[INFRA] Import NVIDIA/CCCL upstream as optimization reference library

CCCL (CUDA C++ Core Libraries) provides:
- CUB: device/block/warp-level GPU primitives (reduce, scan, sort, topk)
- Thrust: high-level parallel algorithms (transform_reduce, sort, scan)
- libcudacxx: CUDA C++ standard library (atomics, barriers, memory)
- cudax: experimental features (memory resources, allocators)
- Tuning policies: per-SM hardware-specific algorithm parameters

Competition optimization vectors mapped to CCCL:
- Output TPS (83% weight): warp_reduce, block_reduce, device_topk
- Input TPS (14% weight): device_scan, block_load, prefetch
- Cache TPS (3% weight): prefix caching strategy patterns
- Memory (0.9 util): pooled/cached/buddy allocators

Source: https://github.com/NVIDIA/cccl (shallow clone, HEAD only)
License: Apache-2.0

cccl_upstream/thrust/testing/complex_transform.cu

@@ -0,0 +1,357 @@
+#include <thrust/complex.h>
+#include <thrust/host_vector.h>
+#include <thrust/transform.h>
+
+#include <iostream>
+
+#include <unittest/unittest.h>
+
+#if THRUST_DEVICE_SYSTEM == THRUST_DEVICE_SYSTEM_CUDA
+#  include <unittest/cuda/testframework.h>
+#endif
+
+struct basic_arithmetic_functor
+{
+  template <typename T>
+  _CCCL_HOST_DEVICE thrust::complex<T> operator()(const thrust::complex<T>& x, const thrust::complex<T>& y)
+  {
+    // exercise unary and binary arithmetic operators
+    // Should return approximately 1
+    return (+x + +y) + (x * y) / (y * x) + (-y + -x);
+  } // end operator()()
+}; // end make_pair_functor
+
+struct complex_plane_functor
+{
+  template <typename T>
+  _CCCL_HOST_DEVICE thrust::complex<T> operator()(const thrust::complex<T>& x)
+  {
+    // Should return a proximately 1
+    return thrust::proj((thrust::polar(abs(x), arg(x)) * conj(x)) / norm(x));
+  } // end operator()()
+}; // end make_pair_functor
+
+struct pow_functor
+{
+  template <typename T>
+  _CCCL_HOST_DEVICE thrust::complex<T> operator()(const thrust::complex<T>& x, const thrust::complex<T>& y)
+  {
+    // exercise power functions
+    return pow(x, y);
+  } // end operator()()
+}; // end make_pair_functor
+
+struct sqrt_functor
+{
+  template <typename T>
+  _CCCL_HOST_DEVICE thrust::complex<T> operator()(const thrust::complex<T>& x)
+  {
+    // exercise power functions
+    return sqrt(x);
+  } // end operator()()
+}; // end make_pair_functor
+
+struct log_functor
+{
+  template <typename T>
+  _CCCL_HOST_DEVICE thrust::complex<T> operator()(const thrust::complex<T>& x)
+  {
+    return log(x);
+  } // end operator()()
+}; // end make_pair_functor
+
+struct exp_functor
+{
+  template <typename T>
+  _CCCL_HOST_DEVICE thrust::complex<T> operator()(const thrust::complex<T>& x)
+  {
+    return exp(x);
+  } // end operator()()
+}; // end make_pair_functor
+
+struct log10_functor
+{
+  template <typename T>
+  _CCCL_HOST_DEVICE thrust::complex<T> operator()(const thrust::complex<T>& x)
+  {
+    return log10(x);
+  } // end operator()()
+}; // end make_pair_functor
+
+struct cos_functor
+{
+  template <typename T>
