@cpp-linter/clang-tidy-linux-arm64 17.0.6 → 18.1.8

package/README.md

@@ -1,7 +1,7 @@
 # @cpp-linter/clang-tidy-linux-arm64
 
-The Linux arm64 binary of clang-tidy from LLVM 17.0.6, for [`@cpp-linter/clang-tidy`](https://www.npmjs.com/package/@cpp-linter/clang-tidy).
+The Linux arm64 binary of clang-tidy from LLVM 18.1.8, for [`@cpp-linter/clang-tidy`](https://www.npmjs.com/package/@cpp-linter/clang-tidy).
 
 Install `@cpp-linter/clang-tidy` instead of this package: npm picks the binary for your platform.
 
-The binary is a static build from the release [`llvm-17.0.6-2`](https://github.com/cpp-linter/clang-tools-static-binaries/releases/tag/llvm-17.0.6-2) of [cpp-linter/clang-tools-static-binaries](https://github.com/cpp-linter/clang-tools-static-binaries). LLVM is licensed under the Apache License v2.0 with LLVM Exceptions (see `LICENSE.TXT`).
+The binary is a static build from the release [`llvm-18.1.8`](https://github.com/cpp-linter/clang-tools-static-binaries/releases/tag/llvm-18.1.8) of [cpp-linter/clang-tools-static-binaries](https://github.com/cpp-linter/clang-tools-static-binaries). LLVM is licensed under the Apache License v2.0 with LLVM Exceptions (see `LICENSE.TXT`).

package/bin/clang-tidy

@@ -1,4 +1,4 @@
 [diffend] Oversized file quarantined before diffing.
 name: package/bin/clang-tidy
-size: 62935440 bytes
-sha256: a86d9341c90f842eedb7c9391c3ce18d71d03ca998a51ce7325b69cde150dd42
+size: 67076560 bytes
+sha256: 7a5d0630cb32633b157197370641f3396afab4e8145fa6c85947b15863792619

package/lib/clang/{17 → 18}/include/__clang_cuda_device_functions.h

@@ -502,8 +502,8 @@ __DEVICE__ unsigned int __pm0(void) { return __nvvm_read_ptx_sreg_pm0(); }
 __DEVICE__ unsigned int __pm1(void) { return __nvvm_read_ptx_sreg_pm1(); }
 __DEVICE__ unsigned int __pm2(void) { return __nvvm_read_ptx_sreg_pm2(); }
 __DEVICE__ unsigned int __pm3(void) { return __nvvm_read_ptx_sreg_pm3(); }
-__DEVICE__ int __popc(int __a) { return __nv_popc(__a); }
-__DEVICE__ int __popcll(long long __a) { return __nv_popcll(__a); }
+__DEVICE__ int __popc(unsigned int __a) { return __nv_popc(__a); }
+__DEVICE__ int __popcll(unsigned long long __a) { return __nv_popcll(__a); }
 __DEVICE__ float __powf(float __a, float __b) {
   return __nv_fast_powf(__a, __b);
 }

package/lib/clang/{17 → 18}/include/__clang_cuda_libdevice_declares.h

@@ -285,8 +285,8 @@ __DEVICE__ double __nv_normcdfinv(double __a);
 __DEVICE__ float __nv_normcdfinvf(float __a);
 __DEVICE__ float __nv_normf(int __a, const float *__b);
 __DEVICE__ double __nv_norm(int __a, const double *__b);
-__DEVICE__ int __nv_popc(int __a);
-__DEVICE__ int __nv_popcll(long long __a);
+__DEVICE__ int __nv_popc(unsigned int __a);
+__DEVICE__ int __nv_popcll(unsigned long long __a);
 __DEVICE__ double __nv_pow(double __a, double __b);
 __DEVICE__ float __nv_powf(float __a, float __b);
 __DEVICE__ double __nv_powi(double __a, int __b);

