[libc] Add support for powi as an LLVM libc extension on the GPU
#98236
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
@llvm/pr-subscribers-backend-amdgpu @llvm/pr-subscribers-libc Author: Joseph Huber (jhuber6) ChangesSummary: Full diff: https://github.com/llvm/llvm-project/pull/98236.diff 14 Files Affected:
diff --git a/libc/config/gpu/entrypoints.txt b/libc/config/gpu/entrypoints.txt
index 62f3f0df247cc..166144d634344 100644
--- a/libc/config/gpu/entrypoints.txt
+++ b/libc/config/gpu/entrypoints.txt
@@ -313,6 +313,8 @@ set(TARGET_LIBM_ENTRYPOINTS
libc.src.math.nexttowardf
libc.src.math.pow
libc.src.math.powf
+ libc.src.math.powi
+ libc.src.math.powif
libc.src.math.remainder
libc.src.math.remainderf
libc.src.math.remquo
diff --git a/libc/docs/math/index.rst b/libc/docs/math/index.rst
index 7914a3d7e6d1a..83daefe785380 100644
--- a/libc/docs/math/index.rst
+++ b/libc/docs/math/index.rst
@@ -320,6 +320,8 @@ Higher Math Functions
+-----------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+
| pow | |check| | | | | | 7.12.7.5 | F.10.4.5 |
+-----------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+
+| powi | |check|\* | |check|\* | | | | | |
++-----------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+
| pown | | | | | | 7.12.7.6 | F.10.4.6 |
+-----------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+
| powr | | | | | | 7.12.7.7 | F.10.4.7 |
diff --git a/libc/spec/llvm_libc_ext.td b/libc/spec/llvm_libc_ext.td
index c0374cb631191..86215029831ca 100644
--- a/libc/spec/llvm_libc_ext.td
+++ b/libc/spec/llvm_libc_ext.td
@@ -76,6 +76,9 @@ def LLVMLibcExt : StandardSpec<"llvm_libc_ext"> {
GuardedFunctionSpec<"f16sqrt", RetValSpec<Float16Type>, [ArgSpec<DoubleType>], "LIBC_TYPES_HAS_FLOAT16">,
GuardedFunctionSpec<"f16sqrtf", RetValSpec<Float16Type>, [ArgSpec<FloatType>], "LIBC_TYPES_HAS_FLOAT16">,
GuardedFunctionSpec<"f16sqrtl", RetValSpec<Float16Type>, [ArgSpec<LongDoubleType>], "LIBC_TYPES_HAS_FLOAT16">,
+
+ FunctionSpec<"powi", RetValSpec<DoubleType>, [ArgSpec<DoubleType>, ArgSpec<IntType>]>,
+ FunctionSpec<"powif", RetValSpec<FloatType>, [ArgSpec<FloatType>, ArgSpec<IntType>]>,
]
>;
diff --git a/libc/src/math/CMakeLists.txt b/libc/src/math/CMakeLists.txt
index 74c2e4efda617..3a13ccbedec5a 100644
--- a/libc/src/math/CMakeLists.txt
+++ b/libc/src/math/CMakeLists.txt
@@ -342,6 +342,8 @@ add_math_entrypoint_object(nextupf128)
add_math_entrypoint_object(pow)
add_math_entrypoint_object(powf)
+add_math_entrypoint_object(powi)
+add_math_entrypoint_object(powif)
add_math_entrypoint_object(remainder)
add_math_entrypoint_object(remainderf)
diff --git a/libc/src/math/amdgpu/CMakeLists.txt b/libc/src/math/amdgpu/CMakeLists.txt
index 202177f54b11a..2ceb12785c607 100644
--- a/libc/src/math/amdgpu/CMakeLists.txt
+++ b/libc/src/math/amdgpu/CMakeLists.txt
@@ -468,6 +468,30 @@ add_entrypoint_object(
VENDOR
)
+add_entrypoint_object(
+ powi
+ SRCS
+ powi.cpp
+ HDRS
+ ../powi.h
+ COMPILE_OPTIONS
+ ${bitcode_link_flags}
+ -O2
+ VENDOR
+)
+
+add_entrypoint_object(
+ powif
+ SRCS
+ powif.cpp
+ HDRS
+ ../powif.h
+ COMPILE_OPTIONS
+ ${bitcode_link_flags}
+ -O2
+ VENDOR
+)
+
add_entrypoint_object(
sinh
SRCS
diff --git a/libc/src/math/amdgpu/declarations.h b/libc/src/math/amdgpu/declarations.h