+  _CCCL_HOST_DEVICE thrust::complex<T> operator()(const thrust::complex<T>& x)
+  {
+    return cos(x);
+  }
+};
+
+struct sin_functor
+{
+  template <typename T>
+  _CCCL_HOST_DEVICE thrust::complex<T> operator()(const thrust::complex<T>& x)
+  {
+    return sin(x);
+  }
+};
+
+struct tan_functor
+{
+  template <typename T>
+  _CCCL_HOST_DEVICE thrust::complex<T> operator()(const thrust::complex<T>& x)
+  {
+    return tan(x);
+  }
+};
+
+struct cosh_functor
+{
+  template <typename T>
+  _CCCL_HOST_DEVICE thrust::complex<T> operator()(const thrust::complex<T>& x)
+  {
+    return cosh(x);
+  }
+};
+
+struct sinh_functor
+{
+  template <typename T>
+  _CCCL_HOST_DEVICE thrust::complex<T> operator()(const thrust::complex<T>& x)
+  {
+    return sinh(x);
+  }
+};
+
+struct tanh_functor
+{
+  template <typename T>
+  _CCCL_HOST_DEVICE thrust::complex<T> operator()(const thrust::complex<T>& x)
+  {
+    return tanh(x);
+  }
+};
+
+struct acos_functor
+{
+  template <typename T>
+  _CCCL_HOST_DEVICE thrust::complex<T> operator()(const thrust::complex<T>& x)
+  {
+    return acos(x);
+  }
+};
+
+struct asin_functor
+{
+  template <typename T>
+  _CCCL_HOST_DEVICE thrust::complex<T> operator()(const thrust::complex<T>& x)
+  {
+    return asin(x);
+  }
+};
+
+struct atan_functor
+{
+  template <typename T>
+  _CCCL_HOST_DEVICE thrust::complex<T> operator()(const thrust::complex<T>& x)
+  {
+    return atan(x);
+  }
+};
+
+struct acosh_functor
+{
+  template <typename T>
+  _CCCL_HOST_DEVICE thrust::complex<T> operator()(const thrust::complex<T>& x)
+  {
+    return acosh(x);
+  }
+};
+
+struct asinh_functor
+{
+  template <typename T>
+  _CCCL_HOST_DEVICE thrust::complex<T> operator()(const thrust::complex<T>& x)
+  {
+    return asinh(x);
+  }
+};
+
+struct atanh_functor
+{
+  template <typename T>
+  _CCCL_HOST_DEVICE thrust::complex<T> operator()(const thrust::complex<T>& x)
+  {
+    return atanh(x);
+  }
+};
+
+template <typename T>
+thrust::host_vector<thrust::complex<T>> random_complex_samples(size_t n)
+{
+  thrust::host_vector<T> real = unittest::random_samples<T>(2 * n);
+  thrust::host_vector<thrust::complex<T>> h_p1(n);
+  for (size_t i = 0; i < n; i++)
+  {
+    h_p1[i].real(real[i]);
+    h_p1[i].imag(real[2 * i]);
+  }
+  return h_p1;
+}
+
+template <typename T>
+struct TestComplexArithmeticTransform
+{
+  void operator()(const size_t n)
+  {
+    using type                     = thrust::complex<T>;
+    thrust::host_vector<type> h_p1 = random_complex_samples<T>(n);
+    thrust::host_vector<type> h_p2 = random_complex_samples<T>(n);
+    thrust::host_vector<type> h_result(n);
+
+    thrust::device_vector<type> d_p1 = h_p1;
+    thrust::device_vector<type> d_p2 = h_p2;
+    thrust::device_vector<type> d_result(n);
+
+    thrust::transform(h_p1.begin(), h_p1.end(), h_p2.begin(), h_result.begin(), basic_arithmetic_functor());
+    thrust::transform(d_p1.begin(), d_p1.end(), d_p2.begin(), d_result.begin(), basic_arithmetic_functor());
+    ASSERT_ALMOST_EQUAL(h_result, d_result);
+  }
+};
+VariableUnitTest<TestComplexArithmeticTransform, FloatingPointTypes> TestComplexArithmeticTransformInstance;
