Joe Ramsay 5bc100bd4b AArch64: Improve codegen in users of AdvSIMD log1pf helper ... log1pf is quite register-intensive - use fewer registers for the polynomial, and make various changes to shorten dependency chains in parent routines. There is now no spilling with GCC 14. Accuracy moves around a little - comments adjusted accordingly but does not require regen-ulps. Use the helper in log1pf as well, instead of having separate implementations. The more accurate polynomial means special-casing can be simplified, and the shorter dependency chain avoids the usual dance around v0, which is otherwise difficult. There is a small duplication of vectors containing 1.0f (or 0x3f800000) - GCC is not currently able to efficiently handle values which fit in FMOV but not MOVI, and are reinterpreted to integer. There may be potential for more optimisation if this is fixed. Reviewed-by: Wilco Dijkstra <Wilco.Dijkstra@arm.com>		2024-09-23 15:44:07 +01:00
..
bits	AArch64: Add vector logp1 alias for log1p	2024-09-19 17:53:34 +01:00
finclude	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
scripts	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
acos_advsimd.c	aarch64/fpu: Sync libmvec routines from 2.39 and before with AOR	2024-02-26 09:45:50 -03:00
acos_sve.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
acosf_advsimd.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
acosf_sve.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
acosh_advsimd.c	aarch64/fpu: Add vector variants of acosh	2024-04-04 10:32:58 +01:00
acosh_sve.c	aarch64/fpu: Add vector variants of acosh	2024-04-04 10:32:58 +01:00
acoshf_advsimd.c	AArch64: Improve codegen in users of AdvSIMD log1pf helper	2024-09-23 15:44:07 +01:00
acoshf_sve.c	aarch64/fpu: Add vector variants of acosh	2024-04-04 10:32:58 +01:00
advsimd_f32_protos.h	AArch64: Add vector logp1 alias for log1p	2024-09-19 17:53:34 +01:00
asin_advsimd.c	aarch64/fpu: Sync libmvec routines from 2.39 and before with AOR	2024-02-26 09:45:50 -03:00
asin_sve.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
asinf_advsimd.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
asinf_sve.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
asinh_advsimd.c	aarch64: Fix AdvSIMD libmvec routines for big-endian	2024-05-14 13:10:33 +01:00
asinh_sve.c	aarch64/fpu: Add vector variants of asinh	2024-04-04 10:33:02 +01:00
asinhf_advsimd.c	AArch64: Improve codegen in users of AdvSIMD log1pf helper	2024-09-23 15:44:07 +01:00
asinhf_sve.c	aarch64/fpu: Add vector variants of asinh	2024-04-04 10:33:02 +01:00
atan2_advsimd.c	aarch64/fpu: Add vector variants of pow	2024-05-21 14:38:49 +01:00
atan2_sve.c	aarch64/fpu: Add vector variants of pow	2024-05-21 14:38:49 +01:00
atan2f_advsimd.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
atan2f_sve.c	aarch64/fpu: Sync libmvec routines from 2.39 and before with AOR	2024-02-26 09:45:50 -03:00
atan_advsimd.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
atan_sve.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
atanf_advsimd.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
atanf_sve.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
atanh_advsimd.c	aarch64/fpu: Add vector variants of atanh	2024-04-04 10:33:12 +01:00
atanh_sve.c	aarch64/fpu: Add vector variants of atanh	2024-04-04 10:33:12 +01:00
atanhf_advsimd.c	AArch64: Improve codegen in users of AdvSIMD log1pf helper	2024-09-23 15:44:07 +01:00
atanhf_sve.c	aarch64/fpu: Add vector variants of atanh	2024-04-04 10:33:12 +01:00
bench-libmvec-arch.h	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
cbrt_advsimd.c	aarch64/fpu: Add vector variants of cbrt	2024-05-16 14:35:06 +01:00
cbrt_sve.c	aarch64/fpu: Add vector variants of cbrt	2024-05-16 14:35:06 +01:00
cbrtf_advsimd.c	aarch64/fpu: Add vector variants of cbrt	2024-05-16 14:35:06 +01:00
cbrtf_sve.c	aarch64/fpu: Add vector variants of cbrt	2024-05-16 14:35:06 +01:00
cos_advsimd.c	aarch64/fpu: Sync libmvec routines from 2.39 and before with AOR	2024-02-26 09:45:50 -03:00
cos_sve.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
