+2017-07-06 Sebastian Peryt <sebastian.peryt@intel.com>
+
+ * config/i386/avx512fintrin.h (_mm_mask_getexp_round_ss,
+ _mm_maskz_getexp_round_ss, _mm_mask_getexp_round_sd,
+ _mm_maskz_getexp_round_sd, _mm_mask_getmant_round_sd,
+ _mm_maskz_getmant_round_sd, _mm_mask_getmant_round_ss,
+ _mm_maskz_getmant_round_ss, _mm_mask_getexp_ss, _mm_maskz_getexp_ss,
+ _mm_mask_getexp_sd, _mm_maskz_getexp_sd, _mm_mask_getmant_sd,
+ _mm_maskz_getmant_sd, _mm_mask_getmant_ss,
+ _mm_maskz_getmant_ss): New intrinsics.
+ (__builtin_ia32_getexpss128_mask): Changed to ...
+ __builtin_ia32_getexpss128_round ... this.
+ (__builtin_ia32_getexpsd128_mask): Changed to ...
+ __builtin_ia32_getexpsd128_round ... this.
+ * config/i386/i386-builtin-types.def
+ ((V2DF, V2DF, V2DF, INT, V2DF, UQI, INT),
+ (V4SF, V4SF, V4SF, INT, V4SF, UQI, INT)): New function type aliases.
+ * config/i386/i386-builtin.def (__builtin_ia32_getexpsd_mask_round,
+ __builtin_ia32_getexpss_mask_round, __builtin_ia32_getmantsd_mask_round,
+ __builtin_ia32_getmantss_mask_round): New builtins.
+ * config/i386/i386.c (V2DF_FTYPE_V2DF_V2DF_INT_V2DF_UQI_INT,
+ V4SF_FTYPE_V4SF_V4SF_INT_V4SF_UQI_INT): Handle new types.
+ (CODE_FOR_avx512f_vgetmantv2df_mask_round,
+ CODE_FOR_avx512f_vgetmantv4sf_mask_round): New cases.
+ * config/i386/sse.md
+ (avx512f_sgetexp<mode><round_saeonly_name>): Changed to ...
+ avx512f_sgetexp<mode><mask_scalar_name>
+ <round_saeonly_scalar_name> ... this.
+ (vgetexp<ssescalarmodesuffix>\t{<round_saeonly_op3>%2, %1, %0|
+ %0, %1, %2<round_saeonly_op3>}): Changed to ...
+ vgetexp<ssescalarmodesuffix>
+ \t{<round_saeonly_scalar_mask_op3>%2, %1, %0<mask_scalar_operand3>|
+ %0<mask_scalar_operand3>, %1, %2<round_saeonly_scalar_mask_op3>} ... this.
+ (avx512f_vgetmant<mode><round_saeonly_name>): Changed to ...
+ avx512f_vgetmant<mode><mask_scalar_name>
+ <round_saeonly_scalar_name> ... this.
+ (vgetmant<ssescalarmodesuffix>\t{%3, <round_saeonly_op4>%2, %1, %0|
+ %0, %1, %2<round_saeonly_op4>, %3}): Changed to ...
+ vgetmant<ssescalarmodesuffix>
+ \t{%3, <round_saeonly_scalar_mask_op4>%2, %1, %0<mask_scalar_operand4>|
+ %0<mask_scalar_operand4>, %1, %2
+ <round_saeonly_scalar_mask_op4>, %3} ... this.
+ * config/i386/subst.md (mask_scalar_operand4,
+ round_saeonly_scalar_mask_operand4, round_saeonly_scalar_mask_op4,
+ round_saeonly_scalar_nimm_predicate): New subst attributes.
+
2017-07-06 Julia Koval <julia.koval@intel.com>
* gcc/config/i386/i386.c (ix86_erase_embedded_rounding):
__R);
}
+extern __inline __m128
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_mask_getexp_round_ss (__m128 __W, __mmask8 __U, __m128 __A,
+ __m128 __B, const int __R)
+{
+ return (__m128) __builtin_ia32_getexpss_mask_round ((__v4sf) __A,
+ (__v4sf) __B,
+ (__v4sf) __W,
+ (__mmask8) __U, __R);
+}
+
+extern __inline __m128
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_maskz_getexp_round_ss (__mmask8 __U, __m128 __A, __m128 __B,
+ const int __R)
+{
+ return (__m128) __builtin_ia32_getexpss_mask_round ((__v4sf) __A,
+ (__v4sf) __B,
+ (__v4sf)
+ _mm_setzero_ps (),
+ (__mmask8) __U, __R);
+}
+
extern __inline __m128d
__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
_mm_getexp_round_sd (__m128d __A, __m128d __B, const int __R)
__R);
}
+extern __inline __m128d
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_mask_getexp_round_sd (__m128d __W, __mmask8 __U, __m128d __A,
+ __m128d __B, const int __R)
+{
+ return (__m128d) __builtin_ia32_getexpsd_mask_round ((__v2df) __A,
+ (__v2df) __B,
+ (__v2df) __W,
+ (__mmask8) __U, __R);
+}
+
+extern __inline __m128d
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_maskz_getexp_round_sd (__mmask8 __U, __m128d __A, __m128d __B,
+ const int __R)
+{
+ return (__m128d) __builtin_ia32_getexpsd_mask_round ((__v2df) __A,
+ (__v2df) __B,
+ (__v2df)
+ _mm_setzero_pd (),
+ (__mmask8) __U, __R);
+}
+
extern __inline __m512
__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
_mm512_getexp_round_ps (__m512 __A, const int __R)
__R);
}
+extern __inline __m128d
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_mask_getmant_round_sd (__m128d __W, __mmask8 __U, __m128d __A,
+ __m128d __B, _MM_MANTISSA_NORM_ENUM __C,
+ _MM_MANTISSA_SIGN_ENUM __D, const int __R)
+{
+ return (__m128d) __builtin_ia32_getmantsd_mask_round ((__v2df) __A,
+ (__v2df) __B,
+ (__D << 2) | __C,
+ (__v2df) __W,
+ __U, __R);
+}
+
+extern __inline __m128d
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_maskz_getmant_round_sd (__mmask8 __U, __m128d __A, __m128d __B,
+ _MM_MANTISSA_NORM_ENUM __C,
+ _MM_MANTISSA_SIGN_ENUM __D, const int __R)
+{
+ return (__m128d) __builtin_ia32_getmantsd_mask_round ((__v2df) __A,
+ (__v2df) __B,
+ (__D << 2) | __C,
+ (__v2df)
+ _mm_setzero_pd(),
+ __U, __R);
+}
+
extern __inline __m128
__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
_mm_getmant_round_ss (__m128 __A, __m128 __B,
__R);
}
+extern __inline __m128
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_mask_getmant_round_ss (__m128 __W, __mmask8 __U, __m128 __A,
+ __m128 __B, _MM_MANTISSA_NORM_ENUM __C,
+ _MM_MANTISSA_SIGN_ENUM __D, const int __R)
+{
+ return (__m128) __builtin_ia32_getmantss_mask_round ((__v4sf) __A,
+ (__v4sf) __B,
+ (__D << 2) | __C,
+ (__v4sf) __W,
+ __U, __R);
+}
+
+extern __inline __m128
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_maskz_getmant_round_ss (__mmask8 __U, __m128 __A, __m128 __B,
+ _MM_MANTISSA_NORM_ENUM __C,
+ _MM_MANTISSA_SIGN_ENUM __D, const int __R)
+{
+ return (__m128) __builtin_ia32_getmantss_mask_round ((__v4sf) __A,
+ (__v4sf) __B,
+ (__D << 2) | __C,
+ (__v4sf)
+ _mm_setzero_ps(),
+ __U, __R);
+}
+
#else
#define _mm512_getmant_round_pd(X, B, C, R) \
((__m512d)__builtin_ia32_getmantpd512_mask ((__v8df)(__m512d)(X), \
(int)(((D)<<2) | (C)), \