+
+template <typename T>
+struct TestComplexPlaneTransform
+{
+  void operator()(const size_t n)
+  {
+    using type                     = thrust::complex<T>;
+    thrust::host_vector<type> h_p1 = random_complex_samples<T>(n);
+    thrust::host_vector<type> h_result(n);
+
+    thrust::device_vector<type> d_p1 = h_p1;
+    thrust::device_vector<type> d_result(n);
+
+    thrust::transform(h_p1.begin(), h_p1.end(), h_result.begin(), complex_plane_functor());
+    thrust::transform(d_p1.begin(), d_p1.end(), d_result.begin(), complex_plane_functor());
+    ASSERT_ALMOST_EQUAL(h_result, d_result);
+  }
+};
+VariableUnitTest<TestComplexPlaneTransform, FloatingPointTypes> TestComplexPlaneTransformInstance;
+
+template <typename T>
+struct TestComplexPowerTransform
+{
+  void operator()(const size_t n)
+  {
+    using type                     = thrust::complex<T>;
+    thrust::host_vector<type> h_p1 = random_complex_samples<T>(n);
+    thrust::host_vector<type> h_p2 = random_complex_samples<T>(n);
+    thrust::host_vector<type> h_result(n);
+
+    thrust::device_vector<type> d_p1 = h_p1;
+    thrust::device_vector<type> d_p2 = h_p2;
+    thrust::device_vector<type> d_result(n);
+
+    thrust::transform(h_p1.begin(), h_p1.end(), h_p2.begin(), h_result.begin(), pow_functor());
+    thrust::transform(d_p1.begin(), d_p1.end(), d_p2.begin(), d_result.begin(), pow_functor());
+    // pow can be very inaccurate there's no point trying to check for equality
+    // Currently just checking for compilation
+    //    ASSERT_ALMOST_EQUAL(h_result, d_result);
+
+    thrust::transform(h_p1.begin(), h_p1.end(), h_result.begin(), sqrt_functor());
+    thrust::transform(d_p1.begin(), d_p1.end(), d_result.begin(), sqrt_functor());
+    ASSERT_ALMOST_EQUAL(h_result, d_result);
+  }
+};
+VariableUnitTest<TestComplexPowerTransform, FloatingPointTypes> TestComplexPowerTransformInstance;
+
+template <typename T>
+struct TestComplexExponentialTransform
+{
+  void operator()(const size_t n)
+  {
+    using type                     = thrust::complex<T>;
+    thrust::host_vector<type> h_p1 = random_complex_samples<T>(n);
+    thrust::host_vector<type> h_result(n);
+
+    thrust::device_vector<type> d_p1 = h_p1;
+    thrust::device_vector<type> d_result(n);
+
+    thrust::transform(h_p1.begin(), h_p1.end(), h_result.begin(), exp_functor());
+    thrust::transform(d_p1.begin(), d_p1.end(), d_result.begin(), exp_functor());
+    ASSERT_ALMOST_EQUAL(h_result, d_result);
+
+    thrust::transform(h_p1.begin(), h_p1.end(), h_result.begin(), log_functor());
+    thrust::transform(d_p1.begin(), d_p1.end(), d_result.begin(), log_functor());
+    ASSERT_ALMOST_EQUAL(h_result, d_result);
+
+    thrust::transform(h_p1.begin(), h_p1.end(), h_result.begin(), log10_functor());
+    thrust::transform(d_p1.begin(), d_p1.end(), d_result.begin(), log10_functor());
+    ASSERT_ALMOST_EQUAL(h_result, d_result);
+  }
+};
+VariableUnitTest<TestComplexExponentialTransform, FloatingPointTypes> TestComplexExponentialTransformInstance;
+