package/lib/clang/{17 → 18}/include/__clang_cuda_math.h

@@ -45,9 +45,9 @@
 // libdevice provides fast low precision and slow full-recision implementations
 // for some functions. Which one gets selected depends on
 // __CLANG_CUDA_APPROX_TRANSCENDENTALS__ which gets defined by clang if
-// -ffast-math or -fcuda-approx-transcendentals are in effect.
+// -ffast-math or -fgpu-approx-transcendentals are in effect.
 #pragma push_macro("__FAST_OR_SLOW")
-#if defined(__CLANG_CUDA_APPROX_TRANSCENDENTALS__)
+#if defined(__CLANG_GPU_APPROX_TRANSCENDENTALS__)
 #define __FAST_OR_SLOW(fast, slow) fast
 #else
 #define __FAST_OR_SLOW(fast, slow) slow

package/lib/clang/{17 → 18}/include/__clang_cuda_runtime_wrapper.h

@@ -196,12 +196,12 @@ inline __host__ double __signbitd(double x) {
 
 // math_function.hpp uses the __USE_FAST_MATH__ macro to determine whether we
 // get the slow-but-accurate or fast-but-inaccurate versions of functions like
-// sin and exp.  This is controlled in clang by -fcuda-approx-transcendentals.
+// sin and exp.  This is controlled in clang by -fgpu-approx-transcendentals.
 //
 // device_functions.hpp uses __USE_FAST_MATH__ for a different purpose (fast vs.
 // slow divides), so we need to scope our define carefully here.
 #pragma push_macro("__USE_FAST_MATH__")
-#if defined(__CLANG_CUDA_APPROX_TRANSCENDENTALS__)
+#if defined(__CLANG_GPU_APPROX_TRANSCENDENTALS__)
 #define __USE_FAST_MATH__ 1
 #endif
 

package/lib/clang/{17 → 18}/include/__clang_hip_math.h

@@ -14,9 +14,6 @@
 #endif
 
 #if !defined(__HIPCC_RTC__)
-#if defined(__cplusplus)
-#include <algorithm>
-#endif
 #include <limits.h>
 #include <stdint.h>
 #ifdef __OPENMP_AMDGCN__
@@ -32,6 +29,17 @@
 #define __DEVICE__ static __device__ inline __attribute__((always_inline))
 #endif
 
+// Device library provides fast low precision and slow full-recision
+// implementations for some functions. Which one gets selected depends on
+// __CLANG_GPU_APPROX_TRANSCENDENTALS__ which gets defined by clang if
+// -ffast-math or -fgpu-approx-transcendentals are in effect.
+#pragma push_macro("__FAST_OR_SLOW")
+#if defined(__CLANG_GPU_APPROX_TRANSCENDENTALS__)
+#define __FAST_OR_SLOW(fast, slow) fast
+#else
+#define __FAST_OR_SLOW(fast, slow) slow
+#endif
+
 // A few functions return bool type starting only in C++11.
 #pragma push_macro("__RETURN_TYPE")
 #ifdef __OPENMP_AMDGCN__
@@ -139,21 +147,180 @@ uint64_t __make_mantissa(const char *__tagp __attribute__((nonnull))) {
 }
 
