Skip to content

[libc] Enable intermediate computation in float for baremetal #163622

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Merged
merged 1 commit into from
Oct 15, 2025
Merged

[libc] Enable intermediate computation in float for baremetal #163622

merged 1 commit into from
Oct 15, 2025

Conversation

@petrhosek
Copy link
Member

This can improve performance on 32-bit baremetal targets.

@petrhosek
[libc] Enable intermediate computation in float for baremetal
fea4978 
This can improve performance on 32-bit baremetal targets.
@petrhosek petrhosek requested a review from Prabhuk October 15, 2025 19:58
@petrhosek petrhosek added the libc label Oct 15, 2025
@llvmbot
Copy link
Member

llvmbot commented Oct 15, 2025

@llvm/pr-subscribers-libc

Author: Petr Hosek (petrhosek)

Changes

This can improve performance on 32-bit baremetal targets.

Full diff: https://github.com/llvm/llvm-project/pull/163622.diff

1 Files Affected:

  • (modified) libc/config/baremetal/config.json (+1-1)
diff --git a/libc/config/baremetal/config.json b/libc/config/baremetal/config.json
index 796b1d8ed1398..ffb4fe6487fdc 100644
--- a/libc/config/baremetal/config.json
+++ b/libc/config/baremetal/config.json
@@ -36,7 +36,7 @@
   },
   "math": {
     "LIBC_CONF_MATH_OPTIMIZATIONS": {
-      "value": "(LIBC_MATH_SKIP_ACCURATE_PASS | LIBC_MATH_SMALL_TABLES)"
+      "value": "(LIBC_MATH_SKIP_ACCURATE_PASS | LIBC_MATH_SMALL_TABLES | LIBC_MATH_INTERMEDIATE_COMP_IN_FLOAT)"
     }
   },
   "general": {

Prabhuk
Prabhuk approved these changes Oct 15, 2025
View reviewed changes
Copy link
Contributor

@Prabhuk Prabhuk left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@petrhosek petrhosek merged commit 49f03ee into llvm:main Oct 15, 2025
22 checks passed
@llvm-ci
Copy link
Collaborator

llvm-ci commented Oct 15, 2025

LLVM Buildbot has detected a new failure on builder fuchsia-x86_64-linux running on fuchsia-debian-64-us-central1-b-1 while building libc at step 4 "annotate".

Full details are available at: https://lab.llvm.org/buildbot/#/builders/11/builds/26090

