > I am trying to emulate the "stack" as like on x86 when using
> > push/pop
> > so afterwards I can use LLVM's optimizer passes to simplify (reduce
> > junk) the code.
> >
> >
> > The LLVM IR code:
> >
> >
> >
> > define { i32, i32, i32 } @test(i32 %foo, i32 %bar, i32 %sp) {
> > ; push foo (On "stack")
> > %sp_1 = sub i32 %sp, 4
> > %sp_1_ptr = inttoptr i32 %sp_1 to i32*
> > store i32 %foo, i32* %sp_1_ptr, align 4
> >
> >
> > ; push bar
> > %sp_2 = sub i32 %sp_1, 4
> > %sp_2_ptr = inttoptr i32 %sp_2 to i32*
> > store i32 %bar, i32* %sp_2_ptr, align 4
> >
> >
> > ; val1 = pop (val1 = bar)
> > %sp_3_ptr = inttoptr i32 %sp_2 to i32*
> > %val1 = load i32, i32* %sp_3_ptr, align 4
> > %sp_3 = add i32 %sp_2, 4
> >
> >
> > ; val2 = pop (val2 = foo)
> > %sp_4_ptr = inttoptr i32 %sp_3 to i32*
> > %val2 = load i32, i32* %sp_4_ptr, align 4
> > %sp_4 = add i32 %sp_3, 4
> >
> >
> > %ret_1 = insertvalue { i32, i32, i32 } undef, i32 %val1, 0
> > %ret_2 = insertvalue { i32, i32, i32 } %ret_1, i32 %val2, 1
> > %ret_3 = insertvalue { i32, i32, i32 } %ret_2, i32 %sp_4, 2
> >
> >
> > ret { i32, i32, i32 } %ret_3
> > }
> >
> >
> > This code will "push" two values onto the stack and pop them in
> > reverse order so afterwards "foo" and "bar" will be swapped and
> > returned back.
> >
> >
> > After running this through "opt -O2 ./test.ll", I am getting this:
> >
> >
> >
> > define { i32, i32, i32 } @test(i32 %foo, i32 %bar, i32 %sp) #0 {
> > %sp_1 = add i32 %sp, -4
> > %1 = zext i32 %sp_1 to i64
> > %sp_1_ptr = inttoptr i64 %1 to i32*
> > store i32 %foo, i32* %sp_1_ptr, align 4
> > %sp_2 = add i32 %sp, -8
> > %2 = zext i32 %sp_2 to i64
> > %sp_2_ptr = inttoptr i64 %2 to i32*
> > store i32 %bar, i32* %sp_2_ptr, align 4
> > %val2 = load i32, i32* %sp_1_ptr, align 4
> > %ret_1 = insertvalue { i32, i32, i32 } undef, i32 %bar, 0 ; Swapped
> > %ret_2 = insertvalue { i32, i32, i32 } %ret_1, i32 %val2, 1; Not
> > Swapped (Not optimized; Should be %foo)
> > %ret_3 = insertvalue { i32, i32, i32 } %ret_2, i32 %sp, 2
> > ret { i32, i32, i32 } %ret_3
> > }
> >
> >
> > As you can see that the IR has got additional code, eg. zext. But
> > the
> > main problem here is that val2 hasn't been optimized.
> > Could anyone show me some hints what is preventing the second val
> > from being optimized? (My guess would be the zext because I am
> > using
> > %sp as a 32bit pointer although the "target" is 64bit).
> >
> >
> > Regards,
> > Paul
> >
> > _______________________________________________
> > LLVM Developers mailing list llvm-dev at lists.llvm.org
> > http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
> >
> >
> > _______________________________________________
> > LLVM Developers mailing list
> > llvm-dev at lists.llvm.org
> > http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
> >
> >
> >
> >
> > _______________________________________________
> > LLVM Developers mailing list
> > llvm-dev at lists.llvm.org
> > http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
> >
>
>
>
> --
> Hal Finkel
> Assistant Computational Scientist
> Leadership Computing Facility
> Argonne National Laboratory
>
>
--
Hal Finkel
Assistant Computational Scientist
Leadership Computing Facility
Argonne National Laboratory
More information about the llvm-dev
mailing list