 // BEGIN FLOAT
+
+// BEGIN INTRINSICS
+
+__DEVICE__
+float __cosf(float __x) { return __ocml_native_cos_f32(__x); }
+
+__DEVICE__
+float __exp10f(float __x) {
+  const float __log2_10 = 0x1.a934f0p+1f;
+  return __builtin_amdgcn_exp2f(__log2_10 * __x);
+}
+
+__DEVICE__
+float __expf(float __x) {
+  const float __log2_e = 0x1.715476p+0;
+  return __builtin_amdgcn_exp2f(__log2_e * __x);
+}
+
+#if defined OCML_BASIC_ROUNDED_OPERATIONS
+__DEVICE__
+float __fadd_rd(float __x, float __y) { return __ocml_add_rtn_f32(__x, __y); }
+__DEVICE__
+float __fadd_rn(float __x, float __y) { return __ocml_add_rte_f32(__x, __y); }
+__DEVICE__
+float __fadd_ru(float __x, float __y) { return __ocml_add_rtp_f32(__x, __y); }
+__DEVICE__
+float __fadd_rz(float __x, float __y) { return __ocml_add_rtz_f32(__x, __y); }
+#else
+__DEVICE__
+float __fadd_rn(float __x, float __y) { return __x + __y; }
+#endif
+
+#if defined OCML_BASIC_ROUNDED_OPERATIONS
+__DEVICE__
+float __fdiv_rd(float __x, float __y) { return __ocml_div_rtn_f32(__x, __y); }
+__DEVICE__
+float __fdiv_rn(float __x, float __y) { return __ocml_div_rte_f32(__x, __y); }
+__DEVICE__
+float __fdiv_ru(float __x, float __y) { return __ocml_div_rtp_f32(__x, __y); }
+__DEVICE__
+float __fdiv_rz(float __x, float __y) { return __ocml_div_rtz_f32(__x, __y); }
+#else
+__DEVICE__
+float __fdiv_rn(float __x, float __y) { return __x / __y; }
+#endif
+
+__DEVICE__
+float __fdividef(float __x, float __y) { return __x / __y; }
+
+#if defined OCML_BASIC_ROUNDED_OPERATIONS
+__DEVICE__
+float __fmaf_rd(float __x, float __y, float __z) {
+  return __ocml_fma_rtn_f32(__x, __y, __z);
+}
+__DEVICE__
+float __fmaf_rn(float __x, float __y, float __z) {
+  return __ocml_fma_rte_f32(__x, __y, __z);
+}
+__DEVICE__
+float __fmaf_ru(float __x, float __y, float __z) {
+  return __ocml_fma_rtp_f32(__x, __y, __z);
+}
+__DEVICE__
+float __fmaf_rz(float __x, float __y, float __z) {
+  return __ocml_fma_rtz_f32(__x, __y, __z);
+}
+#else
+__DEVICE__
+float __fmaf_rn(float __x, float __y, float __z) {
+  return __builtin_fmaf(__x, __y, __z);
+}
+#endif
+
+#if defined OCML_BASIC_ROUNDED_OPERATIONS
+__DEVICE__
+float __fmul_rd(float __x, float __y) { return __ocml_mul_rtn_f32(__x, __y); }
+__DEVICE__
+float __fmul_rn(float __x, float __y) { return __ocml_mul_rte_f32(__x, __y); }
+__DEVICE__
+float __fmul_ru(float __x, float __y) { return __ocml_mul_rtp_f32(__x, __y); }
+__DEVICE__
+float __fmul_rz(float __x, float __y) { return __ocml_mul_rtz_f32(__x, __y); }
+#else
+__DEVICE__
+float __fmul_rn(float __x, float __y) { return __x * __y; }
+#endif
+
+#if defined OCML_BASIC_ROUNDED_OPERATIONS
+__DEVICE__
+float __frcp_rd(float __x) { return __ocml_div_rtn_f32(1.0f, __x); }
+__DEVICE__
+float __frcp_rn(float __x) { return __ocml_div_rte_f32(1.0f, __x); }
+__DEVICE__
+float __frcp_ru(float __x) { return __ocml_div_rtp_f32(1.0f, __x); }
+__DEVICE__
+float __frcp_rz(float __x) { return __ocml_div_rtz_f32(1.0f, __x); }
+#else
+__DEVICE__
+float __frcp_rn(float __x) { return 1.0f / __x; }
+#endif
+
+__DEVICE__
+float __frsqrt_rn(float __x) { return __builtin_amdgcn_rsqf(__x); }
+
+#if defined OCML_BASIC_ROUNDED_OPERATIONS
+__DEVICE__
+float __fsqrt_rd(float __x) { return __ocml_sqrt_rtn_f32(__x); }
+__DEVICE__
+float __fsqrt_rn(float __x) { return __ocml_sqrt_rte_f32(__x); }
+__DEVICE__
+float __fsqrt_ru(float __x) { return __ocml_sqrt_rtp_f32(__x); }
+__DEVICE__
+float __fsqrt_rz(float __x) { return __ocml_sqrt_rtz_f32(__x); }
+#else
+__DEVICE__
+float __fsqrt_rn(float __x) { return __ocml_native_sqrt_f32(__x); }
+#endif
+
+#if defined OCML_BASIC_ROUNDED_OPERATIONS
+__DEVICE__
+float __fsub_rd(float __x, float __y) { return __ocml_sub_rtn_f32(__x, __y); }
+__DEVICE__
+float __fsub_rn(float __x, float __y) { return __ocml_sub_rte_f32(__x, __y); }
+__DEVICE__
+float __fsub_ru(float __x, float __y) { return __ocml_sub_rtp_f32(__x, __y); }
+__DEVICE__
+float __fsub_rz(float __x, float __y) { return __ocml_sub_rtz_f32(__x, __y); }
+#else
+__DEVICE__
+float __fsub_rn(float __x, float __y) { return __x - __y; }
+#endif
+
+__DEVICE__
+float __log10f(float __x) { return __builtin_log10f(__x); }
+
+__DEVICE__
+float __log2f(float __x) { return __builtin_amdgcn_logf(__x); }
+
+__DEVICE__
+float __logf(float __x) { return __builtin_logf(__x); }
+
+__DEVICE__
+float __powf(float __x, float __y) { return __ocml_pow_f32(__x, __y); }
+
+__DEVICE__
+float __saturatef(float __x) { return (__x < 0) ? 0 : ((__x > 1) ? 1 : __x); }
+
+__DEVICE__
+void __sincosf(float __x, float *__sinptr, float *__cosptr) {
+  *__sinptr = __ocml_native_sin_f32(__x);
+  *__cosptr = __ocml_native_cos_f32(__x);
+}
+
+__DEVICE__
+float __sinf(float __x) { return __ocml_native_sin_f32(__x); }
+
+__DEVICE__
+float __tanf(float __x) {
+  return __sinf(__x) * __builtin_amdgcn_rcpf(__cosf(__x));
+}
+// END INTRINSICS
+
 #if defined(__cplusplus)
 __DEVICE__
 int abs(int __x) {
-  int __sgn = __x >> (sizeof(int) * CHAR_BIT - 1);
-  return (__x ^ __sgn) - __sgn;
+  return __builtin_abs(__x);
 }
 __DEVICE__
 long labs(long __x) {
-  long __sgn = __x >> (sizeof(long) * CHAR_BIT - 1);
-  return (__x ^ __sgn) - __sgn;
+  return __builtin_labs(__x);
 }
 __DEVICE__
 long long llabs(long long __x) {
-  long long __sgn = __x >> (sizeof(long long) * CHAR_BIT - 1);
-  return (__x ^ __sgn) - __sgn;
+  return __builtin_llabs(__x);
 }
 #endif
 
