llvm dev - Mar 2017 - -ffast-math optimizations impacted by symbol renaming in header files

Hi, 

I came across an issue where some optimizations that would normally be applied
to standard math function calls are not getting applied when the –ffast-math
option is enabled on the Clang command line on a Linux x86_64 target.

I tracked down the issue to occurring because the –ffast-math option is
triggering Clang to preprocess the math.h header file with the
__FINITE_MATH_ONLY__ macro set to 1. In this case, the Linux header files are
redirecting several math functions using the __asm__ extension to rename the
symbol. An example of what the header files contains is:
      extern double exp (double) __asm__ ("" "__exp_finite")
__attribute__ ((__nothrow__ ));

In effect, ‘exp’ gets converted to be ‘__exp_finite’ in the IR. Because some of
the optimizations like constant folding or vectorization are looking for the
original function name, those optimizations do not get triggered under
–ffast-math due to these alternative function names.

Other affected functions include: acos, asin, exp2, log, log2, log10, and a few
others.

Anybody know if this is intentional, or have input regarding whether these
optimizations should be extended to also look for the renamed versions of these
functions?

Thanks,
Chris



A couple of example cases: 

// File: test_folding.c 

// Constant fold will occur with: 
// clang -O2 -S -emit-llvm test_folding.c 

// Constant fold will not occur with: 
// clang -O2 -S -emit-llvm -ffast-math test_folding.c 

#include <math.h> 

double test_fold_exp() { 
  return exp(0.0); 
} 

double test_fold_acos() { 
  return acos(1.0); 
} 

---------------------------------------------- 

// File: test_vectorize.c 

// Vectorization will occur with: 
// clang -O2 -S -emit-llvm -fno-math-errno test_vectorize.c 

// Vectorization will not occur with: 
// clang -O2 -S -emit-llvm -fno-math-errno -ffast-math test_vectorize.c 

#include <math.h> 

void test_vectorize_exp(long n,  float* restrict y,  float* restrict x) { 
  long i; 
  for (i = 0; i < n; i++) { 
    x[i] = expf(y[i]); 
  } 
}

On 03/20/2017 11:27 AM, Chris Chrulski via llvm-dev
wrote:
> Hi,
>
> I came across an issue where some optimizations that would normally be
applied to standard math function calls are not getting applied when the
–ffast-math option is enabled on the Clang command line on a Linux x86_64
target.
>
> I tracked down the issue to occurring because the –ffast-math option is
triggering Clang to preprocess the math.h header file with the
__FINITE_MATH_ONLY__ macro set to 1. In this case, the Linux header files are
redirecting several math functions using the __asm__ extension to rename the
symbol. An example of what the header files contains is:
>        extern double exp (double) __asm__ (""
"__exp_finite") __attribute__ ((__nothrow__ ));
>
> In effect, ‘exp’ gets converted to be ‘__exp_finite’ in the IR. Because
some of the optimizations like constant folding or vectorization are looking for
the original function name, those optimizations do not get triggered under
–ffast-math due to these alternative function names.
>
> Other affected functions include: acos, asin, exp2, log, log2, log10, and a
few others.
>
> Anybody know if this is intentional, or have input regarding whether these
optimizations should be extended to also look for the renamed versions of these
functions?
It is not intentional; the optimizations should be extended. I imagine 
that we would extend TargetLibraryInfo to have names for these 
functions, and then we could recognize them when available, and generate 
them when available and the inputs have the 'fast' tag.

  -Hal
>
> Thanks,
> Chris
>
>
>
> A couple of example cases:
>
> // File: test_folding.c
>
> // Constant fold will occur with:
> // clang -O2 -S -emit-llvm test_folding.c
>
> // Constant fold will not occur with:
> // clang -O2 -S -emit-llvm -ffast-math test_folding.c
>
> #include <math.h>
>
> double test_fold_exp() {
>    return exp(0.0);
> }
>
> double test_fold_acos() {
>    return acos(1.0);
> }
>
> ----------------------------------------------
>
> // File: test_vectorize.c
>
> // Vectorization will occur with:
> // clang -O2 -S -emit-llvm -fno-math-errno test_vectorize.c
>
> // Vectorization will not occur with:
> // clang -O2 -S -emit-llvm -fno-math-errno -ffast-math test_vectorize.c
>
> #include <math.h>
>
> void test_vectorize_exp(long n,  float* restrict y,  float* restrict x) {
>    long i;
>    for (i = 0; i < n; i++) {
>      x[i] = expf(y[i]);
>    }
> }
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
-- 
Hal Finkel
Lead, Compiler Technology and Programming Languages
Leadership Computing Facility
Argonne National Laboratory

Thanks Hal,
I'll take a look into extending the TargetLibraryInfo.
Chris
 

    On Monday, March 20, 2017 1:20 PM, Hal Finkel <hfinkel at anl.gov>
wrote:
 

 
On 03/20/2017 11:27 AM, Chris Chrulski via llvm-dev
wrote:
> Hi,
>
> I came across an issue where some optimizations that would normally be
applied to standard math function calls are not getting applied when the
–ffast-math option is enabled on the Clang command line on a Linux x86_64
target.
>
> I tracked down the issue to occurring because the –ffast-math option is
triggering Clang to preprocess the math.h header file with the
__FINITE_MATH_ONLY__ macro set to 1. In this case, the Linux header files are
redirecting several math functions using the __asm__ extension to rename the
symbol. An example of what the header files contains is:
>        extern double exp (double) __asm__ (""
"__exp_finite") __attribute__ ((__nothrow__ ));
>
> In effect, ‘exp’ gets converted to be ‘__exp_finite’ in the IR. Because
some of the optimizations like constant folding or vectorization are looking for
the original function name, those optimizations do not get triggered under
–ffast-math due to these alternative function names.
>
> Other affected functions include: acos, asin, exp2, log, log2, log10, and a
few others.
>
> Anybody know if this is intentional, or have input regarding whether these
optimizations should be extended to also look for the renamed versions of these
functions?
It is not intentional; the optimizations should be extended. I imagine 
that we would extend TargetLibraryInfo to have names for these 
functions, and then we could recognize them when available, and generate 
them when available and the inputs have the 'fast' tag.

  -Hal
>
> Thanks,
> Chris
>
>
>
> A couple of example cases:
>
> // File: test_folding.c
>
> // Constant fold will occur with:
> // clang -O2 -S -emit-llvm test_folding.c
>
> // Constant fold will not occur with:
> // clang -O2 -S -emit-llvm -ffast-math test_folding.c
>
> #include <math.h>
>
> double test_fold_exp() {
>    return exp(0.0);
> }
>
> double test_fold_acos() {
>    return acos(1.0);
> }
>
> ----------------------------------------------
>
> // File: test_vectorize.c
>
> // Vectorization will occur with:
> // clang -O2 -S -emit-llvm -fno-math-errno test_vectorize.c
>
> // Vectorization will not occur with:
> // clang -O2 -S -emit-llvm -fno-math-errno -ffast-math test_vectorize.c
>
> #include <math.h>
>
> void test_vectorize_exp(long n,  float* restrict y,  float* restrict x) {
>    long i;
>    for (i = 0; i < n; i++) {
>      x[i] = expf(y[i]);
>    }
> }
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
-- 
Hal Finkel
Lead, Compiler Technology and Programming Languages
Leadership Computing Facility
Argonne National Laboratory



   
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20170320/03435350/attachment.html>