Getting error on make -j4 gpt-2

[praveenkumar91]

Consolidate compiler generated dependencies of target ggml
[ 33%] Built target ggml_utils
[ 50%] Building C object src/CMakeFiles/ggml.dir/ggml.c.o
In file included from /usr/lib/gcc/x86_64-redhat-linux/8/include/immintrin.h:99,
from /home/kumapra/MACHINE/ggml/src/ggml.c:153:
/home/kumapra/MACHINE/ggml/src/ggml.c: In function ‘ggml_vec_dot_f16’:
/usr/lib/gcc/x86_64-redhat-linux/8/include/f16cintrin.h:52:1: error: inlining failed in call to always_inline ‘_mm256_cvtph_ps’: target specific option mismatch
_mm256_cvtph_ps (__m128i __A)
^~~~~~~~~~~~~~~
/home/kumapra/MACHINE/ggml/src/ggml.c:543:33: note: called from here
#define GGML_F32Cx8_LOAD(x) _mm256_cvtph_ps(_mm_loadu_si128((__m128i )(x)))
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/home/kumapra/MACHINE/ggml/src/ggml.c:553:37: note: in expansion of macro ‘GGML_F32Cx8_LOAD’
#define GGML_F16_VEC_LOAD(p, i) GGML_F32Cx8_LOAD(p)
^~~~~~~~~~~~~~~~
/home/kumapra/MACHINE/ggml/src/ggml.c:906:21: note: in expansion of macro ‘GGML_F16_VEC_LOAD’
ay[j] = GGML_F16_VEC_LOAD(y + i + jGGML_F16_EPR, j);
^~~~~~~~~~~~~~~~~
In file included from /usr/lib/gcc/x86_64-redhat-linux/8/include/immintrin.h:99,
from /home/kumapra/MACHINE/ggml/src/ggml.c:153:
/usr/lib/gcc/x86_64-redhat-linux/8/include/f16cintrin.h:52:1: error: inlining failed in call to always_inline ‘_mm256_cvtph_ps’: target specific option mismatch
_mm256_cvtph_ps (__m128i __A)
^~~~~~~~~~~~~~~
/home/kumapra/MACHINE/ggml/src/ggml.c:543:33: note: called from here
#define GGML_F32Cx8_LOAD(x) _mm256_cvtph_ps(_mm_loadu_si128((__m128i )(x)))
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/home/kumapra/MACHINE/ggml/src/ggml.c:553:37: note: in expansion of macro ‘GGML_F32Cx8_LOAD’
#define GGML_F16_VEC_LOAD(p, i) GGML_F32Cx8_LOAD(p)
^~~~~~~~~~~~~~~~
/home/kumapra/MACHINE/ggml/src/ggml.c:905:21: note: in expansion of macro ‘GGML_F16_VEC_LOAD’
ax[j] = GGML_F16_VEC_LOAD(x + i + jGGML_F16_EPR, j);
^~~~~~~~~~~~~~~~~
In file included from /usr/lib/gcc/x86_64-redhat-linux/8/include/immintrin.h:99,
from /home/kumapra/MACHINE/ggml/src/ggml.c:153:
/usr/lib/gcc/x86_64-redhat-linux/8/include/f16cintrin.h:52:1: error: inlining failed in call to always_inline ‘_mm256_cvtph_ps’: target specific option mismatch
_mm256_cvtph_ps (__m128i __A)
^~~~~~~~~~~~~~~
/home/kumapra/MACHINE/ggml/src/ggml.c:543:33: note: called from here
#define GGML_F32Cx8_LOAD(x) _mm256_cvtph_ps(_mm_loadu_si128((__m128i )(x)))
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/home/kumapra/MACHINE/ggml/src/ggml.c:553:37: note: in expansion of macro ‘GGML_F32Cx8_LOAD’
#define GGML_F16_VEC_LOAD(p, i) GGML_F32Cx8_LOAD(p)
^~~~~~~~~~~~~~~~
/home/kumapra/MACHINE/ggml/src/ggml.c:905:21: note: in expansion of macro ‘GGML_F16_VEC_LOAD’
ax[j] = GGML_F16_VEC_LOAD(x + i + jGGML_F16_EPR, j);
^~~~~~~~~~~~~~~~~
In file included from /usr/lib/gcc/x86_64-redhat-linux/8/include/immintrin.h:99,
from /home/kumapra/MACHINE/ggml/src/ggml.c:153:
/usr/lib/gcc/x86_64-redhat-linux/8/include/f16cintrin.h:52:1: error: inlining failed in call to always_inline ‘_mm256_cvtph_ps’: target specific option mismatch
_mm256_cvtph_ps (__m128i __A)
^~~~~~~~~~~~~~~
/home/kumapra/MACHINE/ggml/src/ggml.c:543:33: note: called from here
#define GGML_F32Cx8_LOAD(x) _mm256_cvtph_ps(_mm_loadu_si128((__m128i )(x)))