@@ -188,7 +355,7 @@ __DEVICE__
 float copysignf(float __x, float __y) { return __builtin_copysignf(__x, __y); }
 
 __DEVICE__
-float cosf(float __x) { return __ocml_cos_f32(__x); }
+float cosf(float __x) { return __FAST_OR_SLOW(__cosf, __ocml_cos_f32)(__x); }
 
 __DEVICE__
 float coshf(float __x) { return __ocml_cosh_f32(__x); }
@@ -321,13 +488,13 @@ __DEVICE__
 float log1pf(float __x) { return __ocml_log1p_f32(__x); }
 
 __DEVICE__
-float log2f(float __x) { return __builtin_log2f(__x); }
+float log2f(float __x) { return __FAST_OR_SLOW(__log2f, __ocml_log2_f32)(__x); }
 
 __DEVICE__
 float logbf(float __x) { return __ocml_logb_f32(__x); }
 
 __DEVICE__
-float logf(float __x) { return __builtin_logf(__x); }
+float logf(float __x) { return __FAST_OR_SLOW(__logf, __ocml_log_f32)(__x); }
 
 __DEVICE__
 long int lrintf(float __x) { return __builtin_rintf(__x); }
@@ -401,7 +568,7 @@ float normf(int __dim,
     ++__a;
   }
 