(R)))
+#define _mm_mask_getmant_round_sd(W, U, X, Y, C, D, R) \
+ ((__m128d)__builtin_ia32_getmantsd_mask_round ((__v2df)(__m128d)(X), \
+ (__v2df)(__m128d)(Y), \
+ (int)(((D)<<2) | (C)), \
+ (__v2df)(__m128d)(W), \
+ (__mmask8)(U),\
+ (R)))
+
+#define _mm_maskz_getmant_round_sd(U, X, Y, C, D, R) \
+ ((__m128d)__builtin_ia32_getmantsd_mask_round ((__v2df)(__m128d)(X), \
+ (__v2df)(__m128d)(Y), \
+ (int)(((D)<<2) | (C)), \
+ (__v2df)(__m128d)_mm_setzero_pd(), \
+ (__mmask8)(U),\
+ (R)))
+
#define _mm_getmant_round_ss(X, Y, C, D, R) \
((__m128)__builtin_ia32_getmantss_round ((__v4sf)(__m128)(X), \
(__v4sf)(__m128)(Y), \
(int)(((D)<<2) | (C)), \
(R)))
+#define _mm_mask_getmant_round_ss(W, U, X, Y, C, D, R) \
+ ((__m128)__builtin_ia32_getmantss_mask_round ((__v4sf)(__m128)(X), \
+ (__v4sf)(__m128)(Y), \
+ (int)(((D)<<2) | (C)), \
+ (__v4sf)(__m128)(W), \
+ (__mmask8)(U),\
+ (R)))
+
+#define _mm_maskz_getmant_round_ss(U, X, Y, C, D, R) \
+ ((__m128)__builtin_ia32_getmantss_mask_round ((__v4sf)(__m128)(X), \
+ (__v4sf)(__m128)(Y), \
+ (int)(((D)<<2) | (C)), \
+ (__v4sf)(__m128)_mm_setzero_ps(), \
+ (__mmask8)(U),\
+ (R)))
+
#define _mm_getexp_round_ss(A, B, R) \
((__m128)__builtin_ia32_getexpss128_round((__v4sf)(__m128)(A), (__v4sf)(__m128)(B), R))
+#define _mm_mask_getexp_round_ss(W, U, A, B, C) \
+ (__m128)__builtin_ia32_getexpss_mask_round(A, B, W, U, C)
+
+#define _mm_maskz_getexp_round_ss(U, A, B, C) \
+ (__m128)__builtin_ia32_getexpss_mask_round(A, B, (__v4sf)_mm_setzero_ps(), U, C)
+
#define _mm_getexp_round_sd(A, B, R) \
((__m128d)__builtin_ia32_getexpsd128_round((__v2df)(__m128d)(A), (__v2df)(__m128d)(B), R))
+#define _mm_mask_getexp_round_sd(W, U, A, B, C) \
+ (__m128d)__builtin_ia32_getexpsd_mask_round(A, B, W, U, C)
+
+#define _mm_maskz_getexp_round_sd(U, A, B, C) \
+ (__m128d)__builtin_ia32_getexpsd_mask_round(A, B, (__v2df)_mm_setzero_pd(), U, C)
+
+
#define _mm512_getexp_round_ps(A, R) \
((__m512)__builtin_ia32_getexpps512_mask((__v16sf)(__m512)(A), \
(__v16sf)_mm512_undefined_ps(), (__mmask16)-1, R))
_MM_FROUND_CUR_DIRECTION);
}
+extern __inline __m128
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_mask_getexp_ss (__m128 __W, __mmask8 __U, __m128 __A, __m128 __B)
+{
+ return (__m128) __builtin_ia32_getexpss_mask_round ((__v4sf) __A,
+ (__v4sf) __B,
+ (__v4sf) __W,
+ (__mmask8) __U,
+ _MM_FROUND_CUR_DIRECTION);
+}
+
+extern __inline __m128
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_maskz_getexp_ss (__mmask8 __U, __m128 __A, __m128 __B)
+{
+ return (__m128) __builtin_ia32_getexpss_mask_round ((__v4sf) __A,
+ (__v4sf) __B,
+ (__v4sf)
+ _mm_setzero_ps (),
+ (__mmask8) __U,
+ _MM_FROUND_CUR_DIRECTION);
+}
+
extern __inline __m128d
__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
_mm_getexp_sd (__m128d __A, __m128d __B)
_MM_FROUND_CUR_DIRECTION);
}
+extern __inline __m128d
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_mask_getexp_sd (__m128d __W, __mmask8 __U, __m128d __A, __m128d __B)
+{
+ return (__m128d) __builtin_ia32_getexpsd_mask_round ((__v2df) __A,
+ (__v2df) __B,
+ (__v2df) __W,
+ (__mmask8) __U,
+ _MM_FROUND_CUR_DIRECTION);
+}
+
+extern __inline __m128d
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_maskz_getexp_sd (__mmask8 __U, __m128d __A, __m128d __B)
+{
+ return (__m128d) __builtin_ia32_getexpsd_mask_round ((__v2df) __A,
+ (__v2df) __B,
+ (__v2df)
+ _mm_setzero_pd (),
+ (__mmask8) __U,
+ _MM_FROUND_CUR_DIRECTION);
+}
+
extern __inline __m512d
__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
_mm512_getmant_pd (__m512d __A, _MM_MANTISSA_NORM_ENUM __B,
_MM_FROUND_CUR_DIRECTION);
}
+extern __inline __m128d
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_mask_getmant_sd (__m128d __W, __mmask8 __U, __m128d __A, __m128d __B,
+ _MM_MANTISSA_NORM_ENUM __C, _MM_MANTISSA_SIGN_ENUM __D)
+{
+ return (__m128d) __builtin_ia32_getmantsd_mask_round ((__v2df) __A,
+ (__v2df) __B,
+ (__D << 2) | __C,
+ (__v2df) __W,
+ __U,
+ _MM_FROUND_CUR_DIRECTION);
+}
+
+extern __inline __m128d
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_maskz_getmant_sd (__mmask8 __U, __m128d __A, __m128d __B,
+ _MM_MANTISSA_NORM_ENUM __C, _MM_MANTISSA_SIGN_ENUM __D)
+{
+ return (__m128d) __builtin_ia32_getmantsd_mask_round ((__v2df) __A,
+ (__v2df) __B,
+ (__D << 2) | __C,
+ (__v2df)
+ _mm_setzero_pd(),
+ __U,
+ _MM_FROUND_CUR_DIRECTION);
+}
+
extern __inline __m128
__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
_mm_getmant_ss (__m128 __A, __m128 __B, _MM_MANTISSA_NORM_ENUM __C,
_MM_FROUND_CUR_DIRECTION);
}
+extern __inline __m128
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_mask_getmant_ss (__m128 __W, __mmask8 __U, __m128 __A, __m128 __B,
+ _MM_MANTISSA_NORM_ENUM __C, _MM_MANTISSA_SIGN_ENUM __D)
+{
+ return (__m128) __builtin_ia32_getmantss_mask_round ((__v4sf) __A,
+ (__v4sf) __B,
+ (__D << 2) | __C,
+ (__v4sf) __W,
+ __U,
+ _MM_FROUND_CUR_DIRECTION);
+}
+
+extern __inline __m128
+__attribute__ ((__gnu_inline__, __always_inline__, __artificial__))
+_mm_maskz_getmant_ss (__mmask8 __U, __m128 __A, __m128 __B,
+ _MM_MANTISSA_NORM_ENUM __C, _MM_MANTISSA_SIGN_ENUM __D)
+{
+ return (__m128) __builtin_ia32_getmantss_mask_round ((__v4sf) __A,
+ (__v4sf) __B,
+ (__D << 2) | __C,
+ (__v4sf)
+ _mm_setzero_ps(),
+ __U,
+ _MM_FROUND_CUR_DIRECTION);
+}
+
#else
#define _mm512_getmant_pd(X, B, C) \
((__m512d)__builtin_ia32_getmantpd512_mask ((__v8df)(__m512d)(X), \
(int)(((D)<<2) | (C)), \
_MM_FROUND_CUR_DIRECTION))
+#define _mm_mask_getmant_sd(W, U, X, Y, C, D) \
+ ((__m128d)__builtin_ia32_getmantsd_mask_round ((__v2df)(__m128d)(X), \
+ (__v2df)(__m128d)(Y), \
+ (int)(((D)<<2) | (C)), \
+ (__v2df)(__m128d)(W), \
+ (__mmask8)(U),\
+ _MM_FROUND_CUR_DIRECTION))
+
+#define _mm_maskz_getmant_sd(U, X, Y, C, D) \
