Add blackscholes without AMS support in repo

examples/simple/CMakeLists.txt

@@ -0,0 +1,46 @@
+# Copyright 2021-2023 Lawrence Livermore National Security, LLC and other
+# AMSLib Project Developers
+#
+# SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+
+cmake_minimum_required(VERSION 3.24)
+cmake_policy(SET CMP0104 NEW)
+
+# Define the project
+project(AMSExamples LANGUAGES CXX)
+
+
+option(WITH_AMS_EXAMPLE_CUDA  "Option to enable CUDA" Off)
+option(WITH_AMS_EXAMPLE_HIP "Option to enable HIP" Off)
+option(WITH_AMS_DOWNLOAD_DATA "Download benchmark datasets at configure time" OFF)
+
+if(WITH_AMS_EXAMPLE_CUDA AND WITH_AMS_EXAMPLE_HIP)
+  message(FATAL_ERROR
+    "WITH_AMS_EXAMPLE_CUDA=${WITH_AMS_EXAMPLE_CUDA} and "
+    "WITH_AMS_EXAMPLE_HIP=${WITH_AMS_EXAMPLE_HIP} are mutually exclusive."
+  )
+endif()
+
+
+set(AMS_INPUT_DIR "${CMAKE_BINARY_DIR}/inputs")
+set(CMAKE_CXX_STANDARD 17)
+set(CMAKE_CXX_STANDARD_REQUIRED ON)
+
+if (WITH_AMS_EXAMPLE_CUDA)
+message(STATUS "HERE")
+  set(THREADS_PREFER_PTHREAD_FLAG ON)
+  find_package(Threads REQUIRED)
+  enable_language(CUDA)
+  find_package(CUDAToolkit)
+elseif (WITH_AMS_EXAMPLE_HIP)
+  enable_language(HIP)
+  find_package(hip REQUIRED)
+endif()
+
+if (WITH_AMS_DOWNLOAD_DATA)
+enable_testing()
+add_subdirectory(inputs)
+endif()
+
+
+add_subdirectory(blackscholes)

examples/simple/blackscholes/CMakeLists.txt

@@ -0,0 +1,88 @@
+
+add_executable(blck_cpu driver.cpp cpu_compute.cpp)
+
+if (WITH_AMS_EXAMPLE_HIP)
+  add_executable(blck_hip driver.cpp gpu_compute.cpp)
+  set_source_files_properties(gpu_compute.cpp PROPERTIES LANGUAGE HIP)
+  target_link_libraries(blck_hip PRIVATE hip::device hip::host)
+  target_include_directories(blck_hip
+    PRIVATE
+      ${PROJECT_SOURCE_DIR}/include
+  )
+  target_compile_definitions(blck_hip PRIVATE AMS_EXAMPLE_ENABLE_HIP)
+elseif(WITH_AMS_EXAMPLE_CUDA)
+  add_executable(blck_cuda driver.cpp gpu_compute.cpp)
+  set_source_files_properties(gpu_compute.cpp PROPERTIES LANGUAGE CUDA)
+  target_link_libraries(blck_cuda PRIVATE CUDA::cudart)
+  target_include_directories(blck_cuda
+    PRIVATE
+      ${PROJECT_SOURCE_DIR}/include
+  )
+  target_compile_definitions(blck_cuda PRIVATE AMS_EXAMPLE_ENABLE_CUDA)
+endif()
+
+
+target_include_directories(blck_cpu
+  PRIVATE
+    ${PROJECT_SOURCE_DIR}/include
+)
+
+
+if (WITH_AMS_DOWNLOAD_DATA)
+set(BS_ZST "${AMS_INPUT_DIR}/blackscholes_input.bin.zst")
+set(BS_BIN "${AMS_INPUT_DIR}/blackscholes_input.bin")
+
+  ams_download_dataset(
+    NAME     blackscholes_input_zst
+    URL      "https://zenodo.org/records/17992861/files/random_input.bin.zst?download=1"
+    SHA256   "b895a9e5206ed60e5fd63529ceeb5e63d823149de4559f578fcab58cb2908ee4"
+    OUT_FILE "${BS_ZST}"
+  )
+
+# Always define the build-time decompression target (it will run when needed)
+ams_add_zstd_decompress_target(
+  NAME     inputs_blackscholes
+  ZST_FILE "${BS_ZST}"
+  OUT_FILE "${BS_BIN}"
+)
+
+add_dependencies(blck_cpu inputs_blackscholes)
+add_test(
+  NAME BLACKSCHOLES::CPU
+  COMMAND blck_cpu
+          ${BS_BIN}
+)
+set_tests_properties(BLACKSCHOLES::CPU
+  PROPERTIES
+    LABELS "benchmark;blackscholes;cpu;no-ams"
+    TIMEOUT 600
+)
+
+if (WITH_AMS_EXAMPLE_HIP)
+add_dependencies(blck_hip inputs_blackscholes)
+add_test(
+  NAME BLACKSCHOLES::CPU
+  COMMAND blck_hip
+          ${BS_BIN}
+)
+set_tests_properties(BLACKSCHOLES::GPU
+  PROPERTIES
+    LABELS "benchmark;blackscholes;gpu;no-ams"
+    TIMEOUT 600
+)
+elseif(WITH_AMS_EXAMPLE_CUDA)
+add_dependencies(blck_cpu inputs_blackscholes)
+add_test(
+  NAME BLACKSCHOLES::GPU
+  COMMAND blck_cuda
+          ${BS_BIN}
+)
+set_tests_properties(BLACKSCHOLES::CPU
+  PROPERTIES
+    LABELS "benchmark;blackscholes;gpu;no-ams"
+    TIMEOUT 600
+)
+endif()
+
+
+endif()