cosf_advsimd.c	aarch64/fpu: Sync libmvec routines from 2.39 and before with AOR	2024-02-26 09:45:50 -03:00
cosf_sve.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
cosh_advsimd.c	aarch64: Fix AdvSIMD libmvec routines for big-endian	2024-05-14 13:10:33 +01:00
cosh_sve.c	aarch64/fpu: Add vector variants of cosh	2024-04-04 10:32:52 +01:00
coshf_advsimd.c	aarch64/fpu: Add vector variants of cosh	2024-04-04 10:32:52 +01:00
coshf_sve.c	AArch64: Improve codegen in SVE expf & related routines	2024-09-23 15:44:07 +01:00
erf_advsimd.c	aarch64: Fix AdvSIMD libmvec routines for big-endian	2024-05-14 13:10:33 +01:00
erf_data.c	aarch64/fpu: Add vector variants of erf	2024-04-04 10:32:48 +01:00
erf_sve.c	aarch64/fpu: Add vector variants of erf	2024-04-04 10:32:48 +01:00
erfc_advsimd.c	aarch64: Fix AdvSIMD libmvec routines for big-endian	2024-05-14 13:10:33 +01:00
erfc_data.c	aarch64/fpu: Add vector variants of erfc	2024-04-04 10:33:24 +01:00
erfc_sve.c	aarch64/fpu: Add vector variants of erfc	2024-04-04 10:33:24 +01:00
erfcf_advsimd.c	aarch64: Fix AdvSIMD libmvec routines for big-endian	2024-05-14 13:10:33 +01:00
erfcf_data.c	aarch64/fpu: Add vector variants of erfc	2024-04-04 10:33:24 +01:00
erfcf_sve.c	aarch64/fpu: Add vector variants of erfc	2024-04-04 10:33:24 +01:00
erff_advsimd.c	aarch64: Fix AdvSIMD libmvec routines for big-endian	2024-05-14 13:10:33 +01:00
erff_data.c	aarch64/fpu: Add vector variants of erf	2024-04-04 10:32:48 +01:00
erff_sve.c	aarch64/fpu: Add vector variants of erf	2024-04-04 10:32:48 +01:00
exp2_advsimd.c	aarch64/fpu: Sync libmvec routines from 2.39 and before with AOR	2024-02-26 09:45:50 -03:00
exp2_sve.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
exp2f_advsimd.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
exp2f_sve.c	AArch64: Improve codegen in SVE expf & related routines	2024-09-23 15:44:07 +01:00
exp10_advsimd.c	aarch64/fpu: Sync libmvec routines from 2.39 and before with AOR	2024-02-26 09:45:50 -03:00
exp10_sve.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
exp10f_advsimd.c	aarch64: Fix AdvSIMD libmvec routines for big-endian	2024-05-14 13:10:33 +01:00
exp10f_sve.c	AArch64: Improve codegen in SVE expf & related routines	2024-09-23 15:44:07 +01:00
exp_advsimd.c	aarch64/fpu: Sync libmvec routines from 2.39 and before with AOR	2024-02-26 09:45:50 -03:00
exp_sve.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
expf_advsimd.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
expf_sve.c	AArch64: Improve codegen in SVE expf & related routines	2024-09-23 15:44:07 +01:00
expm1_advsimd.c	aarch64: Fix AdvSIMD libmvec routines for big-endian	2024-05-14 13:10:33 +01:00
expm1_sve.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
expm1f_advsimd.c	aarch64: Fix AdvSIMD libmvec routines for big-endian	2024-05-14 13:10:33 +01:00
expm1f_sve.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
fclrexcpt.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
fedisblxcpt.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
feenablxcpt.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
fegetenv.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
fegetexcept.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
fegetmode.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
fegetround.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
feholdexcpt.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
fenv_private.h	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
fesetenv.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
fesetexcept.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
fesetmode.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
fesetround.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
feupdateenv.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
fgetexcptflg.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
finite_pow.h	aarch64/fpu: Add vector variants of pow	2024-05-21 14:38:49 +01:00
fpu_control.h	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
fraiseexcpt.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
fsetexcptflg.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
ftestexcept.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