-  return __ocml_sqrt_f32(__r);
+  return __builtin_sqrtf(__r);
 }
 
 __DEVICE__
@@ -483,9 +650,13 @@ void sincosf(float __x, float *__sinptr, float *__cosptr) {
 #ifdef __OPENMP_AMDGCN__
 #pragma omp allocate(__tmp) allocator(omp_thread_mem_alloc)
 #endif
+#ifdef __CLANG_CUDA_APPROX_TRANSCENDENTALS__
+  __sincosf(__x, __sinptr, __cosptr);
+#else
   *__sinptr =
       __ocml_sincos_f32(__x, (__attribute__((address_space(5))) float *)&__tmp);
   *__cosptr = __tmp;
+#endif
 }
 
 __DEVICE__
@@ -500,7 +671,7 @@ void sincospif(float __x, float *__sinptr, float *__cosptr) {
 }
 
 __DEVICE__
-float sinf(float __x) { return __ocml_sin_f32(__x); }
+float sinf(float __x) { return __FAST_OR_SLOW(__sinf, __ocml_sin_f32)(__x); }
 
 __DEVICE__
 float sinhf(float __x) { return __ocml_sinh_f32(__x); }
@@ -509,7 +680,7 @@ __DEVICE__
 float sinpif(float __x) { return __ocml_sinpi_f32(__x); }
 
 __DEVICE__
-float sqrtf(float __x) { return __ocml_sqrt_f32(__x); }
+float sqrtf(float __x) { return __builtin_sqrtf(__x); }
 
 __DEVICE__
 float tanf(float __x) { return __ocml_tan_f32(__x); }
@@ -551,158 +722,7 @@ float ynf(int __n, float __x) { // TODO: we could use Ahmes multiplication
   return __x1;
 }
 
-// BEGIN INTRINSICS
-
-__DEVICE__
-float __cosf(float __x) { return __ocml_native_cos_f32(__x); }
-
-__DEVICE__
-float __exp10f(float __x) { return __ocml_native_exp10_f32(__x); }
-
-__DEVICE__
-float __expf(float __x) { return __ocml_native_exp_f32(__x); }
-
-#if defined OCML_BASIC_ROUNDED_OPERATIONS
-__DEVICE__
-float __fadd_rd(float __x, float __y) { return __ocml_add_rtn_f32(__x, __y); }
-__DEVICE__
-float __fadd_rn(float __x, float __y) { return __ocml_add_rte_f32(__x, __y); }
-__DEVICE__
-float __fadd_ru(float __x, float __y) { return __ocml_add_rtp_f32(__x, __y); }
-__DEVICE__
-float __fadd_rz(float __x, float __y) { return __ocml_add_rtz_f32(__x, __y); }
-#else
-__DEVICE__
-float __fadd_rn(float __x, float __y) { return __x + __y; }
-#endif
-
-#if defined OCML_BASIC_ROUNDED_OPERATIONS
-__DEVICE__
-float __fdiv_rd(float __x, float __y) { return __ocml_div_rtn_f32(__x, __y); }
-__DEVICE__
-float __fdiv_rn(float __x, float __y) { return __ocml_div_rte_f32(__x, __y); }
-__DEVICE__
-float __fdiv_ru(float __x, float __y) { return __ocml_div_rtp_f32(__x, __y); }
-__DEVICE__
-float __fdiv_rz(float __x, float __y) { return __ocml_div_rtz_f32(__x, __y); }
-#else
-__DEVICE__
-float __fdiv_rn(float __x, float __y) { return __x / __y; }
-#endif
-
-__DEVICE__
-float __fdividef(float __x, float __y) { return __x / __y; }
-
-#if defined OCML_BASIC_ROUNDED_OPERATIONS
-__DEVICE__
-float __fmaf_rd(float __x, float __y, float __z) {
-  return __ocml_fma_rtn_f32(__x, __y, __z);
-}
-__DEVICE__
-float __fmaf_rn(float __x, float __y, float __z) {
-  return __ocml_fma_rte_f32(__x, __y, __z);
-}
-__DEVICE__
-float __fmaf_ru(float __x, float __y, float __z) {
-  return __ocml_fma_rtp_f32(__x, __y, __z);
-}
-__DEVICE__
-float __fmaf_rz(float __x, float __y, float __z) {
-  return __ocml_fma_rtz_f32(__x, __y, __z);
-}
-#else
-__DEVICE__
-float __fmaf_rn(float __x, float __y, float __z) {
-  return __builtin_fmaf(__x, __y, __z);
-}
-#endif
 