+ ((__m128d)__builtin_ia32_getmantsd_mask_round ((__v2df)(__m128d)(X), \
+ (__v2df)(__m128d)(Y), \
+ (int)(((D)<<2) | (C)), \
+ (__v2df)_mm_setzero_pd(), \
+ (__mmask8)(U),\
+ _MM_FROUND_CUR_DIRECTION))
+
#define _mm_getmant_ss(X, Y, C, D) \
((__m128)__builtin_ia32_getmantss_round ((__v4sf)(__m128)(X), \
(__v4sf)(__m128)(Y), \
(int)(((D)<<2) | (C)), \
_MM_FROUND_CUR_DIRECTION))
+#define _mm_mask_getmant_ss(W, U, X, Y, C, D) \
+ ((__m128)__builtin_ia32_getmantss_mask_round ((__v4sf)(__m128)(X), \
+ (__v4sf)(__m128)(Y), \
+ (int)(((D)<<2) | (C)), \
+ (__v4sf)(__m128)(W), \
+ (__mmask8)(U),\
+ _MM_FROUND_CUR_DIRECTION))
+
+#define _mm_maskz_getmant_ss(U, X, Y, C, D) \
+ ((__m128)__builtin_ia32_getmantss_mask_round ((__v4sf)(__m128)(X), \
+ (__v4sf)(__m128)(Y), \
+ (int)(((D)<<2) | (C)), \
+ (__v4sf)_mm_setzero_ps(), \
+ (__mmask8)(U),\
+ _MM_FROUND_CUR_DIRECTION))
+
#define _mm_getexp_ss(A, B) \
- ((__m128)__builtin_ia32_getexpss128_mask((__v4sf)(__m128)(A), (__v4sf)(__m128)(B), \
+ ((__m128)__builtin_ia32_getexpss128_round((__v4sf)(__m128)(A), (__v4sf)(__m128)(B), \
_MM_FROUND_CUR_DIRECTION))
+#define _mm_mask_getexp_ss(W, U, A, B) \
+ (__m128)__builtin_ia32_getexpss_mask_round(A, B, W, U,\
+ _MM_FROUND_CUR_DIRECTION)
+
+#define _mm_maskz_getexp_ss(U, A, B) \
+ (__m128)__builtin_ia32_getexpss_mask_round(A, B, (__v4sf)_mm_setzero_ps(), U,\
+ _MM_FROUND_CUR_DIRECTION)
+
#define _mm_getexp_sd(A, B) \
- ((__m128d)__builtin_ia32_getexpsd128_mask((__v2df)(__m128d)(A), (__v2df)(__m128d)(B),\
+ ((__m128d)__builtin_ia32_getexpsd128_round((__v2df)(__m128d)(A), (__v2df)(__m128d)(B),\
_MM_FROUND_CUR_DIRECTION))
+#define _mm_mask_getexp_sd(W, U, A, B) \
+ (__m128d)__builtin_ia32_getexpsd_mask_round(A, B, W, U,\
+ _MM_FROUND_CUR_DIRECTION)
+
+#define _mm_maskz_getexp_sd(U, A, B) \
+ (__m128d)__builtin_ia32_getexpsd_mask_round(A, B, (__v2df)_mm_setzero_pd(), U,\
+ _MM_FROUND_CUR_DIRECTION)
+
#define _mm512_getexp_ps(A) \
((__m512)__builtin_ia32_getexpps512_mask((__v16sf)(__m512)(A), \
(__v16sf)_mm512_undefined_ps(), (__mmask16)-1, _MM_FROUND_CUR_DIRECTION))
DEF_FUNCTION_TYPE (V4SF, V4SF, V4SF, V4SI, INT, UQI)
DEF_FUNCTION_TYPE (V4SF, V4SF, V4SF, V4SI, INT, QI, INT)
DEF_FUNCTION_TYPE (V2DF, V2DF, V2DF, V2DI, INT, QI, INT)
+DEF_FUNCTION_TYPE (V2DF, V2DF, V2DF, INT, V2DF, UQI, INT)
+DEF_FUNCTION_TYPE (V4SF, V4SF, V4SF, INT, V4SF, UQI, INT)
DEF_FUNCTION_TYPE (V16SF, V16SF, V4SF, INT)
DEF_FUNCTION_TYPE (V16SF, V16SF, V4SF, INT, V16SF, UHI)
DEF_FUNCTION_TYPE (V16SF, V16SF, V8SF, INT, V16SF, UHI)
BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_avx512f_getexpv8df_mask_round, "__builtin_ia32_getexppd512_mask", IX86_BUILTIN_GETEXPPD512, UNKNOWN, (int) V8DF_FTYPE_V8DF_V8DF_QI_INT)
BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_avx512f_getexpv16sf_mask_round, "__builtin_ia32_getexpps512_mask", IX86_BUILTIN_GETEXPPS512, UNKNOWN, (int) V16SF_FTYPE_V16SF_V16SF_HI_INT)
BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_avx512f_sgetexpv2df_round, "__builtin_ia32_getexpsd128_round", IX86_BUILTIN_GETEXPSD128, UNKNOWN, (int) V2DF_FTYPE_V2DF_V2DF_INT)
+BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_avx512f_sgetexpv2df_mask_round, "__builtin_ia32_getexpsd_mask_round", IX86_BUILTIN_GETEXPSD_MASK_ROUND, UNKNOWN, (int) V2DF_FTYPE_V2DF_V2DF_V2DF_UQI_INT)
BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_avx512f_sgetexpv4sf_round, "__builtin_ia32_getexpss128_round", IX86_BUILTIN_GETEXPSS128, UNKNOWN, (int) V4SF_FTYPE_V4SF_V4SF_INT)
+BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_avx512f_sgetexpv4sf_mask_round, "__builtin_ia32_getexpss_mask_round", IX86_BUILTIN_GETEXPSS_MASK_ROUND, UNKNOWN, (int) V4SF_FTYPE_V4SF_V4SF_V4SF_UQI_INT)
BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_avx512f_getmantv8df_mask_round, "__builtin_ia32_getmantpd512_mask", IX86_BUILTIN_GETMANTPD512, UNKNOWN, (int) V8DF_FTYPE_V8DF_INT_V8DF_QI_INT)
BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_avx512f_getmantv16sf_mask_round, "__builtin_ia32_getmantps512_mask", IX86_BUILTIN_GETMANTPS512, UNKNOWN, (int) V16SF_FTYPE_V16SF_INT_V16SF_HI_INT)
BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_avx512f_vgetmantv2df_round, "__builtin_ia32_getmantsd_round", IX86_BUILTIN_GETMANTSD128, UNKNOWN, (int) V2DF_FTYPE_V2DF_V2DF_INT_INT)
+BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_avx512f_vgetmantv2df_mask_round, "__builtin_ia32_getmantsd_mask_round", IX86_BUILTIN_GETMANTSD_MASK_ROUND, UNKNOWN, (int) V2DF_FTYPE_V2DF_V2DF_INT_V2DF_UQI_INT)
BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_avx512f_vgetmantv4sf_round, "__builtin_ia32_getmantss_round", IX86_BUILTIN_GETMANTSS128, UNKNOWN, (int) V4SF_FTYPE_V4SF_V4SF_INT_INT)
+BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_avx512f_vgetmantv4sf_mask_round, "__builtin_ia32_getmantss_mask_round", IX86_BUILTIN_GETMANTSS_MASK_ROUND, UNKNOWN, (int) V4SF_FTYPE_V4SF_V4SF_INT_V4SF_UQI_INT)
BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_smaxv8df3_mask_round, "__builtin_ia32_maxpd512_mask", IX86_BUILTIN_MAXPD512, UNKNOWN, (int) V8DF_FTYPE_V8DF_V8DF_V8DF_UQI_INT)
BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_smaxv16sf3_mask_round, "__builtin_ia32_maxps512_mask", IX86_BUILTIN_MAXPS512, UNKNOWN, (int) V16SF_FTYPE_V16SF_V16SF_V16SF_HI_INT)
BDESC (OPTION_MASK_ISA_AVX512F, CODE_FOR_sse2_vmsmaxv2df3_round, "__builtin_ia32_maxsd_round", IX86_BUILTIN_MAXSD_ROUND, UNKNOWN, (int) V2DF_FTYPE_V2DF_V2DF_INT)
case V8DF_FTYPE_V8DF_V8DF_INT_V8DF_QI_INT:
case V4SF_FTYPE_V4SF_V4SF_INT_V4SF_QI_INT:
case V2DF_FTYPE_V2DF_V2DF_INT_V2DF_QI_INT:
+ case V2DF_FTYPE_V2DF_V2DF_INT_V2DF_UQI_INT:
+ case V4SF_FTYPE_V4SF_V4SF_INT_V4SF_UQI_INT:
nargs = 6;
nargs_constant = 4;
break;
case CODE_FOR_avx512f_getmantv8df_mask_round:
case CODE_FOR_avx512f_getmantv16sf_mask_round:
case CODE_FOR_avx512f_vgetmantv2df_round:
+ case CODE_FOR_avx512f_vgetmantv2df_mask_round:
case CODE_FOR_avx512f_vgetmantv4sf_round:
+ case CODE_FOR_avx512f_vgetmantv4sf_mask_round:
error ("the immediate argument must be a 4-bit immediate");
return const0_rtx;
case CODE_FOR_avx512f_cmpv8df3_mask_round:
[(set_attr "prefix" "evex")
(set_attr "mode" "<MODE>")])
-(define_insn "avx512f_sgetexp<mode><round_saeonly_name>"
+(define_insn "avx512f_sgetexp<mode><mask_scalar_name><round_saeonly_scalar_name>"
[(set (match_operand:VF_128 0 "register_operand" "=v")
(vec_merge:VF_128
(unspec:VF_128
[(match_operand:VF_128 1 "register_operand" "v")
- (match_operand:VF_128 2 "<round_saeonly_nimm_predicate>" "<round_saeonly_constraint>")]
+ (match_operand:VF_128 2 "<round_saeonly_scalar_nimm_predicate>" "<round_saeonly_scalar_constraint>")]
UNSPEC_GETEXP)
(match_dup 1)
(const_int 1)))]
"TARGET_AVX512F"
- "vgetexp<ssescalarmodesuffix>\t{<round_saeonly_op3>%2, %1, %0|%0, %1, %2<round_saeonly_op3>}";
+ "vgetexp<ssescalarmodesuffix>\t{<round_saeonly_scalar_mask_op3>%2, %1, %0<mask_scalar_operand3>|%0<mask_scalar_operand3>, %1, %2<round_saeonly_scalar_mask_op3>}";
[(set_attr "prefix" "evex")
(set_attr "mode" "<ssescalarmode>")])
[(set_attr "prefix" "evex")
(set_attr "mode" "<MODE>")])
-(define_insn "avx512f_vgetmant<mode><round_saeonly_name>"
+(define_insn "avx512f_vgetmant<mode><mask_scalar_name><round_saeonly_scalar_name>"
[(set (match_operand:VF_128 0 "register_operand" "=v")
(vec_merge:VF_128
(unspec:VF_128
[(match_operand:VF_128 1 "register_operand" "v")
- (match_operand:VF_128 2 "<round_saeonly_nimm_predicate>" "<round_saeonly_constraint>")
+ (match_operand:VF_128 2 "<round_saeonly_scalar_nimm_predicate>" "<round_saeonly_scalar_constraint>")
(match_operand:SI 3 "const_0_to_15_operand")]
UNSPEC_GETMANT)
(match_dup 1)
(const_int 1)))]
"TARGET_AVX512F"
- "vgetmant<ssescalarmodesuffix>\t{%3, <round_saeonly_op4>%2, %1, %0|%0, %1, %2<round_saeonly_op4>, %3}";
+ "vgetmant<ssescalarmodesuffix>\t{%3, <round_saeonly_scalar_mask_op4>%2, %1, %0<mask_scalar_operand4>|%0<mask_scalar_operand4>, %1, %2<round_saeonly_scalar_mask_op4>, %3}";
[(set_attr "prefix" "evex")
(set_attr "mode" "<ssescalarmode>")])
(define_subst_attr "mask_scalar_name" "mask_scalar" "" "_mask")
(define_subst_attr "mask_scalar_operand3" "mask_scalar" "" "%{%4%}%N3")
+(define_subst_attr "mask_scalar_operand4" "mask_scalar" "" "%{%5%}%N4")
(define_subst "mask_scalar"
[(set (match_operand:SUBST_V 0)
(define_subst_attr "round_saeonly_scalar_name" "round_saeonly_scalar" "" "_round")
(define_subst_attr "round_saeonly_scalar_mask_operand3" "mask_scalar" "%r3" "%r5")
+(define_subst_attr "round_saeonly_scalar_mask_operand4" "mask_scalar" "%r4" "%r6")
(define_subst_attr "round_saeonly_scalar_mask_op3" "round_saeonly_scalar" "" "<round_saeonly_scalar_mask_operand3>")
+(define_subst_attr "round_saeonly_scalar_mask_op4" "round_saeonly_scalar" "" "<round_saeonly_scalar_mask_operand4>")
(define_subst_attr "round_saeonly_scalar_constraint" "round_saeonly_scalar" "vm" "v")
(define_subst_attr "round_saeonly_scalar_prefix" "round_saeonly_scalar" "vex" "evex")
+(define_subst_attr "round_saeonly_scalar_nimm_predicate" "round_saeonly_scalar" "vector_operand" "register_operand")
(define_subst "round_saeonly_scalar"
[(set (match_operand:SUBST_V 0)
+2017-07-06 Sebastian Peryt <sebastian.peryt@intel.com>
+
+ * gcc.target/i386/avx512f-vgetexpsd-1.c (_mm_mask_getexp_sd,
+ _mm_maskz_getexp_sd, _mm_mask_getexp_round_sd,
+ _mm_maskz_getexp_round_sd): Test new intrinsics.
+ * gcc.target/i386/avx512f-vgetexpss-1.c (_mm_mask_getexp_ss,
+ _mm_maskz_getexp_ss, _mm_mask_getexp_round_ss,
+ _mm_maskz_getexp_round_ss): Ditto.
+ * gcc.target/i386/avx512f-vgetmantsd-1.c (_mm_mask_getmant_sd,
+ _mm_maskz_getmant_sd, _mm_mask_getmant_round_sd,
+ _mm_maskz_getmant_round_sd): Ditto.
+ * gcc.target/i386/avx512f-vgetmantss-1.c (_mm_mask_getmant_ss,
+ _mm_maskz_getmant_ss, _mm_mask_getmant_round_ss,
+ _mm_maskz_getmant_round_ss): Ditto.
+ * gcc.target/i386/avx512f-vgetexpsd-2.c (_mm_mask_getexp_sd,
+ _mm_maskz_getexp_sd, _mm_getexp_round_sd, _mm_mask_getexp_round_sd,
+ _mm_maskz_getexp_round_sd): New runtime tests.
+ * gcc.target/i386/avx512f-vgetexpss-2.c (_mm_mask_getexp_ss,
+ _mm_maskz_getexp_ss, _mm_getexp_round_ss, _mm_mask_getexp_round_ss,
+ _mm_maskz_getexp_round_ss): Ditto.
+ * gcc.target/i386/avx512f-vgetmantsd-2.c (_mm_mask_getmant_sd,
+ _mm_maskz_getmant_sd, _mm_getmant_round_sd, _mm_mask_getmant_round_sd,
+ _mm_maskz_getmant_round_sd): Ditto.
+ * gcc.target/i386/avx512f-vgetmantss-2.c (_mm_mask_getmant_ss,
+ _mm_maskz_getmant_ss, _mm_getmant_round_ss, _mm_mask_getmant_round_ss,
+ _mm_maskz_getmant_round_ss): Ditto.
+ * gcc.target/i386/avx-1.c (__builtin_ia32_getexpsd_mask_round,
+ __builtin_ia32_getexpss_mask_round, __builtin_ia32_getmantsd_mask_round,
+ __builtin_ia32_getmantss_mask_round): Test new builtins.
+ * gcc.target/i386/sse-13.c : Ditto.
+ * gcc.target/i386/sse-23.c: Ditto.
+ * gcc.target/i386/sse-14.c (_mm_maskz_getexp_round_sd,
+ _mm_maskz_getexp_round_ss, _mm_mask_getmant_round_sd,
+ _mm_maskz_getmant_round_sd, _mm_mask_getmant_round_ss,
+ _mm_maskz_getmant_round_ss, _mm_mask_getexp_round_sd,
+ _mm_mask_getexp_round_ss): Test new intrinsics.
+ * gcc.target/i386/testround-1.c: Ditto.
+ * gcc.target/i386/sse-22.c (_mm_maskz_getmant_round_sd,
+ _mm_maskz_getmant_round_ss, _mm_mask_getmant_round_sd,
+ _mm_mask_getmant_round_ss): Test new intrinsics
+ * gcc.target/i386/testimm-10.c (_mm_mask_getmant_sd,
+ _mm_maskz_getmant_sd, _mm_mask_getmant_ss,
+ _mm_maskz_getmant_ss): Test new intrinsics.
+
2017-07-06 Christophe Lyon <christophe.lyon@linaro.org>
* lib/target-supports-dg.exp (dg-require-stack-check): New.