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/home/kumapra/MACHINE/ggml/src/ggml.c:553:37: note: in expansion of macro ‘GGML_F32Cx8_LOAD’
#define GGML_F16_VEC_LOAD(p, i) GGML_F32Cx8_LOAD(p)
^~~~~~~~~~~~~~~~
/home/kumapra/MACHINE/ggml/src/ggml.c:906:21: note: in expansion of macro ‘GGML_F16_VEC_LOAD’
ay[j] = GGML_F16_VEC_LOAD(y + i + jGGML_F16_EPR, j);
^~~~~~~~~~~~~~~~~
In file included from /usr/lib/gcc/x86_64-redhat-linux/8/include/immintrin.h:99,
from /home/kumapra/MACHINE/ggml/src/ggml.c:153:
/usr/lib/gcc/x86_64-redhat-linux/8/include/f16cintrin.h:52:1: error: inlining failed in call to always_inline ‘_mm256_cvtph_ps’: target specific option mismatch
_mm256_cvtph_ps (__m128i __A)
^~~~~~~~~~~~~~~
/home/kumapra/MACHINE/ggml/src/ggml.c:543:33: note: called from here
#define GGML_F32Cx8_LOAD(x) _mm256_cvtph_ps(_mm_loadu_si128((__m128i )(x)))
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/home/kumapra/MACHINE/ggml/src/ggml.c:553:37: note: in expansion of macro ‘GGML_F32Cx8_LOAD’
#define GGML_F16_VEC_LOAD(p, i) GGML_F32Cx8_LOAD(p)
^~~~~~~~~~~~~~~~
/home/kumapra/MACHINE/ggml/src/ggml.c:905:21: note: in expansion of macro ‘GGML_F16_VEC_LOAD’
ax[j] = GGML_F16_VEC_LOAD(x + i + jGGML_F16_EPR, j);
^~~~~~~~~~~~~~~~~
In file included from /usr/lib/gcc/x86_64-redhat-linux/8/include/immintrin.h:99,
from /home/kumapra/MACHINE/ggml/src/ggml.c:153:
/usr/lib/gcc/x86_64-redhat-linux/8/include/f16cintrin.h:52:1: error: inlining failed in call to always_inline ‘_mm256_cvtph_ps’: target specific option mismatch
_mm256_cvtph_ps (__m128i __A)
^~~~~~~~~~~~~~~
/home/kumapra/MACHINE/ggml/src/ggml.c:543:33: note: called from here
#define GGML_F32Cx8_LOAD(x) _mm256_cvtph_ps(_mm_loadu_si128((__m128i )(x)))
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/home/kumapra/MACHINE/ggml/src/ggml.c:553:37: note: in expansion of macro ‘GGML_F32Cx8_LOAD’
#define GGML_F16_VEC_LOAD(p, i) GGML_F32Cx8_LOAD(p)
^~~~~~~~~~~~~~~~
/home/kumapra/MACHINE/ggml/src/ggml.c:906:21: note: in expansion of macro ‘GGML_F16_VEC_LOAD’
ay[j] = GGML_F16_VEC_LOAD(y + i + jGGML_F16_EPR, j);
^~~~~~~~~~~~~~~~~
make[3]: *** [src/CMakeFiles/ggml.dir/build.make:76: src/CMakeFiles/ggml.dir/ggml.c.o] Error 1
make[2]: *** [CMakeFiles/Makefile2:205: src/CMakeFiles/ggml.dir/all] Error 2
make[1]: *** [CMakeFiles/Makefile2:499: examples/gpt-2/CMakeFiles/gpt-2.dir/rule] Error 2
make: *** [Makefile:322: gpt-2] Error 2

[Computeiful]

I was able to overcome this issue by undefine-ing the AVX2 macro in ggml.c, specifically above the definition for the vec_dot functions, for example something like:

// Undefine the __AVX2__ macro so the function uses __AVX__ instead.
#undef __AVX2__

static void ggml_vec_dot_q4_0_q8_0(const int n, float * restrict s, const void * restrict vx, const void * restrict vy);
static void ggml_vec_dot_q4_1_q8_1(const int n, float * restrict s, const void * restrict vx, const void * restrict vy);
static void ggml_vec_dot_q5_0_q8_0(const int n, float * restrict s, const void * restrict vx, const void * restrict vy);
static void ggml_vec_dot_q5_1_q8_1(const int n, float * restrict s, const void * restrict vx, const void * restrict vy);
static void ggml_vec_dot_q8_0_q8_0(const int n, float * restrict s, const void * restrict vx, const void * restrict vy);