get-rounding-mode.h	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
hypot_advsimd.c	aarch64/fpu: Add vector variants of hypot	2024-05-16 14:34:43 +01:00
hypot_sve.c	aarch64/fpu: Add vector variants of hypot	2024-05-16 14:34:43 +01:00
hypotf_advsimd.c	aarch64/fpu: Add vector variants of hypot	2024-05-16 14:34:43 +01:00
hypotf_sve.c	aarch64/fpu: Add vector variants of hypot	2024-05-16 14:34:43 +01:00
log1p_advsimd.c	AArch64: Add vector logp1 alias for log1p	2024-09-19 17:53:34 +01:00
log1p_sve.c	AArch64: Add vector logp1 alias for log1p	2024-09-19 17:53:34 +01:00
log1pf_advsimd.c	AArch64: Improve codegen in users of AdvSIMD log1pf helper	2024-09-23 15:44:07 +01:00
log1pf_sve.c	AArch64: Add vector logp1 alias for log1p	2024-09-19 17:53:34 +01:00
log2_advsimd.c	aarch64: Fix AdvSIMD libmvec routines for big-endian	2024-05-14 13:10:33 +01:00
log2_sve.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
log2f_advsimd.c	aarch64: Avoid redundant MOVs in AdvSIMD F32 logs	2024-09-09 13:03:49 +01:00
log2f_sve.c	AArch64: Improve codegen in SVE F32 logs	2024-09-23 15:44:07 +01:00
log10_advsimd.c	aarch64: Fix AdvSIMD libmvec routines for big-endian	2024-05-14 13:10:33 +01:00
log10_sve.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
log10f_advsimd.c	aarch64: Avoid redundant MOVs in AdvSIMD F32 logs	2024-09-09 13:03:49 +01:00
log10f_sve.c	AArch64: Improve codegen in SVE F32 logs	2024-09-23 15:44:07 +01:00
log_advsimd.c	aarch64: Fix AdvSIMD libmvec routines for big-endian	2024-05-14 13:10:33 +01:00
log_sve.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
logf_advsimd.c	aarch64: Avoid redundant MOVs in AdvSIMD F32 logs	2024-09-09 13:03:49 +01:00
logf_sve.c	AArch64: Improve codegen in SVE F32 logs	2024-09-23 15:44:07 +01:00
Makefile	aarch64/fpu: Add vector variants of pow	2024-05-21 14:38:49 +01:00
math_private.h	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
math-barriers.h	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
math-tests-arch.h	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
math-use-builtins-ceil.h	math: Decompose math-use-builtins.h	2020-06-22 11:09:45 -03:00
math-use-builtins-floor.h	math: Decompose math-use-builtins.h	2020-06-22 11:09:45 -03:00
math-use-builtins-fma.h	math: Decompose math-use-builtins.h	2020-06-22 11:09:45 -03:00
math-use-builtins-fmax.h	aarch64: Add math-use-builtins-f{max,min}.h	2021-12-13 10:08:46 -03:00
math-use-builtins-fmin.h	aarch64: Add math-use-builtins-f{max,min}.h	2021-12-13 10:08:46 -03:00
math-use-builtins-nearbyint.h	math: Decompose math-use-builtins.h	2020-06-22 11:09:45 -03:00
math-use-builtins-rint.h	math: Decompose math-use-builtins.h	2020-06-22 11:09:45 -03:00
math-use-builtins-round.h	math: Decompose math-use-builtins.h	2020-06-22 11:09:45 -03:00
math-use-builtins-sqrt.h	math: Decompose math-use-builtins.h	2020-06-22 11:09:45 -03:00
math-use-builtins-trunc.h	math: Decompose math-use-builtins.h	2020-06-22 11:09:45 -03:00
poly_advsimd_f32.h	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
poly_advsimd_f64.h	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
poly_generic.h	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
poly_sve_f32.h	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
poly_sve_f64.h	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
poly_sve_generic.h	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
pow_advsimd.c	aarch64/fpu: Add vector variants of pow	2024-05-21 14:38:49 +01:00
pow_sve.c	aarch64/fpu: Add vector variants of pow	2024-05-21 14:38:49 +01:00
powf_advsimd.c	aarch64/fpu: Add vector variants of pow	2024-05-21 14:38:49 +01:00
powf_sve.c	aarch64/fpu: Add vector variants of pow	2024-05-21 14:38:49 +01:00
s_llrint.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
s_llrintf.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
s_llround.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
s_llroundf.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
s_lrint.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
s_lrintf.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
s_lround.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
