167 Commits

Author SHA1 Message Date
Fraser Cormack
d12a8da1de
[libclc] Move min/max/clamp into the CLC builtins library (#114386)
These functions are "shared" between integer and floating-point types,
hence the directory name. They are used in several CLC internal
functions such as __clc_ldexp.

Note that clspv and spirv targets don't want to define these functions,
so pre-processor macros replace calls to __clc_min with regular min, for
example. This means they can use as much of the generic CLC source files
as possible, but where CLC functions would usually call out to an
external __clc_min symbol, they call out to an external min symbol. Then
they opt out of defining __clc_min itself in their CLC builtins library.

Preprocessor definitions for these targets have also been changed
somewhat: what used to be CLC_SPIRV (the 32-bit target) is now
CLC_SPIRV32, and CLC_SPIRV now represents either CLC_SPIRV32 or
CLC_SPIRV64. Same goes for CLC_CLSPV.

There are no differences (measured with llvm-diff) in any of the final
builtins libraries for nvptx, amdgpu, or clspv. Neither are there
differences in the SPIR-V targets' LLVM IR before it's actually lowered
to SPIR-V.
2024-10-31 16:45:37 +00:00
Jan Vesely
4b23a2e8e9 libclc: Move rsqrt implementation to a .cl file
Reviewer: awatry
Differential Revision: https://reviews.llvm.org/D74013
2020-02-09 14:42:09 -05:00
Jan Vesely
4a725996e5 sincos: Simplify declaration headers.
This follows the same pattern as modf and fract.
Reviewer: Aaron Watry
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>

llvm-svn: 356028
2019-03-13 07:13:34 +00:00
Jan Vesely
e7c0c37a31 fdim: Use binary_decl_tt.inc instead of custom inc file.
Reviewer: Aaron Watry
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 356027
2019-03-13 07:13:32 +00:00
Jan Vesely
5b0600c277 nextafter: Use binary_decl_tt.inc instead of custom inc file.
Reviewer: Aaron Watry
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 356026
2019-03-13 07:13:30 +00:00
Jan Vesely
e438b58cd0 copysign: Use binary_decl_tt.inc instead of custom inc file.
Reviewer: Aaron Watry
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 356025
2019-03-13 07:13:28 +00:00
Jan Vesely
81bc9ee81c atan2pi: Use binary_decl_tt.inc instead of custom inc file.
Reviewer: Aaron Watry
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 356024
2019-03-13 07:13:26 +00:00
Jan Vesely
9526e02021 atan2: Use binary_decl_tt.inc instead of custom inc file.
Reviewer: Aaron Watry
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 356023
2019-03-13 07:13:24 +00:00
Jan Vesely
8985c9c212 hypot: Use binary_decl_tt.inc instead of custom inc file
Reviewer: Aaron Watry
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 356022
2019-03-13 07:13:22 +00:00
Jan Vesely
5b136ca125 Move unary_instrinsic.inc to private headers.
Reviewer: Aaron Watry
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 356021
2019-03-13 07:06:19 +00:00
Jan Vesely
2aa333f3d1 Move binary_intrinsic.h to private headers.
Reviewer: Aaron Watry
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 356020
2019-03-13 07:06:15 +00:00
Jan Vesely
1f4a8a9158 Move ternary_intrinsic.h to private headers.
Reviewer: Aaron Watry
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 356019
2019-03-13 07:06:13 +00:00
Jan Vesely
ee555aa992 trunc: Remove llvm intrinsic from the header.
Reviewer: Aaron Watry
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 356018
2019-03-13 07:06:10 +00:00
Jan Vesely
1c395b74bf round: Remove llvm intrinsic from the header
Reviewer: Aaron Watry
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 356017
2019-03-13 07:06:08 +00:00
Jan Vesely
b3d64e4a83 rint: Remove llvm intrinsic from the header.
Reviewer: Aaron Watry
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 356016
2019-03-13 07:06:06 +00:00
Jan Vesely
fd199f0139 floor: Remove llvm isntrinsic from the header.
Reviewer: Aaron Watry
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 356015
2019-03-13 07:06:03 +00:00
Jan Vesely
fda15e56a6 fabs: Remove llvm intrinsic from the header.
Reviewer: Aaron Watry
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 356014
2019-03-13 07:06:00 +00:00
Jan Vesely
54eb4d3a6d ceil: Remove llvm intrinsic from the header.
Reviewer: Aaron Watry
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 356013
2019-03-13 07:05:58 +00:00
Jan Vesely
4b0b9a727e mad: Convert to standard ternary header
Reviewer: Aaron Watry
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 356011
2019-03-13 07:05:53 +00:00
Jan Vesely
6e85e6309d math/fma: Add fp32 software implementation
Passes CTS on carrizo (when forced to use sw fma) and turks.
Reviewer: Tom Stellard <tstellar@redhat.com>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>