Here is the relevant piece of the build log for the reference 
Step 4 (annotate) failure: 'python ../llvm-zorg/zorg/buildbot/builders/annotated/fuchsia-linux.py ...' (failure)
...
[859/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.nextdownbf16.dir/nextdownbf16.cpp.obj
[860/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.ufromfpxl.dir/ufromfpxl.cpp.obj
[861/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.tanhf.dir/tanhf.cpp.obj
[862/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.atanhf.dir/atanhf.cpp.obj
[863/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.nextupbf16.dir/nextupbf16.cpp.obj
[864/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.fdimbf16.dir/fdimbf16.cpp.obj
[865/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.ffma.dir/ffma.cpp.obj
[866/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.ffmal.dir/ffmal.cpp.obj
[867/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.fmodbf16.dir/fmodbf16.cpp.obj
[868/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.atan2f.dir/atan2f.cpp.obj
FAILED: libc/src/math/generic/CMakeFiles/libc.src.math.generic.atan2f.dir/atan2f.cpp.obj 
/var/lib/buildbot/fuchsia-x86_64-linux/build/llvm-build-52kagzfa/./bin/clang++ --target=armv6m-none-eabi -DLIBC_NAMESPACE=__llvm_libc_22_0_0_git -I/var/lib/buildbot/fuchsia-x86_64-linux/llvm-project/libc -isystem /var/lib/buildbot/fuchsia-x86_64-linux/build/llvm-build-52kagzfa/include/armv6m-unknown-none-eabi --target=armv6m-none-eabi -Wno-atomic-alignment "-Dvfprintf(stream, format, vlist)=vprintf(format, vlist)" "-Dfprintf(stream, format, ...)=printf(format)" "-Dfputs(string, stream)=puts(string)" -D_LIBCPP_PRINT=1 -mthumb -fPIC -fno-semantic-interposition -fvisibility-inlines-hidden -Werror=date-time -Werror=unguarded-availability-new -Wall -Wextra -Wno-unused-parameter -Wwrite-strings -Wcast-qual -Wmissing-field-initializers -Wimplicit-fallthrough -Wcovered-switch-default -Wno-noexcept-type -Wnon-virtual-dtor -Wdelete-non-virtual-dtor -Wsuggest-override -Wstring-conversion -Wno-pass-failed -Wmisleading-indentation -Wctad-maybe-unsupported -ffunction-sections -fdata-sections -ffile-prefix-map=/var/lib/buildbot/fuchsia-x86_64-linux/build/llvm-build-52kagzfa/runtimes/runtimes-armv6m-none-eabi-bins=../../../../llvm-project -ffile-prefix-map=/var/lib/buildbot/fuchsia-x86_64-linux/llvm-project/= -no-canonical-prefixes -Os -DNDEBUG --target=armv6m-none-eabi -DLIBC_QSORT_IMPL=LIBC_QSORT_HEAP_SORT -DLIBC_TYPES_TIME_T_IS_32_BIT "-DLIBC_MATH=(LIBC_MATH_SKIP_ACCURATE_PASS | LIBC_MATH_SMALL_TABLES | LIBC_MATH_INTERMEDIATE_COMP_IN_FLOAT)" -DLIBC_ERRNO_MODE=LIBC_ERRNO_MODE_EXTERNAL -DLIBC_THREAD_MODE=LIBC_THREAD_MODE_SINGLE -DLIBC_TARGET_OS_IS_BAREMETAL -fpie -ffreestanding -DLIBC_FULL_BUILD -nostdlibinc -ffixed-point -fno-builtin -fno-exceptions -fno-lax-vector-conversions -fno-unwind-tables -fno-asynchronous-unwind-tables -fno-rtti -ftrivial-auto-var-init=pattern -fno-omit-frame-pointer -Wall -Wextra -Werror -Wconversion -Wno-sign-conversion -Wdeprecated -Wno-c99-extensions -Wno-gnu-imaginary-constant -Wno-pedantic -Wimplicit-fallthrough -Wwrite-strings -Wextra-semi -Wnewline-eof -Wnonportable-system-include-path -Wstrict-prototypes -Wthread-safety -Wglobal-constructors -DLIBC_COPT_PUBLIC_PACKAGING -MD -MT libc/src/math/generic/CMakeFiles/libc.src.math.generic.atan2f.dir/atan2f.cpp.obj -MF libc/src/math/generic/CMakeFiles/libc.src.math.generic.atan2f.dir/atan2f.cpp.obj.d -o libc/src/math/generic/CMakeFiles/libc.src.math.generic.atan2f.dir/atan2f.cpp.obj -c /var/lib/buildbot/fuchsia-x86_64-linux/llvm-project/libc/src/math/generic/atan2f.cpp
In file included from /var/lib/buildbot/fuchsia-x86_64-linux/llvm-project/libc/src/math/generic/atan2f.cpp:10:
In file included from /var/lib/buildbot/fuchsia-x86_64-linux/llvm-project/libc/src/__support/math/atan2f.h:16:
/var/lib/buildbot/fuchsia-x86_64-linux/llvm-project/libc/src/__support/FPUtil/double_double.h:83:13: error: implicit conversion loses floating-point precision: 'double' to 'float' [-Werror,-Wimplicit-float-conversion]
   83 |   r.hi = t1 + t2;
      |        ~ ~~~^~~~
/var/lib/buildbot/fuchsia-x86_64-linux/llvm-project/libc/src/__support/FPUtil/double_double.h:139:24: note: in instantiation of function template specialization '__llvm_libc_22_0_0_git::fputil::split<float, 12U>' requested here
  139 |     NumberPair<T> as = split(a);
      |                        ^
/var/lib/buildbot/fuchsia-x86_64-linux/llvm-project/libc/src/__support/math/atan2f_float.h:223:30: note: in instantiation of function template specialization '__llvm_libc_22_0_0_git::fputil::exact_mult<float, 12U>' requested here
  223 |   FloatFloat num_k = fputil::exact_mult(num, k);
      |                              ^
1 error generated.
[869/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.atanf.dir/atanf.cpp.obj
[870/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.totalordermag.dir/totalordermag.cpp.obj
[871/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.totalorder.dir/totalorder.cpp.obj