s_lroundf.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
s_roundeven.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
s_roundevenf.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
sin_advsimd.c	aarch64/fpu: Sync libmvec routines from 2.39 and before with AOR	2024-02-26 09:45:50 -03:00
sin_sve.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
sinf_advsimd.c	aarch64/fpu: Sync libmvec routines from 2.39 and before with AOR	2024-02-26 09:45:50 -03:00
sinf_sve.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
sinh_advsimd.c	aarch64: Fix AdvSIMD libmvec routines for big-endian	2024-05-14 13:10:33 +01:00
sinh_sve.c	aarch64/fpu: Add vector variants of sinh	2024-04-04 10:33:16 +01:00
sinhf_advsimd.c	aarch64/fpu: Add vector variants of sinh	2024-04-04 10:33:16 +01:00
sinhf_sve.c	aarch64/fpu: Add vector variants of sinh	2024-04-04 10:33:16 +01:00
sv_erf_data.c	aarch64/fpu: Add vector variants of erf	2024-04-04 10:32:48 +01:00
sv_erff_data.c	aarch64/fpu: Add vector variants of erf	2024-04-04 10:32:48 +01:00
sv_expf_inline.h	AArch64: Improve codegen in SVE expf & related routines	2024-09-23 15:44:07 +01:00
sv_expm1f_inline.h	aarch64/fpu: Add vector variants of sinh	2024-04-04 10:33:16 +01:00
sv_log1p_inline.h	aarch64/fpu: Add vector variants of acosh	2024-04-04 10:32:58 +01:00
sv_log1pf_inline.h	aarch64/fpu: Add vector variants of acosh	2024-04-04 10:32:58 +01:00
sv_math.h	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
tan_advsimd.c	aarch64: Fix AdvSIMD libmvec routines for big-endian	2024-05-14 13:10:33 +01:00
tan_sve.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
tanf_advsimd.c	aarch64: Fix AdvSIMD libmvec routines for big-endian	2024-05-14 13:10:33 +01:00
tanf_sve.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
tanh_advsimd.c	aarch64/fpu: Add vector variants of tanh	2024-04-04 10:33:20 +01:00
tanh_sve.c	aarch64/fpu: Add vector variants of tanh	2024-04-04 10:33:20 +01:00
tanhf_advsimd.c	aarch64/fpu: Add vector variants of tanh	2024-04-04 10:33:20 +01:00
tanhf_sve.c	aarch64/fpu: Add vector variants of tanh	2024-04-04 10:33:20 +01:00
test-double-advsimd-wrappers.c	aarch64/fpu: Add vector variants of pow	2024-05-21 14:38:49 +01:00
test-double-advsimd.h	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
test-double-sve-wrappers.c	aarch64/fpu: Add vector variants of pow	2024-05-21 14:38:49 +01:00
test-double-sve.h	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
test-float-advsimd-wrappers.c	aarch64/fpu: Add vector variants of pow	2024-05-21 14:38:49 +01:00
test-float-advsimd.h	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
test-float-sve-wrappers.c	aarch64/fpu: Add vector variants of pow	2024-05-21 14:38:49 +01:00
test-float-sve.h	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
test-vpcs-vector-wrapper.h	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
v_exp_data.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
v_exp_tail_data.c	aarch64/fpu: Add vector variants of cosh	2024-04-04 10:32:52 +01:00
v_expf_inline.h	aarch64: Fix AdvSIMD libmvec routines for big-endian	2024-05-14 13:10:33 +01:00
v_expm1f_inline.h	aarch64: Fix AdvSIMD libmvec routines for big-endian	2024-05-14 13:10:33 +01:00
v_log1p_inline.h	aarch64/fpu: Add vector variants of acosh	2024-04-04 10:32:58 +01:00
v_log1pf_inline.h	AArch64: Improve codegen in users of AdvSIMD log1pf helper	2024-09-23 15:44:07 +01:00
v_log2_data.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
v_log10_data.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
v_log_data.c	Update copyright dates with scripts/update-copyrights	2024-01-01 10:53:40 -08:00
v_math.h	aarch64/fpu: Add vector variants of acosh	2024-04-04 10:32:58 +01:00
v_pow_exp_data.c	aarch64/fpu: Add vector variants of pow	2024-05-21 14:38:49 +01:00
v_pow_log_data.c	aarch64/fpu: Add vector variants of pow	2024-05-21 14:38:49 +01:00
v_powf_data.c	aarch64/fpu: Add vector variants of pow	2024-05-21 14:38:49 +01:00
vecmath_config.h	aarch64/fpu: Add vector variants of pow	2024-05-21 14:38:49 +01:00
Versions	AArch64: Add vector logp1 alias for log1p	2024-09-19 17:53:34 +01:00