-#if defined OCML_BASIC_ROUNDED_OPERATIONS
-__DEVICE__
-float __fmul_rd(float __x, float __y) { return __ocml_mul_rtn_f32(__x, __y); }
-__DEVICE__
-float __fmul_rn(float __x, float __y) { return __ocml_mul_rte_f32(__x, __y); }
-__DEVICE__
-float __fmul_ru(float __x, float __y) { return __ocml_mul_rtp_f32(__x, __y); }
-__DEVICE__
-float __fmul_rz(float __x, float __y) { return __ocml_mul_rtz_f32(__x, __y); }
-#else
-__DEVICE__
-float __fmul_rn(float __x, float __y) { return __x * __y; }
-#endif
-
-#if defined OCML_BASIC_ROUNDED_OPERATIONS
-__DEVICE__
-float __frcp_rd(float __x) { return __ocml_div_rtn_f32(1.0f, __x); }
-__DEVICE__
-float __frcp_rn(float __x) { return __ocml_div_rte_f32(1.0f, __x); }
-__DEVICE__
-float __frcp_ru(float __x) { return __ocml_div_rtp_f32(1.0f, __x); }
-__DEVICE__
-float __frcp_rz(float __x) { return __ocml_div_rtz_f32(1.0f, __x); }
-#else
-__DEVICE__
-float __frcp_rn(float __x) { return 1.0f / __x; }
-#endif
-
-__DEVICE__
-float __frsqrt_rn(float __x) { return __builtin_amdgcn_rsqf(__x); }
-
-#if defined OCML_BASIC_ROUNDED_OPERATIONS
-__DEVICE__
-float __fsqrt_rd(float __x) { return __ocml_sqrt_rtn_f32(__x); }
-__DEVICE__
-float __fsqrt_rn(float __x) { return __ocml_sqrt_rte_f32(__x); }
-__DEVICE__
-float __fsqrt_ru(float __x) { return __ocml_sqrt_rtp_f32(__x); }
-__DEVICE__
-float __fsqrt_rz(float __x) { return __ocml_sqrt_rtz_f32(__x); }
-#else
-__DEVICE__
-float __fsqrt_rn(float __x) { return __ocml_native_sqrt_f32(__x); }
-#endif
-
-#if defined OCML_BASIC_ROUNDED_OPERATIONS
-__DEVICE__
-float __fsub_rd(float __x, float __y) { return __ocml_sub_rtn_f32(__x, __y); }
-__DEVICE__
-float __fsub_rn(float __x, float __y) { return __ocml_sub_rte_f32(__x, __y); }
-__DEVICE__
-float __fsub_ru(float __x, float __y) { return __ocml_sub_rtp_f32(__x, __y); }
-__DEVICE__
-float __fsub_rz(float __x, float __y) { return __ocml_sub_rtz_f32(__x, __y); }
-#else
-__DEVICE__
-float __fsub_rn(float __x, float __y) { return __x - __y; }
-#endif
-
-__DEVICE__
-float __log10f(float __x) { return __ocml_native_log10_f32(__x); }
-
-__DEVICE__
-float __log2f(float __x) { return __ocml_native_log2_f32(__x); }
-
-__DEVICE__
-float __logf(float __x) { return __ocml_native_log_f32(__x); }
-
-__DEVICE__
-float __powf(float __x, float __y) { return __ocml_pow_f32(__x, __y); }
-
-__DEVICE__
-float __saturatef(float __x) { return (__x < 0) ? 0 : ((__x > 1) ? 1 : __x); }
-
-__DEVICE__
-void __sincosf(float __x, float *__sinptr, float *__cosptr) {
-  *__sinptr = __ocml_native_sin_f32(__x);
-  *__cosptr = __ocml_native_cos_f32(__x);
-}
-
-__DEVICE__
-float __sinf(float __x) { return __ocml_native_sin_f32(__x); }
-
-__DEVICE__
-float __tanf(float __x) { return __ocml_tan_f32(__x); }
-// END INTRINSICS
 // END FLOAT
 