[872/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.totalorderf.dir/totalorderf.cpp.obj
[873/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.totalorderl.dir/totalorderl.cpp.obj
[874/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.totalordermagf.dir/totalordermagf.cpp.obj
[875/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.fmaf.dir/fmaf.cpp.obj
[876/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.totalordermagl.dir/totalordermagl.cpp.obj
[877/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.scalblnf.dir/scalblnf.cpp.obj
[878/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.remquobf16.dir/remquobf16.cpp.obj
[879/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.remainderbf16.dir/remainderbf16.cpp.obj
[880/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.acoshf.dir/acoshf.cpp.obj
[881/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.getpayload.dir/getpayload.cpp.obj
[882/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.scalbln.dir/scalbln.cpp.obj
[883/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.getpayloadl.dir/getpayloadl.cpp.obj
[884/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.getpayloadf.dir/getpayloadf.cpp.obj
[885/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.asinhf.dir/asinhf.cpp.obj
[886/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.scalbn.dir/scalbn.cpp.obj
[887/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.setpayloadf.dir/setpayloadf.cpp.obj
[888/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.cbrtf.dir/cbrtf.cpp.obj
[889/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.scalbnl.dir/scalbnl.cpp.obj
[890/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.scalblnl.dir/scalblnl.cpp.obj
[891/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.setpayload.dir/setpayload.cpp.obj
[892/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.setpayloadl.dir/setpayloadl.cpp.obj
[893/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.scalbnf.dir/scalbnf.cpp.obj
Step 6 (build) failure: build (failure)
...
[859/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.nextdownbf16.dir/nextdownbf16.cpp.obj
[860/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.ufromfpxl.dir/ufromfpxl.cpp.obj
[861/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.tanhf.dir/tanhf.cpp.obj
[862/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.atanhf.dir/atanhf.cpp.obj
[863/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.nextupbf16.dir/nextupbf16.cpp.obj
[864/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.fdimbf16.dir/fdimbf16.cpp.obj
[865/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.ffma.dir/ffma.cpp.obj
[866/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.ffmal.dir/ffmal.cpp.obj
[867/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.fmodbf16.dir/fmodbf16.cpp.obj
[868/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.atan2f.dir/atan2f.cpp.obj
FAILED: libc/src/math/generic/CMakeFiles/libc.src.math.generic.atan2f.dir/atan2f.cpp.obj 
/var/lib/buildbot/fuchsia-x86_64-linux/build/llvm-build-52kagzfa/./bin/clang++ --target=armv6m-none-eabi -DLIBC_NAMESPACE=__llvm_libc_22_0_0_git -I/var/lib/buildbot/fuchsia-x86_64-linux/llvm-project/libc -isystem /var/lib/buildbot/fuchsia-x86_64-linux/build/llvm-build-52kagzfa/include/armv6m-unknown-none-eabi --target=armv6m-none-eabi -Wno-atomic-alignment "-Dvfprintf(stream, format, vlist)=vprintf(format, vlist)" "-Dfprintf(stream, format, ...)=printf(format)" "-Dfputs(string, stream)=puts(string)" -D_LIBCPP_PRINT=1 -mthumb -fPIC -fno-semantic-interposition -fvisibility-inlines-hidden -Werror=date-time -Werror=unguarded-availability-new -Wall -Wextra -Wno-unused-parameter -Wwrite-strings -Wcast-qual -Wmissing-field-initializers -Wimplicit-fallthrough -Wcovered-switch-default -Wno-noexcept-type -Wnon-virtual-dtor -Wdelete-non-virtual-dtor -Wsuggest-override -Wstring-conversion -Wno-pass-failed -Wmisleading-indentation -Wctad-maybe-unsupported -ffunction-sections -fdata-sections -ffile-prefix-map=/var/lib/buildbot/fuchsia-x86_64-linux/build/llvm-build-52kagzfa/runtimes/runtimes-armv6m-none-eabi-bins=../../../../llvm-project -ffile-prefix-map=/var/lib/buildbot/fuchsia-x86_64-linux/llvm-project/= -no-canonical-prefixes -Os -DNDEBUG --target=armv6m-none-eabi -DLIBC_QSORT_IMPL=LIBC_QSORT_HEAP_SORT -DLIBC_TYPES_TIME_T_IS_32_BIT "-DLIBC_MATH=(LIBC_MATH_SKIP_ACCURATE_PASS | LIBC_MATH_SMALL_TABLES | LIBC_MATH_INTERMEDIATE_COMP_IN_FLOAT)" -DLIBC_ERRNO_MODE=LIBC_ERRNO_MODE_EXTERNAL -DLIBC_THREAD_MODE=LIBC_THREAD_MODE_SINGLE -DLIBC_TARGET_OS_IS_BAREMETAL -fpie -ffreestanding -DLIBC_FULL_BUILD -nostdlibinc -ffixed-point -fno-builtin -fno-exceptions -fno-lax-vector-conversions -fno-unwind-tables -fno-asynchronous-unwind-tables -fno-rtti -ftrivial-auto-var-init=pattern -fno-omit-frame-pointer -Wall -Wextra -Werror -Wconversion -Wno-sign-conversion -Wdeprecated -Wno-c99-extensions -Wno-gnu-imaginary-constant -Wno-pedantic -Wimplicit-fallthrough -Wwrite-strings -Wextra-semi -Wnewline-eof -Wnonportable-system-include-path -Wstrict-prototypes -Wthread-safety -Wglobal-constructors -DLIBC_COPT_PUBLIC_PACKAGING -MD -MT libc/src/math/generic/CMakeFiles/libc.src.math.generic.atan2f.dir/atan2f.cpp.obj -MF libc/src/math/generic/CMakeFiles/libc.src.math.generic.atan2f.dir/atan2f.cpp.obj.d -o libc/src/math/generic/CMakeFiles/libc.src.math.generic.atan2f.dir/atan2f.cpp.obj -c /var/lib/buildbot/fuchsia-x86_64-linux/llvm-project/libc/src/math/generic/atan2f.cpp