#define __builtin_ia32_getexppd512_mask(A, B, C, D) __builtin_ia32_getexppd512_mask(A, B, C, 8)
#define __builtin_ia32_getexpps512_mask(A, B, C, D) __builtin_ia32_getexpps512_mask(A, B, C, 8)
#define __builtin_ia32_getexpsd128_round(A, B, C) __builtin_ia32_getexpsd128_round(A, B, 4)
+#define __builtin_ia32_getexpsd_mask_round(A, B, C, D, E) __builtin_ia32_getexpsd_mask_round(A, B, C, D, 4)
#define __builtin_ia32_getexpss128_round(A, B, C) __builtin_ia32_getexpss128_round(A, B, 4)
+#define __builtin_ia32_getexpss_mask_round(A, B, C, D, E) __builtin_ia32_getexpss_mask_round(A, B, C, D, 4)
#define __builtin_ia32_getmantpd512_mask(A, F, C, D, E) __builtin_ia32_getmantpd512_mask(A, 1, C, D, 8)
#define __builtin_ia32_getmantps512_mask(A, F, C, D, E) __builtin_ia32_getmantps512_mask(A, 1, C, D, 8)
#define __builtin_ia32_getmantsd_round(A, B, C, D) __builtin_ia32_getmantsd_round(A, B, 1, 4)
+#define __builtin_ia32_getmantsd_mask_round(A, B, C, W, U, D) __builtin_ia32_getmantsd_mask_round(A, B, 1, W, U, 4)
#define __builtin_ia32_getmantss_round(A, B, C, D) __builtin_ia32_getmantss_round(A, B, 1, 4)
+#define __builtin_ia32_getmantss_mask_round(A, B, C, W, U, D) __builtin_ia32_getmantss_mask_round(A, B, 1, W, U, 4)
#define __builtin_ia32_insertf32x4_mask(A, B, F, D, E) __builtin_ia32_insertf32x4_mask(A, B, 1, D, E)
#define __builtin_ia32_insertf64x4_mask(A, B, F, D, E) __builtin_ia32_insertf64x4_mask(A, B, 1, D, E)
#define __builtin_ia32_inserti32x4_mask(A, B, F, D, E) __builtin_ia32_inserti32x4_mask(A, B, 1, D, E)
/* { dg-do compile } */
/* { dg-options "-mavx512f -O2" } */
/* { dg-final { scan-assembler-times "vgetexpsd\[ \\t\]+\[^\{\n\]*%xmm\[0-9\]+\, %xmm\[0-9\]+(?:\n|\[ \\t\]+#)" 1 } } */
+/* { dg-final { scan-assembler-times "vgetexpsd\[ \\t\]+\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}(?:\n|\[ \\t\]+#)" 1 } } */
+/* { dg-final { scan-assembler-times "vgetexpsd\[ \\t\]+\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}\{z\}(?:\n|\[ \\t\]+#)" 1 } } */
/* { dg-final { scan-assembler-times "vgetexpsd\[ \\t\]+\[^\{\n\]*\{sae\}\[^\n\]*%xmm\[0-9\]+\, %xmm\[0-9\]+(?:\n|\[ \\t\]+#)" 1 } } */
+/* { dg-final { scan-assembler-times "vgetexpsd\[ \\t\]+\[^\n\]*\{sae\}\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}(?:\n|\[ \\t\]+#)" 1 } } */
+/* { dg-final { scan-assembler-times "vgetexpsd\[ \\t\]+\[^\n\]*\{sae\}\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}\{z\}(?:\n|\[ \\t\]+#)" 1 } } */
#include <immintrin.h>
volatile __m128d x;
+volatile __mmask8 m;
void extern
avx512f_test (void)
{
x = _mm_getexp_sd (x, x);
+ x = _mm_mask_getexp_sd (x, m, x, x);
+ x = _mm_maskz_getexp_sd (m, x, x);
x = _mm_getexp_round_sd (x, x, _MM_FROUND_NO_EXC);
+ x = _mm_mask_getexp_round_sd (x, m, x, x, _MM_FROUND_NO_EXC);
+ x = _mm_maskz_getexp_round_sd (m, x, x, _MM_FROUND_NO_EXC);
}
#include <math.h>
#include "avx512f-check.h"
+#include "avx512f-helper.h"
+#include "avx512f-mask-type.h"
static void
compute_vgetexpsd (double *s, double *r)
{
r[0] = floor (log (s[0]) / log (2));
+ r[1] = s[1];
}
void static
avx512f_test (void)
{
int i;
- union128d res1, s1;
+ union128d res1, res2, res3, res4, res5, res6, s1;
+ MASK_TYPE mask = MASK_VALUE;
double res_ref[SIZE];
for (i = 0; i < SIZE; i++)
{
s1.a[i] = 5.0 - i;
- res_ref[i] = s1.a[i];
+ res2.a[i] = DEFAULT_VALUE;
+ res5.a[i] = DEFAULT_VALUE;
}
res1.x = _mm_getexp_sd (s1.x, s1.x);
+ res2.x = _mm_mask_getexp_sd (res2.x, mask, s1.x, s1.x);
+ res3.x = _mm_maskz_getexp_sd (mask, s1.x, s1.x);
+ res4.x = _mm_getexp_round_sd (s1.x, s1.x, _MM_FROUND_NO_EXC);
+ res5.x = _mm_mask_getexp_round_sd (res5.x, mask, s1.x, s1.x, _MM_FROUND_NO_EXC);
+ res6.x = _mm_maskz_getexp_round_sd (mask, s1.x, s1.x, _MM_FROUND_NO_EXC);
compute_vgetexpsd (s1.a, res_ref);
if (check_fp_union128d (res1, res_ref))
abort ();
+
+ MASK_MERGE (d) (res_ref, mask, 1);
+ if (check_fp_union128d (res2, res_ref))
+ abort ();
+
+ MASK_ZERO (d) (res_ref, mask, 1);
+ if (check_fp_union128d (res3, res_ref))
+ abort ();
+
+ compute_vgetexpsd (s1.a, res_ref);
+
+ if (check_fp_union128d (res4, res_ref))
+ abort ();
+
+ MASK_MERGE (d) (res_ref, mask, 1);
+ if (check_fp_union128d (res5, res_ref))
+ abort ();
+
+ MASK_ZERO (d) (res_ref, mask, 1);
+ if (check_fp_union128d (res6, res_ref))
+ abort ();
}
/* { dg-do compile } */
/* { dg-options "-mavx512f -O2" } */
/* { dg-final { scan-assembler-times "vgetexpss\[ \\t\]+\[^\{\n\]*%xmm\[0-9\]+\, %xmm\[0-9\]+(?:\n|\[ \\t\]+#)" 1 } } */
+/* { dg-final { scan-assembler-times "vgetexpss\[ \\t\]+\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}(?:\n|\[ \\t\]+#)" 1 } } */
+/* { dg-final { scan-assembler-times "vgetexpss\[ \\t\]+\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}\{z\}(?:\n|\[ \\t\]+#)" 1 } } */
/* { dg-final { scan-assembler-times "vgetexpss\[ \\t\]+\[^\{\n\]*\{sae\}\[^\n\]*%xmm\[0-9\]+\, %xmm\[0-9\]+(?:\n|\[ \\t\]+#)" 1 } } */
+/* { dg-final { scan-assembler-times "vgetexpss\[ \\t\]+\[^\n\]*\{sae\}\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}(?:\n|\[ \\t\]+#)" 1 } } */
+/* { dg-final { scan-assembler-times "vgetexpss\[ \\t\]+\[^\n\]*\{sae\}\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}\{z\}(?:\n|\[ \\t\]+#)" 1 } } */
#include <immintrin.h>
volatile __m128 x;
+volatile __mmask8 m;
void extern
avx512f_test (void)
{
x = _mm_getexp_ss (x, x);
+ x = _mm_mask_getexp_ss (x, m, x, x);
+ x = _mm_maskz_getexp_ss (m, x, x);
x = _mm_getexp_round_ss (x, x, _MM_FROUND_NO_EXC);
+ x = _mm_mask_getexp_round_ss (x, m, x, x, _MM_FROUND_NO_EXC);
+ x = _mm_maskz_getexp_round_ss (m, x, x, _MM_FROUND_NO_EXC);
}
#include <math.h>
#include "avx512f-check.h"
+#include "avx512f-helper.h"
+#include "avx512f-mask-type.h"
static void
compute_vgetexpss (float *s, float *r)
{
+ int i;
r[0] = floor (log (s[0]) / log (2));
+ for(i = 1; i < SIZE; i++)
+ r[i] = s[i];
}
void static
avx512f_test (void)
{
int i;
- union128 res1, s1;
+ union128 res1, res2, res3, res4, res5, res6, s1;
+ MASK_TYPE mask = MASK_VALUE;
float res_ref[SIZE];
for (i = 0; i < SIZE; i++)
{
s1.a[i] = 5.0 - i;
- res_ref[i] = s1.a[i];
+ res2.a[i] = DEFAULT_VALUE;
+ res5.a[i] = DEFAULT_VALUE;
}
res1.x = _mm_getexp_ss (s1.x, s1.x);
+ res2.x = _mm_mask_getexp_ss (res2.x, mask, s1.x, s1.x);
+ res3.x = _mm_maskz_getexp_ss (mask, s1.x, s1.x);