 // BEGIN DOUBLE
@@ -941,7 +961,7 @@ double norm(int __dim,
     ++__a;
   }
 
-  return __ocml_sqrt_f64(__r);
+  return __builtin_sqrt(__r);
 }
 
 __DEVICE__
@@ -1064,7 +1084,7 @@ __DEVICE__
 double sinpi(double __x) { return __ocml_sinpi_f64(__x); }
 
 __DEVICE__
-double sqrt(double __x) { return __ocml_sqrt_f64(__x); }
+double sqrt(double __x) { return __builtin_sqrt(__x); }
 
 __DEVICE__
 double tan(double __x) { return __ocml_tan_f64(__x); }
@@ -1198,7 +1218,7 @@ __DEVICE__
 double __dsqrt_rz(double __x) { return __ocml_sqrt_rtz_f64(__x); }
 #else
 __DEVICE__
-double __dsqrt_rn(double __x) { return __ocml_sqrt_f64(__x); }
+double __dsqrt_rn(double __x) { return __builtin_sqrt(__x); }
 #endif
 
 #if defined OCML_BASIC_ROUNDED_OPERATIONS
@@ -1288,16 +1308,17 @@ double min(double __x, double __y) { return __builtin_fmin(__x, __y); }
 
 #if !defined(__HIPCC_RTC__) && !defined(__OPENMP_AMDGCN__)
 __host__ inline static int min(int __arg1, int __arg2) {
-  return std::min(__arg1, __arg2);
+  return __arg1 < __arg2 ? __arg1 : __arg2;
 }
 
 __host__ inline static int max(int __arg1, int __arg2) {
-  return std::max(__arg1, __arg2);
+  return __arg1 > __arg2 ? __arg1 : __arg2;
 }
 #endif // !defined(__HIPCC_RTC__) && !defined(__OPENMP_AMDGCN__)
 #endif
 
 #pragma pop_macro("__DEVICE__")
 #pragma pop_macro("__RETURN_TYPE")
+#pragma pop_macro("__FAST_OR_SLOW")
 
 #endif // __CLANG_HIP_MATH_H__

package/lib/clang/{17 → 18}/include/__clang_hip_runtime_wrapper.h

@@ -47,28 +47,9 @@ extern "C" {
 #endif //__cplusplus
 
 #if !defined(__HIPCC_RTC__)
-#include <cmath>
-#include <cstdlib>
-#include <stdlib.h>
 #if __has_include("hip/hip_version.h")
 #include "hip/hip_version.h"
 #endif // __has_include("hip/hip_version.h")
-#else
-typedef __SIZE_TYPE__ size_t;
-// Define macros which are needed to declare HIP device API's without standard
-// C/C++ headers. This is for readability so that these API's can be written
-// the same way as non-hipRTC use case. These macros need to be popped so that
-// they do not pollute users' name space.
-#pragma push_macro("NULL")
-#pragma push_macro("uint32_t")
-#pragma push_macro("uint64_t")
-#pragma push_macro("CHAR_BIT")
-#pragma push_macro("INT_MAX")
-#define NULL (void *)0
-#define uint32_t __UINT32_TYPE__
-#define uint64_t __UINT64_TYPE__
-#define CHAR_BIT __CHAR_BIT__
-#define INT_MAX __INTMAX_MAX__
 #endif // __HIPCC_RTC__
 
 typedef __SIZE_TYPE__ __hip_size_t;
@@ -78,11 +59,13 @@ extern "C" {
 #endif //__cplusplus
 