In file included from /var/lib/buildbot/fuchsia-x86_64-linux/llvm-project/libc/src/math/generic/atan2f.cpp:10:
In file included from /var/lib/buildbot/fuchsia-x86_64-linux/llvm-project/libc/src/__support/math/atan2f.h:16:
/var/lib/buildbot/fuchsia-x86_64-linux/llvm-project/libc/src/__support/FPUtil/double_double.h:83:13: error: implicit conversion loses floating-point precision: 'double' to 'float' [-Werror,-Wimplicit-float-conversion]
   83 |   r.hi = t1 + t2;
      |        ~ ~~~^~~~
/var/lib/buildbot/fuchsia-x86_64-linux/llvm-project/libc/src/__support/FPUtil/double_double.h:139:24: note: in instantiation of function template specialization '__llvm_libc_22_0_0_git::fputil::split<float, 12U>' requested here
  139 |     NumberPair<T> as = split(a);
      |                        ^
/var/lib/buildbot/fuchsia-x86_64-linux/llvm-project/libc/src/__support/math/atan2f_float.h:223:30: note: in instantiation of function template specialization '__llvm_libc_22_0_0_git::fputil::exact_mult<float, 12U>' requested here
  223 |   FloatFloat num_k = fputil::exact_mult(num, k);
      |                              ^
1 error generated.
[869/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.atanf.dir/atanf.cpp.obj
[870/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.totalordermag.dir/totalordermag.cpp.obj
[871/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.totalorder.dir/totalorder.cpp.obj
[872/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.totalorderf.dir/totalorderf.cpp.obj
[873/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.totalorderl.dir/totalorderl.cpp.obj
[874/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.totalordermagf.dir/totalordermagf.cpp.obj
[875/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.fmaf.dir/fmaf.cpp.obj
[876/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.totalordermagl.dir/totalordermagl.cpp.obj
[877/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.scalblnf.dir/scalblnf.cpp.obj
[878/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.remquobf16.dir/remquobf16.cpp.obj
[879/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.remainderbf16.dir/remainderbf16.cpp.obj
[880/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.acoshf.dir/acoshf.cpp.obj
[881/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.getpayload.dir/getpayload.cpp.obj
[882/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.scalbln.dir/scalbln.cpp.obj
[883/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.getpayloadl.dir/getpayloadl.cpp.obj
[884/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.getpayloadf.dir/getpayloadf.cpp.obj
[885/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.asinhf.dir/asinhf.cpp.obj
[886/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.scalbn.dir/scalbn.cpp.obj
[887/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.setpayloadf.dir/setpayloadf.cpp.obj
[888/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.cbrtf.dir/cbrtf.cpp.obj
[889/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.scalbnl.dir/scalbnl.cpp.obj
[890/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.scalblnl.dir/scalblnl.cpp.obj
[891/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.setpayload.dir/setpayload.cpp.obj
[892/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.setpayloadl.dir/setpayloadl.cpp.obj
[893/2598] Building CXX object libc/src/math/generic/CMakeFiles/libc.src.math.generic.scalbnf.dir/scalbnf.cpp.obj

@petrhosek petrhosek mentioned this pull request Oct 16, 2025
Closed
petrhosek added a commit that referenced this pull request Oct 16, 2025
@petrhosek
Revert "[libc] Enable intermediate computation in float for baremetal (…
0181877 
…#163622)"

This reverts commit 49f03ee.
@petrhosek petrhosek mentioned this pull request Oct 16, 2025
Merged
petrhosek added a commit that referenced this pull request Oct 16, 2025
@petrhosek
Revert "[libc] Enable intermediate computation in float for baremetal" (
d50423e 
#163712)

Reverts #163622, see #163711 for details.
llvm-sync bot pushed a commit to arm/arm-toolchain that referenced this pull request Oct 16, 2025
@petrhosek @github-actions
Automerge: Revert "[libc] Enable intermediate computation in float fo…
36e2e42 
…r baremetal" (#163712)

Reverts llvm/llvm-project#163622, see #163711 for details.
@petrhosek petrhosek deleted the libc-baremetal-intermediate-math-in-float branch October 16, 2025 17:30
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@Prabhuk Prabhuk Prabhuk approved these changes

Assignees

No one assigned

Labels

libc

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

4 participants

@petrhosek @llvmbot @llvm-ci @Prabhuk