+ res4.x = _mm_getexp_round_ss (s1.x, s1.x, _MM_FROUND_NO_EXC);
+ res5.x = _mm_mask_getexp_round_ss (res5.x, mask, s1.x, s1.x, _MM_FROUND_NO_EXC);
+ res6.x = _mm_maskz_getexp_round_ss (mask, s1.x, s1.x, _MM_FROUND_NO_EXC);
compute_vgetexpss (s1.a, res_ref);
if (check_fp_union128 (res1, res_ref))
abort ();
+
+ MASK_MERGE () (res_ref, mask, 1);
+ if (check_fp_union128 (res2, res_ref))
+ abort ();
+
+ MASK_ZERO () (res_ref, mask, 1);
+ if (check_fp_union128 (res3, res_ref))
+ abort ();
+
+ compute_vgetexpss (s1.a, res_ref);
+
+ if (check_fp_union128 (res4, res_ref))
+ abort ();
+
+ MASK_MERGE () (res_ref, mask, 1);
+ if (check_fp_union128 (res5, res_ref))
+ abort ();
+
+ MASK_ZERO () (res_ref, mask, 1);
+ if (check_fp_union128 (res6, res_ref))
+ abort ();
}
/* { dg-do compile } */
/* { dg-options "-O2 -mavx512f" } */
/* { dg-final { scan-assembler-times "vgetmantsd\[ \\t\]+\[^\{\n\]*%xmm\[0-9\]+(?:\n|\[ \\t\]+#)" 1 } } */
+/* { dg-final { scan-assembler-times "vgetmantsd\[ \\t\]+\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}(?:\n|\[ \\t\]+#)" 1 } } */
+/* { dg-final { scan-assembler-times "vgetmantsd\[ \\t\]+\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}\{z\}(?:\n|\[ \\t\]+#)" 1 } } */
/* { dg-final { scan-assembler-times "vgetmantsd\[ \\t\]+\[^\{\n\]*\{sae\}\[^\n\]*%xmm\[0-9\]+(?:\n|\[ \\t\]+#)" 1 } } */
+/* { dg-final { scan-assembler-times "vgetmantsd\[ \\t\]+\[^\n\]*\{sae\}\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}(?:\n|\[ \\t\]+#)" 1 } } */
+/* { dg-final { scan-assembler-times "vgetmantsd\[ \\t\]+\[^\n\]*\{sae\}\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}\{z\}(?:\n|\[ \\t\]+#)" 1 } } */
#include <immintrin.h>
volatile __m128d x, y, z;
+volatile __mmask8 m;
void extern
avx512f_test (void)
{
x = _mm_getmant_sd (y, z, _MM_MANT_NORM_p75_1p5, _MM_MANT_SIGN_src);
- x = _mm_getmant_round_sd (y, z, _MM_MANT_NORM_p75_1p5, _MM_MANT_SIGN_src,
- _MM_FROUND_NO_EXC);
+ x = _mm_mask_getmant_sd (x, m, y, z, _MM_MANT_NORM_p75_1p5, _MM_MANT_SIGN_src);
+ x = _mm_maskz_getmant_sd (m, y, z, _MM_MANT_NORM_p75_1p5, _MM_MANT_SIGN_src);
+ x = _mm_getmant_round_sd (y, z, _MM_MANT_NORM_p75_1p5, _MM_MANT_SIGN_src,_MM_FROUND_NO_EXC);
+ x = _mm_mask_getmant_round_sd (x, m, y, z, _MM_MANT_NORM_p75_1p5, _MM_MANT_SIGN_src, _MM_FROUND_NO_EXC);
+ x = _mm_maskz_getmant_round_sd (m, y, z, _MM_MANT_NORM_p75_1p5, _MM_MANT_SIGN_src, _MM_FROUND_NO_EXC);
}
#include "avx512f-check.h"
#include <math.h>
+#include "avx512f-helper.h"
+
+#define SIZE (128/64)
+#include "avx512f-mask-type.h"
union fp_int_t
{
static void
avx512f_test (void)
{
- union128d res1, src1, src2;
+ union128d res1, res2, res3, res4, res5, res6, src1, src2;
double res_ref[2];
- int interv = _MM_MANT_NORM_p5_1;
- int signctrl = _MM_MANT_SIGN_src;
+ MASK_TYPE mask = MASK_VALUE;
src1.x = _mm_set_pd (-3.0, 111.111);
src2.x = _mm_set_pd (222.222, -2.0);
+
+ int i;
+ for (i = 0; i < SIZE; i++)
+ {
+ res2.a[i] = DEFAULT_VALUE;
+ res5.a[i] = DEFAULT_VALUE;
+ }
- res1.x = _mm_getmant_sd (src1.x, src2.x, interv, signctrl);
+ res1.x = _mm_getmant_sd (src1.x, src2.x, _MM_MANT_NORM_p5_1, _MM_MANT_SIGN_src);
+ res2.x = _mm_mask_getmant_sd (res2.x, mask, src1.x, src2.x, _MM_MANT_NORM_p5_1, _MM_MANT_SIGN_src);
+ res3.x = _mm_maskz_getmant_sd (mask, src1.x, src2.x, _MM_MANT_NORM_p5_1, _MM_MANT_SIGN_src);
+ res4.x = _mm_getmant_round_sd (src1.x, src2.x, _MM_MANT_NORM_p5_1, _MM_MANT_SIGN_src, _MM_FROUND_NO_EXC);
+ res5.x = _mm_mask_getmant_round_sd (res5.x, mask, src1.x, src2.x, _MM_MANT_NORM_p5_1, _MM_MANT_SIGN_src, _MM_FROUND_NO_EXC);
+ res6.x = _mm_maskz_getmant_round_sd (mask, src1.x, src2.x, _MM_MANT_NORM_p5_1, _MM_MANT_SIGN_src, _MM_FROUND_NO_EXC);
- compute_vgetmantsd (res_ref, src1.a, src2.a, interv, signctrl);
+ compute_vgetmantsd (res_ref, src1.a, src2.a, _MM_MANT_NORM_p5_1, _MM_MANT_SIGN_src);
if (check_union128d (res1, res_ref))
abort ();
+
+ MASK_MERGE (d) (res_ref, mask, 1);
+ if (check_union128d (res2, res_ref))
+ abort ();
+
+ MASK_ZERO (d) (res_ref, mask, 1);
+ if (check_union128d (res3, res_ref))
+ abort ();
+
+ compute_vgetmantsd (res_ref, src1.a, src2.a, _MM_MANT_NORM_p5_1, _MM_MANT_SIGN_src);
+
+ if (check_union128d (res4, res_ref))
+ abort ();
+
+ MASK_MERGE (d) (res_ref, mask, 1);
+ if (check_union128d (res5, res_ref))
+ abort ();
+
+ MASK_ZERO (d) (res_ref, mask, 1);
+ if (check_union128d (res6, res_ref))
+ abort ();
+
}
/* { dg-do compile } */
/* { dg-options "-O2 -mavx512f" } */
/* { dg-final { scan-assembler-times "vgetmantss\[ \\t\]+\[^\{\n\]*%xmm\[0-9\]+(?:\n|\[ \\t\]+#)" 1 } } */
+/* { dg-final { scan-assembler-times "vgetmantss\[ \\t\]+\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}(?:\n|\[ \\t\]+#)" 1 } } */
+/* { dg-final { scan-assembler-times "vgetmantss\[ \\t\]+\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}\{z\}(?:\n|\[ \\t\]+#)" 1 } } */
/* { dg-final { scan-assembler-times "vgetmantss\[ \\t\]+\[^\{\n\]*\{sae\}\[^\n\]*%xmm\[0-9\]+(?:\n|\[ \\t\]+#)" 1 } } */
+/* { dg-final { scan-assembler-times "vgetmantss\[ \\t\]+\[^\n\]*\{sae\}\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}(?:\n|\[ \\t\]+#)" 1 } } */
+/* { dg-final { scan-assembler-times "vgetmantss\[ \\t\]+\[^\n\]*\{sae\}\[^\{\n\]*%xmm\[0-9\]+\{%k\[1-7\]\}\{z\}(?:\n|\[ \\t\]+#)" 1 } } */
#include <immintrin.h>
volatile __m128 x, y, z;
+volatile __mmask8 m;
void extern
avx512f_test (void)
{
x = _mm_getmant_ss (y, z, _MM_MANT_NORM_p75_1p5, _MM_MANT_SIGN_src);
- x = _mm_getmant_round_ss (y, z, _MM_MANT_NORM_p75_1p5, _MM_MANT_SIGN_src,
- _MM_FROUND_NO_EXC);
+ x = _mm_mask_getmant_ss (x, m, y, z, _MM_MANT_NORM_p75_1p5, _MM_MANT_SIGN_src);
+ x = _mm_maskz_getmant_ss (m, y, z, _MM_MANT_NORM_p75_1p5, _MM_MANT_SIGN_src);
+ x = _mm_getmant_round_ss (y, z, _MM_MANT_NORM_p75_1p5, _MM_MANT_SIGN_src, _MM_FROUND_NO_EXC);
+ x = _mm_mask_getmant_round_ss (x, m, y, z, _MM_MANT_NORM_p75_1p5, _MM_MANT_SIGN_src, _MM_FROUND_NO_EXC);
+ x = _mm_maskz_getmant_round_ss (m, y, z, _MM_MANT_NORM_p75_1p5, _MM_MANT_SIGN_src, _MM_FROUND_NO_EXC);
}
#include "avx512f-check.h"
#include <math.h>
+#include "avx512f-helper.h"
+
+#define SIZE (128/32)
+#include "avx512f-mask-type.h"
union fp_int_t
{
{
int i;
r[0] = get_norm_mant (s2[0], signctrl, interv);
- for (i = 1; i < 4; i++)
+ for (i = 1; i < SIZE; i++)
{