llvm-svn: 334226
2018-06-07 20:27:43 +00:00
Jan Vesely
70a270da5f Add initial support for half precision builtins
v2: fix fmax implementation
    use consistent checks for __CLC_FP_SIZE
    add missing TODOs
    fix whitespace in definitions.h
v3: undef ZERO in modf.inc

Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
reviewer: Jeroen Ketema <j.ketema@xs4all.nl>
Reviewed-by: Aaron Watry <awatry@gmail.com>
Tested-by: Aaron Watry <awatry@gmail.com>
llvm-svn: 332677
2018-05-17 22:55:30 +00:00
Jan Vesely
8db45e4cf1 remquo: Port from amd builtins
double version passes on carrizo. float version fails on denormals.

Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
Reviewer: Aaron Watry <awatry@gmail.com>
llvm-svn: 331434
2018-05-03 05:44:26 +00:00
Jan Vesely
93af966747 fmod: Port from amd_builtins
Uses only denormal path for fp32.
Passes CTS on carrizo and turks.

v2: whitespace fix

Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
Reviewer: Aaron Watry <awatry@gmail.com>
llvm-svn: 329433
2018-04-06 17:43:08 +00:00
Jan Vesely
5b10494fa8 remainder: Port from amd builtins
Mostly ported from amd_builtins, uses only denormal path for fp32.
Passes CTS on carrizo and turks

Reviewer: Aaron Watry <awatry@gmail.com>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 327818
2018-03-19 01:01:10 +00:00
Jan Vesely
b672f7a251 nan: Implement
Passes CTS on carrizo and turks

Reviewer: Aaron Watry <awatry@gmail.com>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 327324
2018-03-12 19:46:52 +00:00
Jan Vesely
dbaf6d0f7c Move cl_khr_fp64 exntension enablement to gentype include lists
This will make adding cl_khr_fp16 support easier

Reviewed-by: Aaron Watry <awatry@gmail.com>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 326816
2018-03-06 17:48:35 +00:00
Jan Vesely
3b8b4eb64d half_powr: Implement using powr
v2: Use full precision implementation

Reviewer: Jeroen Ketema <j.ketema@xs4all.nl>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 323942
2018-02-01 03:00:35 +00:00
Jan Vesely
7013857f95 tanpi: Port from amd_builtins
Passes piglit on turks and carrizo.
Passes CTS on carrizo.

Acked-By: Aaron Watry <awatry@gmail.com>
Tested-By: Aaron Watry <awatry@gmail.com>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 322980
2018-01-19 18:57:22 +00:00
Jan Vesely
44e0522c09 half_divide: Implement using x/y
Passes CTS on carrizo
v2: Use full precision implementation

Reviewer: Jeroen Ketema <j.ketema@xs4all.nl>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 322899
2018-01-18 21:12:06 +00:00
Jan Vesely
2813b4f8d9 half_tan: Implement using tan
v2: Use full precision implementation

Reviewer: Jeroen Ketema <j.ketema@xs4all.nl>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 322898
2018-01-18 21:12:04 +00:00
Jan Vesely
bf38fae8de half_sin: Implement using sin
Passes CTS on carrizo
v2: Use full precision implementation

Reviewer: Jeroen Ketema <j.ketema@xs4all.nl>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 322897
2018-01-18 21:12:01 +00:00
Jan Vesely
398108b91e half_recip: Implement using 1/x
Passes CTS on carrizo
v2: Use full precision implementation

Reviewer: Jeroen Ketema <j.ketema@xs4all.nl>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 322896
2018-01-18 21:11:58 +00:00
Jan Vesely
a1aba44ffa half_log2: Implement using log2
Passes CTS on carrizo
v2: Use full precision implementation

Reviewer: Jeroen Ketema <j.ketema@xs4all.nl>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 322895
2018-01-18 21:11:56 +00:00
Jan Vesely
b3b72af4b9 half_log10: Implement using log10
Passes CTS on carrizo
v2: Use full precision implementation