+template <typename T>
+struct TestComplexTrigonometricTransform
+{
+  void operator()(const size_t n)
+  {
+    using type                     = thrust::complex<T>;
+    thrust::host_vector<type> h_p1 = random_complex_samples<T>(n);
+    thrust::host_vector<type> h_result(n);
+
+    thrust::device_vector<type> d_p1 = h_p1;
+    thrust::device_vector<type> d_result(n);
+
+    thrust::transform(h_p1.begin(), h_p1.end(), h_result.begin(), sin_functor());
+    thrust::transform(d_p1.begin(), d_p1.end(), d_result.begin(), sin_functor());
+    ASSERT_ALMOST_EQUAL(h_result, d_result);
+
+    thrust::transform(h_p1.begin(), h_p1.end(), h_result.begin(), cos_functor());
+    thrust::transform(d_p1.begin(), d_p1.end(), d_result.begin(), cos_functor());
+    ASSERT_ALMOST_EQUAL(h_result, d_result);
+
+    thrust::transform(h_p1.begin(), h_p1.end(), h_result.begin(), tan_functor());
+    thrust::transform(d_p1.begin(), d_p1.end(), d_result.begin(), tan_functor());
+    ASSERT_ALMOST_EQUAL(h_result, d_result);
+
+    thrust::transform(h_p1.begin(), h_p1.end(), h_result.begin(), sinh_functor());
+    thrust::transform(d_p1.begin(), d_p1.end(), d_result.begin(), sinh_functor());
+    ASSERT_ALMOST_EQUAL(h_result, d_result);
+
+    thrust::transform(h_p1.begin(), h_p1.end(), h_result.begin(), cosh_functor());
+    thrust::transform(d_p1.begin(), d_p1.end(), d_result.begin(), cosh_functor());
+    ASSERT_ALMOST_EQUAL(h_result, d_result);
+
+    thrust::transform(h_p1.begin(), h_p1.end(), h_result.begin(), tanh_functor());
+    thrust::transform(d_p1.begin(), d_p1.end(), d_result.begin(), tanh_functor());
+    ASSERT_ALMOST_EQUAL(h_result, d_result);
+
+    thrust::transform(h_p1.begin(), h_p1.end(), h_result.begin(), asin_functor());
+    thrust::transform(d_p1.begin(), d_p1.end(), d_result.begin(), asin_functor());
+    ASSERT_ALMOST_EQUAL(h_result, d_result);
+
+    thrust::transform(h_p1.begin(), h_p1.end(), h_result.begin(), acos_functor());
+    thrust::transform(d_p1.begin(), d_p1.end(), d_result.begin(), acos_functor());
+    ASSERT_ALMOST_EQUAL(h_result, d_result);
+
+    thrust::transform(h_p1.begin(), h_p1.end(), h_result.begin(), atan_functor());
+    thrust::transform(d_p1.begin(), d_p1.end(), d_result.begin(), atan_functor());
+    ASSERT_ALMOST_EQUAL(h_result, d_result);
+
+    thrust::transform(h_p1.begin(), h_p1.end(), h_result.begin(), asinh_functor());
+    thrust::transform(d_p1.begin(), d_p1.end(), d_result.begin(), asinh_functor());
+    ASSERT_ALMOST_EQUAL(h_result, d_result);
+
+    thrust::transform(h_p1.begin(), h_p1.end(), h_result.begin(), acosh_functor());
+    thrust::transform(d_p1.begin(), d_p1.end(), d_result.begin(), acosh_functor());
+    ASSERT_ALMOST_EQUAL(h_result, d_result);
+
+    thrust::transform(h_p1.begin(), h_p1.end(), h_result.begin(), atanh_functor());
+    thrust::transform(d_p1.begin(), d_p1.end(), d_result.begin(), atanh_functor());
+    ASSERT_ALMOST_EQUAL(h_result, d_result);
+  }
+};
+VariableUnitTest<TestComplexTrigonometricTransform, FloatingPointTypes> TestComplexTrigonometricTransformInstance;