r[i] = s1[i];
}
static void
avx512f_test (void)
{
- union128 res1, src1, src2;
+ union128 res1, res2, res3, res4, res5, res6, src1, src2;
float res_ref[4];
- int interv = _MM_MANT_NORM_p5_1;
- int signctrl = _MM_MANT_SIGN_src;
+ MASK_TYPE mask = MASK_VALUE;
src1.x = _mm_set_ps (-24.043, 68.346, -43.35, 546.46);
src2.x = _mm_set_ps (222.222, 333.333, 444.444, -2.0);
- res1.x = _mm_getmant_ss (src1.x, src2.x, interv, signctrl);
+ int i;
+ for (i = 0; i < SIZE; i++)
+ {
+ res2.a[i] = DEFAULT_VALUE;
+ res5.a[i] = DEFAULT_VALUE;
+ }
+
+ res1.x = _mm_getmant_ss (src1.x, src2.x, _MM_MANT_NORM_p5_1, _MM_MANT_SIGN_src);
+ res2.x = _mm_mask_getmant_ss (res2.x, mask, src1.x, src2.x, _MM_MANT_NORM_p5_1, _MM_MANT_SIGN_src);
+ res3.x = _mm_maskz_getmant_ss (mask, src1.x, src2.x, _MM_MANT_NORM_p5_1, _MM_MANT_SIGN_src);
+ res4.x = _mm_getmant_round_ss (src1.x, src2.x, _MM_MANT_NORM_p5_1, _MM_MANT_SIGN_src, _MM_FROUND_NO_EXC);
+ res5.x = _mm_mask_getmant_round_ss (res5.x, mask, src1.x, src2.x, _MM_MANT_NORM_p5_1, _MM_MANT_SIGN_src, _MM_FROUND_NO_EXC);
+ res6.x = _mm_maskz_getmant_round_ss (mask, src1.x, src2.x, _MM_MANT_NORM_p5_1, _MM_MANT_SIGN_src, _MM_FROUND_NO_EXC);
- compute_vgetmantss (res_ref, src1.a, src2.a, interv, signctrl);
+ compute_vgetmantss (res_ref, src1.a, src2.a, _MM_MANT_NORM_p5_1, _MM_MANT_SIGN_src);
if (check_union128 (res1, res_ref))
abort ();
+
+ MASK_MERGE () (res_ref, mask, 1);
+ if (check_union128 (res2, res_ref))
+ abort ();
+
+ MASK_ZERO () (res_ref, mask, 1);
+ if (check_union128 (res3, res_ref))
+ abort ();
+
+ compute_vgetmantss (res_ref, src1.a, src2.a, _MM_MANT_NORM_p5_1, _MM_MANT_SIGN_src);
+
+ if (check_union128 (res4, res_ref))
+ abort ();
+
+ MASK_MERGE () (res_ref, mask, 1);
+ if (check_union128 (res5, res_ref))
+ abort ();
+
+ MASK_ZERO () (res_ref, mask, 1);
+ if (check_union128 (res6, res_ref))
+ abort ();
}
#define __builtin_ia32_getexppd512_mask(A, B, C, D) __builtin_ia32_getexppd512_mask(A, B, C, 8)
#define __builtin_ia32_getexpps512_mask(A, B, C, D) __builtin_ia32_getexpps512_mask(A, B, C, 8)
#define __builtin_ia32_getexpsd128_round(A, B, C) __builtin_ia32_getexpsd128_round(A, B, 4)
+#define __builtin_ia32_getexpsd_mask_round(A, B, C, D, E) __builtin_ia32_getexpsd_mask_round(A, B, C, D, 4)
#define __builtin_ia32_getexpss128_round(A, B, C) __builtin_ia32_getexpss128_round(A, B, 4)
+#define __builtin_ia32_getexpss_mask_round(A, B, C, D, E) __builtin_ia32_getexpss_mask_round(A, B, C, D, 4)
#define __builtin_ia32_getmantpd512_mask(A, F, C, D, E) __builtin_ia32_getmantpd512_mask(A, 1, C, D, 8)
#define __builtin_ia32_getmantps512_mask(A, F, C, D, E) __builtin_ia32_getmantps512_mask(A, 1, C, D, 8)
#define __builtin_ia32_getmantsd_round(A, B, C, D) __builtin_ia32_getmantsd_round(A, B, 1, 4)
+#define __builtin_ia32_getmantsd_mask_round(A, B, C, W, U, D) __builtin_ia32_getmantsd_mask_round(A, B, 1, W, U, 4)
#define __builtin_ia32_getmantss_round(A, B, C, D) __builtin_ia32_getmantss_round(A, B, 1, 4)
+#define __builtin_ia32_getmantss_mask_round(A, B, C, W, U, D) __builtin_ia32_getmantss_mask_round(A, B, 1, W, U, 4)
#define __builtin_ia32_insertf32x4_mask(A, B, F, D, E) __builtin_ia32_insertf32x4_mask(A, B, 1, D, E)
#define __builtin_ia32_insertf64x4_mask(A, B, F, D, E) __builtin_ia32_insertf64x4_mask(A, B, 1, D, E)
#define __builtin_ia32_inserti32x4_mask(A, B, F, D, E) __builtin_ia32_inserti32x4_mask(A, B, 1, D, E)
test_2 (_mm512_maskz_extracti32x4_epi32, __m128i, __mmask8, __m512i, 1)
test_2 (_mm512_maskz_extracti64x4_epi64, __m256i, __mmask8, __m512i, 1)
test_2 (_mm512_maskz_getexp_round_pd, __m512d, __mmask8, __m512d, 8)
+test_3 (_mm_maskz_getexp_round_sd, __m128d, __mmask8, __m128d, __m128d, 8)
test_2 (_mm512_maskz_getexp_round_ps, __m512, __mmask16, __m512, 8)
+test_3 (_mm_maskz_getexp_round_ss, __m128, __mmask8, __m128, __m128, 8)
test_2y (_mm512_maskz_getmant_round_pd, __m512d, __mmask8, __m512d, 1, 1, 8)
test_2y (_mm512_maskz_getmant_round_ps, __m512, __mmask16, __m512, 1, 1, 8)
test_2 (_mm512_maskz_permute_pd, __m512d, __mmask8, __m512d, 1)
test_2 (_mm_getexp_round_sd, __m128d, __m128d, __m128d, 8)
test_2 (_mm_getexp_round_ss, __m128, __m128, __m128, 8)
test_2y (_mm_getmant_round_sd, __m128d, __m128d, __m128d, 1, 1, 8)
+test_4y (_mm_mask_getmant_round_sd, __m128d, __m128d, __mmask8, __m128d, __m128d, 1, 1, 8)
+test_3y (_mm_maskz_getmant_round_sd, __m128d, __mmask8, __m128d, __m128d, 1, 1, 8)
test_2y (_mm_getmant_round_ss, __m128, __m128, __m128, 1, 1, 8)
+test_4y (_mm_mask_getmant_round_ss, __m128, __m128, __mmask8, __m128, __m128, 1, 1, 8)
+test_3y (_mm_maskz_getmant_round_ss, __m128, __mmask8, __m128, __m128, 1, 1, 8)
test_2 (_mm_mul_round_sd, __m128d, __m128d, __m128d, 9)
test_2 (_mm_mul_round_ss, __m128, __m128, __m128, 9)
test_2 (_mm_scalef_round_sd, __m128d, __m128d, __m128d, 9)
test_3 (_mm512_mask_extracti32x4_epi32, __m128i, __m128i, __mmask8, __m512i, 1)
test_3 (_mm512_mask_extracti64x4_epi64, __m256i, __m256i, __mmask8, __m512i, 1)
test_3 (_mm512_mask_getexp_round_pd, __m512d, __m512d, __mmask8, __m512d, 8)
+test_4 (_mm_mask_getexp_round_sd, __m128d, __m128d, __mmask8, __m128d, __m128d, 8)
test_3 (_mm512_mask_getexp_round_ps, __m512, __m512, __mmask16, __m512, 8)
+test_4 (_mm_mask_getexp_round_ss, __m128, __m128, __mmask8, __m128, __m128, 8)
test_3y (_mm512_mask_getmant_round_pd, __m512d, __m512d, __mmask8, __m512d, 1, 1, 8)
test_3y (_mm512_mask_getmant_round_ps, __m512, __m512, __mmask16, __m512, 1, 1, 8)
test_3 (_mm512_mask_permute_pd, __m512d, __m512d, __mmask8, __m512d, 1)
test_2 (_mm512_maskz_getexp_round_pd, __m512d, __mmask8, __m512d, 8)
test_2 (_mm512_maskz_getexp_round_ps, __m512, __mmask16, __m512, 8)
test_2y (_mm512_maskz_getmant_round_pd, __m512d, __mmask8, __m512d, 1, 1, 8)
+test_3y (_mm_maskz_getmant_round_sd, __m128d, __mmask8, __m128d, __m128d, 1, 1, 8)
test_2y (_mm512_maskz_getmant_round_ps, __m512, __mmask16, __m512, 1, 1, 8)
+test_3y (_mm_maskz_getmant_round_ss, __m128, __mmask8, __m128, __m128, 1, 1, 8)
test_2 (_mm512_maskz_permute_pd, __m512d, __mmask8, __m512d, 1)
test_2 (_mm512_maskz_permute_ps, __m512, __mmask16, __m512, 1)
test_2 (_mm512_maskz_permutex_epi64, __m512i, __mmask8, __m512i, 1)
test_3 (_mm512_mask_getexp_round_pd, __m512d, __m512d, __mmask8, __m512d, 8)