 #if HIP_VERSION_MAJOR * 100 + HIP_VERSION_MINOR >= 405
-extern "C" __device__ unsigned long long __ockl_dm_alloc(unsigned long long __size);
-extern "C" __device__ void __ockl_dm_dealloc(unsigned long long __addr);
+__device__ unsigned long long __ockl_dm_alloc(unsigned long long __size);
+__device__ void __ockl_dm_dealloc(unsigned long long __addr);
 #if __has_feature(address_sanitizer)
-extern "C" __device__ unsigned long long __asan_malloc_impl(unsigned long long __size, unsigned long long __pc);
-extern "C" __device__ void __asan_free_impl(unsigned long long __addr, unsigned long long __pc);
+__device__ unsigned long long __asan_malloc_impl(unsigned long long __size,
+                                                 unsigned long long __pc);
+__device__ void __asan_free_impl(unsigned long long __addr,
+                                 unsigned long long __pc);
 __attribute__((noinline, weak)) __device__ void *malloc(__hip_size_t __size) {
   unsigned long long __pc = (unsigned long long)__builtin_return_address(0);
   return (void *)__asan_malloc_impl(__size, __pc);
@@ -91,7 +74,7 @@ __attribute__((noinline, weak)) __device__ void free(void *__ptr) {
   unsigned long long __pc = (unsigned long long)__builtin_return_address(0);
   __asan_free_impl((unsigned long long)__ptr, __pc);
 }
-#else
+#else // __has_feature(address_sanitizer)
 __attribute__((weak)) inline __device__ void *malloc(__hip_size_t __size) {
   return (void *) __ockl_dm_alloc(__size);
 }
@@ -109,7 +92,7 @@ __attribute__((weak)) inline __device__ void *malloc(__hip_size_t __size) {
 __attribute__((weak)) inline __device__ void free(void *__ptr) {
   __hip_free(__ptr);
 }
-#else
+#else  // __HIP_ENABLE_DEVICE_MALLOC__
 __attribute__((weak)) inline __device__ void *malloc(__hip_size_t __size) {
   __builtin_trap();
   return (void *)0;
@@ -117,13 +100,38 @@ __attribute__((weak)) inline __device__ void *malloc(__hip_size_t __size) {
 __attribute__((weak)) inline __device__ void free(void *__ptr) {
   __builtin_trap();
 }
-#endif
+#endif // __HIP_ENABLE_DEVICE_MALLOC__
 #endif // HIP version check
 
 #ifdef __cplusplus
 } // extern "C"
 #endif //__cplusplus
 
+#if !defined(__HIPCC_RTC__)
+#include <cmath>
+#include <cstdlib>
+#include <stdlib.h>
+#if __has_include("hip/hip_version.h")
+#include "hip/hip_version.h"
+#endif // __has_include("hip/hip_version.h")
+#else
+typedef __SIZE_TYPE__ size_t;
+// Define macros which are needed to declare HIP device API's without standard
+// C/C++ headers. This is for readability so that these API's can be written
+// the same way as non-hipRTC use case. These macros need to be popped so that
+// they do not pollute users' name space.
+#pragma push_macro("NULL")
+#pragma push_macro("uint32_t")
+#pragma push_macro("uint64_t")
+#pragma push_macro("CHAR_BIT")
+#pragma push_macro("INT_MAX")
+#define NULL (void *)0
+#define uint32_t __UINT32_TYPE__
+#define uint64_t __UINT64_TYPE__
+#define CHAR_BIT __CHAR_BIT__
+#define INT_MAX __INTMAX_MAX__
+#endif // __HIPCC_RTC__
+
 #include <__clang_hip_libdevice_declares.h>
 #include <__clang_hip_math.h>
 #include <__clang_hip_stdlib.h>

package/lib/clang/18/include/__stdarg___gnuc_va_list.h

@@ -0,0 +1,13 @@
+/*===---- __stdarg___gnuc_va_list.h - Definition of __gnuc_va_list ---------===
+ *
+ * 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 __GNUC_VA_LIST
+#define __GNUC_VA_LIST
+typedef __builtin_va_list __gnuc_va_list;
+#endif

package/lib/clang/18/include/__stdarg___va_copy.h

@@ -0,0 +1,12 @@
+/*===---- __stdarg___va_copy.h - Definition of __va_copy -------------------===
+ *
+ * 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 __va_copy
+#define __va_copy(d, s) __builtin_va_copy(d, s)
+#endif