Message ID | 20230908041616.3574595-1-mingli.yu@windriver.com |
---|---|
State | New |
Headers | show |
Series | [mickledore,v2] fixup! python3-numpy: remove NPY_INLINE, use inline instead | expand |
when do we see these errors ? how does this impact performance? On Thu, Sep 7, 2023 at 9:16 PM Yu, Mingli <mingli.yu@eng.windriver.com> wrote: > > From: Mingli Yu <mingli.yu@windriver.com> > > Fixes: > /usr/lib/gcc/x86_64-redhat-linux/13/include/avx512fintrin.h:314:1: error: inlining failed in call to 'always_inline' '_mm512_setzero_ps': target specific option mismatch > 314 | _mm512_setzero_ps (void) > | ^~~~~~~~~~~~~~~~~ > numpy/core/src/umath/simd.inc.src:977:20: note: called from here > 977 | @vtype@ zeros = _mm512_setzero_@vsuffix@(); > | ^~~~~~~~~~~~~~~~~~~ > numpy/core/src/umath/simd.inc.src:596:1: error: inlining failed in call to ‘always_inline’ ‘avx512_get_full_load_mask_ps’: target specific option mismatch > 596 | avx512_get_full_load_mask_ps(void) > | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~ > numpy/core/src/umath/simd.inc.src:976:27: note: called from here > 976 | @mask@ load_mask = avx512_get_full_load_mask_@vsuffix@(); > | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > > Signed-off-by: Mingli Yu <mingli.yu@windriver.com> > --- > .../0001-simd.inc.src-Change-NPY_INLINE-to-inline.patch | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/meta/recipes-devtools/python/python3-numpy/0001-simd.inc.src-Change-NPY_INLINE-to-inline.patch b/meta/recipes-devtools/python/python3-numpy/0001-simd.inc.src-Change-NPY_INLINE-to-inline.patch > index d733dda333..73c9c79bdd 100644 > --- a/meta/recipes-devtools/python/python3-numpy/0001-simd.inc.src-Change-NPY_INLINE-to-inline.patch > +++ b/meta/recipes-devtools/python/python3-numpy/0001-simd.inc.src-Change-NPY_INLINE-to-inline.patch > @@ -117,7 +117,7 @@ index d6c9a7e..39aec9a 100644 > > #if defined HAVE_ATTRIBUTE_TARGET_AVX512F_WITH_INTRINSICS && defined NPY_HAVE_SSE2_INTRINSICS > -static NPY_GCC_OPT_3 NPY_INLINE NPY_GCC_TARGET_AVX512F void > -+static NPY_GCC_OPT_3 inline NPY_GCC_TARGET_AVX512F void > ++static inline NPY_GCC_TARGET_AVX512F void > AVX512F_@func@_@TYPE@(@type@ * op, > @type@ * ip, > const npy_intp array_size, > @@ -126,7 +126,7 @@ index d6c9a7e..39aec9a 100644 > > #if defined HAVE_ATTRIBUTE_TARGET_AVX512F_WITH_INTRINSICS && defined NPY_HAVE_SSE2_INTRINSICS > -static NPY_GCC_OPT_3 NPY_INLINE NPY_GCC_TARGET_AVX512F void > -+static NPY_GCC_OPT_3 inline NPY_GCC_TARGET_AVX512F void > ++static inline NPY_GCC_TARGET_AVX512F void > AVX512F_absolute_@TYPE@(@type@ * op, > @type@ * ip, > const npy_intp array_size, > -- > 2.25.1 > > > -=-=-=-=-=-=-=-=-=-=-=- > Links: You receive all messages sent to this group. > View/Reply Online (#187402): https://lists.openembedded.org/g/openembedded-core/message/187402 > Mute This Topic: https://lists.openembedded.org/mt/101231101/1997914 > Group Owner: openembedded-core+owner@lists.openembedded.org > Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [raj.khem@gmail.com] > -=-=-=-=-=-=-=-=-=-=-=- >
We will encounter the build failure when debug enabled(add DEBUG_BUILD = "1" in local.conf) [1]. [1] https://git.openembedded.org/openembedded-core/commit/?h=mickledore&id=77a64a8686b6c9ef3bc6adbce6cdc442096decfd Thanks, ________________________________ From: Khem Raj <raj.khem@gmail.com> Sent: Friday, September 8, 2023 12:46 To: Yu, Mingli <Mingli.Yu@windriver.com> Cc: openembedded-core@lists.openembedded.org <openembedded-core@lists.openembedded.org> Subject: Re: [OE-core] [mickledore][PATCH v2] fixup! python3-numpy: remove NPY_INLINE, use inline instead CAUTION: This email comes from a non Wind River email account! Do not click links or open attachments unless you recognize the sender and know the content is safe. when do we see these errors ? how does this impact performance? On Thu, Sep 7, 2023 at 9:16 PM Yu, Mingli <mingli.yu@eng.windriver.com> wrote: > > From: Mingli Yu <mingli.yu@windriver.com> > > Fixes: > /usr/lib/gcc/x86_64-redhat-linux/13/include/avx512fintrin.h:314:1: error: inlining failed in call to 'always_inline' '_mm512_setzero_ps': target specific option mismatch > 314 | _mm512_setzero_ps (void) > | ^~~~~~~~~~~~~~~~~ > numpy/core/src/umath/simd.inc.src:977:20: note: called from here > 977 | @vtype@ zeros = _mm512_setzero_@vsuffix@(); > | ^~~~~~~~~~~~~~~~~~~ > numpy/core/src/umath/simd.inc.src:596:1: error: inlining failed in call to ‘always_inline’ ‘avx512_get_full_load_mask_ps’: target specific option mismatch > 596 | avx512_get_full_load_mask_ps(void) > | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~ > numpy/core/src/umath/simd.inc.src:976:27: note: called from here > 976 | @mask@ load_mask = avx512_get_full_load_mask_@vsuffix@(); > | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > > Signed-off-by: Mingli Yu <mingli.yu@windriver.com> > --- > .../0001-simd.inc.src-Change-NPY_INLINE-to-inline.patch | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/meta/recipes-devtools/python/python3-numpy/0001-simd.inc.src-Change-NPY_INLINE-to-inline.patch b/meta/recipes-devtools/python/python3-numpy/0001-simd.inc.src-Change-NPY_INLINE-to-inline.patch > index d733dda333..73c9c79bdd 100644 > --- a/meta/recipes-devtools/python/python3-numpy/0001-simd.inc.src-Change-NPY_INLINE-to-inline.patch > +++ b/meta/recipes-devtools/python/python3-numpy/0001-simd.inc.src-Change-NPY_INLINE-to-inline.patch > @@ -117,7 +117,7 @@ index d6c9a7e..39aec9a 100644 > > #if defined HAVE_ATTRIBUTE_TARGET_AVX512F_WITH_INTRINSICS && defined NPY_HAVE_SSE2_INTRINSICS > -static NPY_GCC_OPT_3 NPY_INLINE NPY_GCC_TARGET_AVX512F void > -+static NPY_GCC_OPT_3 inline NPY_GCC_TARGET_AVX512F void > ++static inline NPY_GCC_TARGET_AVX512F void > AVX512F_@func@_@TYPE@(@type@ * op, > @type@ * ip, > const npy_intp array_size, > @@ -126,7 +126,7 @@ index d6c9a7e..39aec9a 100644 > > #if defined HAVE_ATTRIBUTE_TARGET_AVX512F_WITH_INTRINSICS && defined NPY_HAVE_SSE2_INTRINSICS > -static NPY_GCC_OPT_3 NPY_INLINE NPY_GCC_TARGET_AVX512F void > -+static NPY_GCC_OPT_3 inline NPY_GCC_TARGET_AVX512F void > ++static inline NPY_GCC_TARGET_AVX512F void > AVX512F_absolute_@TYPE@(@type@ * op, > @type@ * ip, > const npy_intp array_size, > -- > 2.25.1 > > > -=-=-=-=-=-=-=-=-=-=-=- > Links: You receive all messages sent to this group. > View/Reply Online (#187402): https://lists.openembedded.org/g/openembedded-core/message/187402 > Mute This Topic: https://lists.openembedded.org/mt/101231101/1997914 > Group Owner: openembedded-core+owner@lists.openembedded.org > Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [raj.khem@gmail.com] > -=-=-=-=-=-=-=-=-=-=-=- >
On Thu, Sep 7, 2023 at 10:29 PM Yu, Mingli <Mingli.Yu@windriver.com> wrote: > > We will encounter the build failure when debug enabled(add DEBUG_BUILD = "1" in local.conf) [1]. > I would suggest to keep it limited to this option then > [1] https://git.openembedded.org/openembedded-core/commit/?h=mickledore&id=77a64a8686b6c9ef3bc6adbce6cdc442096decfd > > Thanks, > ________________________________ > From: Khem Raj <raj.khem@gmail.com> > Sent: Friday, September 8, 2023 12:46 > To: Yu, Mingli <Mingli.Yu@windriver.com> > Cc: openembedded-core@lists.openembedded.org <openembedded-core@lists.openembedded.org> > Subject: Re: [OE-core] [mickledore][PATCH v2] fixup! python3-numpy: remove NPY_INLINE, use inline instead > > CAUTION: This email comes from a non Wind River email account! > Do not click links or open attachments unless you recognize the sender and know the content is safe. > > when do we see these errors ? how does this impact performance? > > On Thu, Sep 7, 2023 at 9:16 PM Yu, Mingli <mingli.yu@eng.windriver.com> wrote: > > > > From: Mingli Yu <mingli.yu@windriver.com> > > > > Fixes: > > /usr/lib/gcc/x86_64-redhat-linux/13/include/avx512fintrin.h:314:1: error: inlining failed in call to 'always_inline' '_mm512_setzero_ps': target specific option mismatch > > 314 | _mm512_setzero_ps (void) > > | ^~~~~~~~~~~~~~~~~ > > numpy/core/src/umath/simd.inc.src:977:20: note: called from here > > 977 | @vtype@ zeros = _mm512_setzero_@vsuffix@(); > > | ^~~~~~~~~~~~~~~~~~~ > > numpy/core/src/umath/simd.inc.src:596:1: error: inlining failed in call to ‘always_inline’ ‘avx512_get_full_load_mask_ps’: target specific option mismatch > > 596 | avx512_get_full_load_mask_ps(void) > > | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~ > > numpy/core/src/umath/simd.inc.src:976:27: note: called from here > > 976 | @mask@ load_mask = avx512_get_full_load_mask_@vsuffix@(); > > | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > > > > Signed-off-by: Mingli Yu <mingli.yu@windriver.com> > > --- > > .../0001-simd.inc.src-Change-NPY_INLINE-to-inline.patch | 4 ++-- > > 1 file changed, 2 insertions(+), 2 deletions(-) > > > > diff --git a/meta/recipes-devtools/python/python3-numpy/0001-simd.inc.src-Change-NPY_INLINE-to-inline.patch b/meta/recipes-devtools/python/python3-numpy/0001-simd.inc.src-Change-NPY_INLINE-to-inline.patch > > index d733dda333..73c9c79bdd 100644 > > --- a/meta/recipes-devtools/python/python3-numpy/0001-simd.inc.src-Change-NPY_INLINE-to-inline.patch > > +++ b/meta/recipes-devtools/python/python3-numpy/0001-simd.inc.src-Change-NPY_INLINE-to-inline.patch > > @@ -117,7 +117,7 @@ index d6c9a7e..39aec9a 100644 > > > > #if defined HAVE_ATTRIBUTE_TARGET_AVX512F_WITH_INTRINSICS && defined NPY_HAVE_SSE2_INTRINSICS > > -static NPY_GCC_OPT_3 NPY_INLINE NPY_GCC_TARGET_AVX512F void > > -+static NPY_GCC_OPT_3 inline NPY_GCC_TARGET_AVX512F void > > ++static inline NPY_GCC_TARGET_AVX512F void > > AVX512F_@func@_@TYPE@(@type@ * op, > > @type@ * ip, > > const npy_intp array_size, > > @@ -126,7 +126,7 @@ index d6c9a7e..39aec9a 100644 > > > > #if defined HAVE_ATTRIBUTE_TARGET_AVX512F_WITH_INTRINSICS && defined NPY_HAVE_SSE2_INTRINSICS > > -static NPY_GCC_OPT_3 NPY_INLINE NPY_GCC_TARGET_AVX512F void > > -+static NPY_GCC_OPT_3 inline NPY_GCC_TARGET_AVX512F void > > ++static inline NPY_GCC_TARGET_AVX512F void > > AVX512F_absolute_@TYPE@(@type@ * op, > > @type@ * ip, > > const npy_intp array_size, > > -- > > 2.25.1 > > > > > > -=-=-=-=-=-=-=-=-=-=-=- > > Links: You receive all messages sent to this group. > > View/Reply Online (#187402): https://lists.openembedded.org/g/openembedded-core/message/187402 > > Mute This Topic: https://lists.openembedded.org/mt/101231101/1997914 > > Group Owner: openembedded-core+owner@lists.openembedded.org > > Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [raj.khem@gmail.com] > > -=-=-=-=-=-=-=-=-=-=-=- > >
diff --git a/meta/recipes-devtools/python/python3-numpy/0001-simd.inc.src-Change-NPY_INLINE-to-inline.patch b/meta/recipes-devtools/python/python3-numpy/0001-simd.inc.src-Change-NPY_INLINE-to-inline.patch index d733dda333..73c9c79bdd 100644 --- a/meta/recipes-devtools/python/python3-numpy/0001-simd.inc.src-Change-NPY_INLINE-to-inline.patch +++ b/meta/recipes-devtools/python/python3-numpy/0001-simd.inc.src-Change-NPY_INLINE-to-inline.patch @@ -117,7 +117,7 @@ index d6c9a7e..39aec9a 100644 #if defined HAVE_ATTRIBUTE_TARGET_AVX512F_WITH_INTRINSICS && defined NPY_HAVE_SSE2_INTRINSICS -static NPY_GCC_OPT_3 NPY_INLINE NPY_GCC_TARGET_AVX512F void -+static NPY_GCC_OPT_3 inline NPY_GCC_TARGET_AVX512F void ++static inline NPY_GCC_TARGET_AVX512F void AVX512F_@func@_@TYPE@(@type@ * op, @type@ * ip, const npy_intp array_size, @@ -126,7 +126,7 @@ index d6c9a7e..39aec9a 100644 #if defined HAVE_ATTRIBUTE_TARGET_AVX512F_WITH_INTRINSICS && defined NPY_HAVE_SSE2_INTRINSICS -static NPY_GCC_OPT_3 NPY_INLINE NPY_GCC_TARGET_AVX512F void -+static NPY_GCC_OPT_3 inline NPY_GCC_TARGET_AVX512F void ++static inline NPY_GCC_TARGET_AVX512F void AVX512F_absolute_@TYPE@(@type@ * op, @type@ * ip, const npy_intp array_size,