test_3 (_mm512_mask_getexp_round_ps, __m512, __m512, __mmask16, __m512, 8)
test_3y (_mm512_mask_getmant_round_pd, __m512d, __m512d, __mmask8, __m512d, 1, 1, 8)
+test_4y (_mm_mask_getmant_round_sd, __m128d, __m128d, __mmask8, __m128d, __m128d, 1, 1, 8)
test_3y (_mm512_mask_getmant_round_ps, __m512, __m512, __mmask16, __m512, 1, 1, 8)
+test_4y (_mm_mask_getmant_round_ss, __m128, __m128, __mmask8, __m128, __m128, 1, 1, 8)
test_3 (_mm512_mask_permute_pd, __m512d, __m512d, __mmask8, __m512d, 1)
test_3 (_mm512_mask_permute_ps, __m512, __m512, __mmask16, __m512, 1)
test_3 (_mm512_mask_permutex_epi64, __m512i, __m512i, __mmask8, __m512i, 1)
#define __builtin_ia32_getexppd512_mask(A, B, C, D) __builtin_ia32_getexppd512_mask(A, B, C, 8)
#define __builtin_ia32_getexpps512_mask(A, B, C, D) __builtin_ia32_getexpps512_mask(A, B, C, 8)
#define __builtin_ia32_getexpsd128_round(A, B, C) __builtin_ia32_getexpsd128_round(A, B, 4)
+#define __builtin_ia32_getexpsd_mask_round(A, B, C, D, E) __builtin_ia32_getexpsd_mask_round(A, B, C, D, 4)
#define __builtin_ia32_getexpss128_round(A, B, C) __builtin_ia32_getexpss128_round(A, B, 4)
+#define __builtin_ia32_getexpss_mask_round(A, B, C, D, E) __builtin_ia32_getexpss_mask_round(A, B, C, D, 4)
#define __builtin_ia32_getmantpd512_mask(A, F, C, D, E) __builtin_ia32_getmantpd512_mask(A, 1, C, D, 8)
#define __builtin_ia32_getmantps512_mask(A, F, C, D, E) __builtin_ia32_getmantps512_mask(A, 1, C, D, 8)
#define __builtin_ia32_getmantsd_round(A, B, C, D) __builtin_ia32_getmantsd_round(A, B, 1, 4)
+#define __builtin_ia32_getmantsd_mask_round(A, B, C, W, U, D) __builtin_ia32_getmantsd_mask_round(A, B, 1, W, U, 4)
#define __builtin_ia32_getmantss_round(A, B, C, D) __builtin_ia32_getmantss_round(A, B, 1, 4)
+#define __builtin_ia32_getmantss_mask_round(A, B, C, W, U, D) __builtin_ia32_getmantss_mask_round(A, B, 1, W, U, 4)
#define __builtin_ia32_insertf32x4_mask(A, B, F, D, E) __builtin_ia32_insertf32x4_mask(A, B, 1, D, E)
#define __builtin_ia32_insertf64x4_mask(A, B, F, D, E) __builtin_ia32_insertf64x4_mask(A, B, 1, D, E)
#define __builtin_ia32_inserti32x4_mask(A, B, F, D, E) __builtin_ia32_inserti32x4_mask(A, B, 1, D, E)
m512 = _mm512_maskz_getmant_ps (mmask16, m512, 1, 64); /* { dg-error "the immediate argument must be a 4-bit immediate" } */
m128d = _mm_getmant_sd (m128d, m128d, 1, 64); /* { dg-error "the immediate argument must be a 4-bit immediate" } */
+ m128d = _mm_mask_getmant_sd (m128d, mmask8, m128d, m128d, 1, 64); /* { dg-error "the immediate argument must be a 4-bit immediate" } */
+ m128d = _mm_maskz_getmant_sd (mmask8, m128d, m128d, 1, 64); /* { dg-error "the immediate argument must be a 4-bit immediate" } */
m128 = _mm_getmant_ss (m128, m128, 1, 64); /* { dg-error "the immediate argument must be a 4-bit immediate" } */
+ m128 = _mm_mask_getmant_ss (m128, mmask8, m128, m128, 1, 64); /* { dg-error "the immediate argument must be a 4-bit immediate" } */
+ m128 = _mm_maskz_getmant_ss (mmask8, m128, m128, 1, 64); /* { dg-error "the immediate argument must be a 4-bit immediate" } */
}
m128d = _mm_cvt_roundss_sd (m128d, m128, 7); /* { dg-error "incorrect rounding operand" } */
m128 = _mm_getexp_round_ss (m128, m128, 7); /* { dg-error "incorrect rounding operand" } */
+ m128 = _mm_mask_getexp_round_ss (m128, mmask8, m128, m128, 7); /* { dg-error "incorrect rounding operand" } */
+ m128 = _mm_maskz_getexp_round_ss (mmask8, m128, m128, 7); /* { dg-error "incorrect rounding operand" } */
m128d = _mm_getexp_round_sd (m128d, m128d, 7); /* { dg-error "incorrect rounding operand" } */
+ m128d = _mm_mask_getexp_round_sd (m128d, mmask8, m128d, m128d, 7); /* { dg-error "incorrect rounding operand" } */
+ m128d = _mm_maskz_getexp_round_sd (mmask8, m128d, m128d, 7); /* { dg-error "incorrect rounding operand" } */
m512 = _mm512_getexp_round_ps (m512, 7); /* { dg-error "incorrect rounding operand" } */
m512 = _mm512_mask_getexp_round_ps (m512, mmask16, m512, 7); /* { dg-error "incorrect rounding operand" } */
m512 = _mm512_maskz_getexp_round_ps (mmask16, m512, 7); /* { dg-error "incorrect rounding operand" } */
m512 = _mm512_mask_getmant_round_ps (m512, mmask16, m512, 0, 0, 7); /* { dg-error "incorrect rounding operand" } */
m512 = _mm512_maskz_getmant_round_ps (mmask16, m512, 0, 0, 7); /* { dg-error "incorrect rounding operand" } */
m128d = _mm_getmant_round_sd (m128d, m128d, 0, 0, 7); /* { dg-error "incorrect rounding operand" } */
+ m128d = _mm_mask_getmant_round_sd (m128d, mmask8, m128d, m128d, 0, 0, 7); /* { dg-error "incorrect rounding operand" } */
+ m128d = _mm_maskz_getmant_round_sd (mmask8, m128d, m128d, 0, 0, 7); /* { dg-error "incorrect rounding operand" } */
m128 = _mm_getmant_round_ss (m128, m128, 0, 0, 7); /* { dg-error "incorrect rounding operand" } */
+ m128 = _mm_mask_getmant_round_ss (m128, mmask8, m128, m128, 0, 0, 7); /* { dg-error "incorrect rounding operand" } */
+ m128 = _mm_maskz_getmant_round_ss (mmask8, m128, m128, 0, 0, 7); /* { dg-error "incorrect rounding operand" } */
m512 = _mm512_roundscale_round_ps (m512, 4, 7); /* { dg-error "incorrect rounding operand" } */
m512 = _mm512_mask_roundscale_round_ps (m512, mmask16, m512, 4, 7); /* { dg-error "incorrect rounding operand" } */
m128d = _mm_cvt_roundss_sd (m128d, m128, 3); /* { dg-error "incorrect rounding operand" } */
m128 = _mm_getexp_round_ss (m128, m128, 3); /* { dg-error "incorrect rounding operand" } */
+ m128 = _mm_mask_getexp_round_ss (m128, mmask8, m128, m128, 3); /* { dg-error "incorrect rounding operand" } */
+ m128 = _mm_maskz_getexp_round_ss (mmask8, m128, m128, 3); /* { dg-error "incorrect rounding operand" } */
m128d = _mm_getexp_round_sd (m128d, m128d, 3); /* { dg-error "incorrect rounding operand" } */
+ m128d = _mm_mask_getexp_round_sd (m128d, mmask8, m128d, m128d, 5); /* { dg-error "incorrect rounding operand" } */
+ m128d = _mm_maskz_getexp_round_sd (mmask8, m128d, m128d, 5); /* { dg-error "incorrect rounding operand" } */
m512 = _mm512_getexp_round_ps (m512, 3); /* { dg-error "incorrect rounding operand" } */
m512 = _mm512_mask_getexp_round_ps (m512, mmask16, m512, 3); /* { dg-error "incorrect rounding operand" } */
m512 = _mm512_maskz_getexp_round_ps (mmask16, m512, 3); /* { dg-error "incorrect rounding operand" } */
projects / gcc.git / commitdiff