index 5d7f3c9609d23..81d788d943f22 100644
--- a/libc/src/math/amdgpu/declarations.h
+++ b/libc/src/math/amdgpu/declarations.h
@@ -65,6 +65,8 @@ float __ocml_nextafter_f32(float, float);
double __ocml_nextafter_f64(double, double);
float __ocml_pow_f32(float, float);
double __ocml_pow_f64(double, double);
+float __ocml_pown_f32(float, int);
+double __ocml_pown_f64(double, int);
float __ocml_sin_f32(float);
double __ocml_sin_f64(double);
float __ocml_sincos_f32(float, float *);
diff --git a/libc/src/math/amdgpu/powi.cpp b/libc/src/math/amdgpu/powi.cpp
new file mode 100644
index 0000000000000..b13cb52aaaf6a
--- /dev/null
+++ b/libc/src/math/amdgpu/powi.cpp
@@ -0,0 +1,20 @@
+//===-- Implementation of the powi function for GPU -----------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "src/math/powi.h"
+#include "src/__support/common.h"
+
+#include "declarations.h"
+
+namespace LIBC_NAMESPACE {
+
+LLVM_LIBC_FUNCTION(double, powi, (double x, int y)) {
+ return __ocml_pown_f64(x, y);
+}
+
+} // namespace LIBC_NAMESPACE
diff --git a/libc/src/math/amdgpu/powif.cpp b/libc/src/math/amdgpu/powif.cpp
new file mode 100644
index 0000000000000..7682f0e1a27a6
--- /dev/null
+++ b/libc/src/math/amdgpu/powif.cpp
@@ -0,0 +1,20 @@
+//===-- Implementation of the powi function for GPU -----------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "src/math/powif.h"
+#include "src/__support/common.h"
+
+#include "declarations.h"
+
+namespace LIBC_NAMESPACE {
+
+LLVM_LIBC_FUNCTION(float, powif, (float x, int y)) {
+ return __ocml_pown_f32(x, y);
+}
+
+} // namespace LIBC_NAMESPACE
diff --git a/libc/src/math/nvptx/CMakeLists.txt b/libc/src/math/nvptx/CMakeLists.txt
index bf37c52f09e44..4295ebf9ff630 100644
--- a/libc/src/math/nvptx/CMakeLists.txt
+++ b/libc/src/math/nvptx/CMakeLists.txt
@@ -421,6 +421,30 @@ add_entrypoint_object(
VENDOR
)
+add_entrypoint_object(
+ powi
+ SRCS
+ powi.cpp
+ HDRS
+ ../powi.h
+ COMPILE_OPTIONS
+ ${bitcode_link_flags}
+ -O2
+ VENDOR
+)
+
+add_entrypoint_object(
+ powif
+ SRCS
+ powif.cpp
+ HDRS
+ ../powif.h
+ COMPILE_OPTIONS
+ ${bitcode_link_flags}
+ -O2
+ VENDOR
+)
+
add_entrypoint_object(
sinh
SRCS
diff --git a/libc/src/math/nvptx/declarations.h b/libc/src/math/nvptx/declarations.h
index d41b16c8eec9f..cf8f6e09e492a 100644
--- a/libc/src/math/nvptx/declarations.h
+++ b/libc/src/math/nvptx/declarations.h
@@ -64,6 +64,8 @@ double __nv_nextafter(double, double);
float __nv_nextafterf(float, float);
double __nv_pow(double, double);
float __nv_powf(float, float);
+double __nv_powi(double, int);
+float __nv_powif(float, int);
double __nv_sin(double);
float __nv_sinf(float);
void __nv_sincos(double, double *, double *);
diff --git a/libc/src/math/nvptx/powi.cpp b/libc/src/math/nvptx/powi.cpp
new file mode 100644
index 0000000000000..b7871ab9da92e
--- /dev/null
+++ b/libc/src/math/nvptx/powi.cpp
@@ -0,0 +1,18 @@
+//===-- Implementation of the powi function for GPU -----------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "src/math/powi.h"
+#include "src/__support/common.h"
+
+#include "declarations.h"
+
+namespace LIBC_NAMESPACE {
+
+LLVM_LIBC_FUNCTION(double, powi, (double x, int y)) { return __nv_powi(x, y); }
+