Reviewer: Jeroen Ketema <j.ketema@xs4all.nl>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 322894
2018-01-18 21:11:53 +00:00
Jan Vesely
6852023802 half_log: Implement using log
Passes CTS on carrizo
v2: Use full precision implementation

Reviewer: Jeroen Ketema <j.ketema@xs4all.nl>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 322893
2018-01-18 21:11:50 +00:00
Jan Vesely
aa4c3899b5 half_exp10: Implement using exp10
Passes CTS on carrizo
v2: Use full precision implementation

Reviewer: Jeroen Ketema <j.ketema@xs4all.nl>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 322892
2018-01-18 21:11:48 +00:00
Jan Vesely
3c0e19b61a half_exp2: Implement using exp2
Passes CTS on carrizo
v2: Use full precision implementation

Reviewer: Jeroen Ketema <j.ketema@xs4all.nl>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 322891
2018-01-18 21:11:45 +00:00
Jan Vesely
caa9000b1c half_exp: Implement using exp
Passes CTS on carrizo
v2: Use full precision implementation

Reviewer: Jeroen Ketema <j.ketema@xs4all.nl>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 322890
2018-01-18 21:11:43 +00:00
Jan Vesely
b5d556061d half_cos: Implement using cos
Passes CTS on carrizo
v2: Use full precision implementation

Reviewer: Jeroen Ketema <j.ketema@xs4all.nl>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 322889
2018-01-18 21:11:40 +00:00
Jan Vesely
e53ae3b596 half_sqrt: Cleanup implementation
Passes CTS on carrizo
v2: Use full precision implementation

Reviewer: Jeroen Ketema <j.ketema@xs4all.nl>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 322888
2018-01-18 21:11:38 +00:00
Jan Vesely
a95db14461 half_rsqrt: Cleanup implementation
Passes CTS on carrizo
v2: Use full precision implementation

Reviewer: Jeroen Ketema <j.ketema@xs4all.nl>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 322887
2018-01-18 21:11:35 +00:00
Jan Vesely
fe8e00bc3c rootn: Port from amd_builtins
Passes piglit on turks and carrizo
fp64 passes ctx on carrizo

v2: fix formatting
    check fp32 denormal support at runtime

Reviewer: Jeroen Ketema <j.ketema@xs4all.nl>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 322763
2018-01-17 21:22:14 +00:00
Jan Vesely
c45ec604f5 powr: Port from amd_builtins
Passes piglit on turks and carrizo
fp64 passes cts on carrizo

v2: fix formatting
    check fp32 denormal support at runtime

Reviewer: Jeroen Ketema <j.ketema@xs4all.nl>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 322762
2018-01-17 21:22:06 +00:00
Jan Vesely
5efc8fe321 pown: Port from amd_builtins
Passes piglit on turks and carrizo
fp64 passes CTS on carrizo

v2: fix formatting
    check fp32 denormal support at runtime

Reviewer: Jeroen Ketema <j.ketema@xs4all.nl>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 322761
2018-01-17 21:22:03 +00:00
Jan Vesely
cc5c65b2c2 pow: Port from amd_builtins
Passes piglit on turks and carrizo
fp64 passes CTS on carrizo

v2: fix formatting
    check fp32 denormal support at runtime

Reviewer: Jeroen Ketema <j.ketema@xs4all.nl>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 322760
2018-01-17 21:21:35 +00:00
Jan Vesely
fe7c045753 math: Implement minmag
Reviewer: Aaron Watry
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 318265
2017-11-15 04:10:39 +00:00
Jan Vesely
7ba243cc3d math: Implement maxmag
Reviewer: Aaron Watry
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 318264
2017-11-15 04:10:37 +00:00
Jan Vesely
383fbd050c native_powr: Switch implementation to native_exp2 and native_log2
v2: don't use assume
    check only for x<0, the other conditions are handled transparently
v3: don't check inputs at all, nan propagation works as expected

Reviewer: Jeroen Ketema
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 318204
2017-11-14 21:55:41 +00:00
Jan Vesely
f38b40daf7 native_divide: provide function implementation instead of macro
Reviewer: Jeroen Ketema
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 318067
2017-11-13 18:28:56 +00:00
Jan Vesely
1b9825f982 native_recip: provide function implementation instead of macro
Reviewer: Jeroen Ketema
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 318066
2017-11-13 18:28:53 +00:00