+} // namespace LIBC_NAMESPACE
diff --git a/libc/src/math/nvptx/powif.cpp b/libc/src/math/nvptx/powif.cpp
new file mode 100644
index 0000000000000..0143745479cec
--- /dev/null
+++ b/libc/src/math/nvptx/powif.cpp
@@ -0,0 +1,18 @@
+//===-- Implementation of the powif function for GPU ----------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "src/math/powif.h"
+#include "src/__support/common.h"
+
+#include "declarations.h"
+
+namespace LIBC_NAMESPACE {
+
+LLVM_LIBC_FUNCTION(float, powif, (float x, int y)) { return __nv_powif(x, y); }
+
+} // namespace LIBC_NAMESPACE
diff --git a/libc/src/math/powi.h b/libc/src/math/powi.h
new file mode 100644
index 0000000000000..71f6562d59154
--- /dev/null
+++ b/libc/src/math/powi.h
@@ -0,0 +1,18 @@
+//===-- Implementation header for powi --------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_LIBC_SRC_MATH_POW_H
+#define LLVM_LIBC_SRC_MATH_POW_H
+
+namespace LIBC_NAMESPACE {
+
+double powi(double x, int y);
+
+} // namespace LIBC_NAMESPACE
+
+#endif // LLVM_LIBC_SRC_MATH_POW_H
diff --git a/libc/src/math/powif.h b/libc/src/math/powif.h
new file mode 100644
index 0000000000000..08833dec33ab3
--- /dev/null
+++ b/libc/src/math/powif.h
@@ -0,0 +1,18 @@
+//===-- Implementation header for powif -------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_LIBC_SRC_MATH_POW_H
+#define LLVM_LIBC_SRC_MATH_POW_H
+
+namespace LIBC_NAMESPACE {
+
+float powif(float x, int y);
+
+} // namespace LIBC_NAMESPACE
+
+#endif // LLVM_LIBC_SRC_MATH_POW_H
|
lntue
reviewed
Jul 10, 2024
libc/src/math/powif.h
Outdated
| // | ||
| //===----------------------------------------------------------------------===// | ||
|
|
||
| #ifndef LLVM_LIBC_SRC_MATH_POW_H |
lntue
reviewed
Jul 10, 2024
libc/src/math/powi.h
Outdated
| // | ||
| //===----------------------------------------------------------------------===// | ||
|
|
||
| #ifndef LLVM_LIBC_SRC_MATH_POW_H |
Summary: This function is used by the CUDA / HIP / OpenMP headers and exists as an NVIDIA extension basically. This function is implemented in the C23 standard as `pown`, but for now we need to provide `powi` for backwards compatibility. In the future this entrypoint will just be a redirect to `pown` once that is implemented.
lntue
approved these changes
Jul 10, 2024
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/183/builds/1137 Here is the relevant piece of the build log for the reference: |
aaryanshukla
pushed a commit
to aaryanshukla/llvm-project
that referenced
this pull request
Jul 14, 2024
…lvm#98236) Summary: This function is used by the CUDA / HIP / OpenMP headers and exists as an NVIDIA extension basically. This function is implemented in the C23 standard as `pown`, but for now we need to provide `powi` for backwards compatibility. In the future this entrypoint will just be a redirect to `pown` once that is implemented.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary:
This function is used by the CUDA / HIP / OpenMP headers and exists as
an NVIDIA extension basically. This function is implemented in the C23
standard as
pown, but for now we need to providepowifor backwardscompatibility. In the future this entrypoint will just be a redirect to
pownonce that is implemented.