examples/simple/blackscholes/constants.hpp

@@ -0,0 +1,12 @@
+#define fptype double
+
+const fptype inv_sqrt_2xPI = 0.39894228040143270286f;
+const fptype zero = 0.0;
+const fptype half = 0.5;
+const fptype const1 = 0.2316419;
+const fptype one = 1.0;
+const fptype const2 = 0.319381530;
+const fptype const3 = 0.356563782;
+const fptype const4 = 1.781477937;
+const fptype const5 = 1.821255978;
+const fptype const6 = 1.330274429;

examples/simple/blackscholes/cpu_compute.cpp

@@ -0,0 +1,179 @@
+#include <cmath>
+
+#include "constants.hpp"
+
+////////////////////////////////////////////////////////////////////////////////
+////////////////////////////////////////////////////////////////////////////////
+///////////////////////////////////////////////////////////////////////////////
+////////////////////////////////////////////////////////////////////////////////
+// Cumulative Normal Distribution Function
+// See Hull, Section 11.8, P.243-244
+//
+
+fptype CNDF(fptype InputX)
+{
+  int sign;
+
+  fptype OutputX;
+  fptype xInput;
+  fptype xNPrimeofX;
+  fptype expValues;
+  fptype xK2;
+  fptype xK2_2, xK2_3;
+  fptype xK2_4, xK2_5;
+  fptype xLocal, xLocal_1;
+  fptype xLocal_2, xLocal_3;
+  fptype temp;
+
+  // Check for negative value of InputX
+  if (InputX < zero) {
+    InputX = -InputX;
+    sign = 1;
+  } else
+    sign = 0;
+
+  xInput = InputX;
+
+  // Compute NPrimeX term common to both four & six decimal accuracy calcs
+  temp = -half * InputX * InputX;
+
+  expValues = exp(temp);
+
+  xNPrimeofX = expValues;
+  xNPrimeofX = xNPrimeofX * inv_sqrt_2xPI;
+
+  xK2 = const1 * xInput;
+  xK2 = one + xK2;
+  xK2 = one / xK2;
+  xK2_2 = xK2 * xK2;
+  xK2_3 = xK2_2 * xK2;
+  xK2_4 = xK2_3 * xK2;
+  xK2_5 = xK2_4 * xK2;
+
+  xLocal_1 = xK2 * const2;
+  xLocal_2 = xK2_2 * (-const3);
+  xLocal_3 = xK2_3 * const4;
+  xLocal_2 = xLocal_2 + xLocal_3;
+  xLocal_3 = xK2_4 * (-const5);
+  xLocal_2 = xLocal_2 + xLocal_3;
+  xLocal_3 = xK2_5 * const6;
+  xLocal_2 = xLocal_2 + xLocal_3;
+
+  xLocal_1 = xLocal_2 + xLocal_1;
+  xLocal = xLocal_1 * xNPrimeofX;
+  xLocal = one - xLocal;
+
+  OutputX = xLocal;
+
+  if (sign) {
+    OutputX = one - OutputX;
+  }
+
+  return OutputX;
+}
+
+fptype BlkSchlsEqEuroNoDiv(fptype sptprice,
+                           fptype strike,
+                           fptype rate,
+                           fptype volatility,
+                           fptype time,
+                           int otype,
+                           float timet)
+{
+  fptype OptionPrice;
+
+  // local private working variables for the calculation
+  fptype xStockPrice;
+  fptype xStrikePrice;
+  fptype xRiskFreeRate;
+  fptype xVolatility;
+
+  fptype xTime;
+  fptype xSqrtTime;
+
+  fptype logValues;
+  fptype xLogTerm;
+  fptype xD1;
+  fptype xD2;
+  fptype xPowerTerm;
+  fptype xDen;
+  fptype d1;
+  fptype d2;
+  fptype FutureValueX;
+  fptype NofXd1;
+  fptype NofXd2;
+  fptype NegNofXd1;
+  fptype NegNofXd2;
+  fptype temp;
+
+  xStockPrice = sptprice;
+  xStrikePrice = strike;
+  xRiskFreeRate = rate;
+  xVolatility = volatility;
+
+  xTime = time;
+  xSqrtTime = sqrt(xTime);
+
+  temp = sptprice / strike;
+
+  logValues = log(sptprice / strike);
+
+  xLogTerm = logValues;
+
+
+  xPowerTerm = xVolatility * xVolatility;
+  xPowerTerm = xPowerTerm * half;
+
+  xD1 = xRiskFreeRate + xPowerTerm;
+  xD1 = xD1 * xTime;
+  xD1 = xD1 + xLogTerm;
+
+  xDen = xVolatility * xSqrtTime;
+  xD1 = xD1 / xDen;
+  xD2 = xD1 - xDen;
+
+  d1 = xD1;
+  d2 = xD2;
+
+  //@APPROX LABEL("CNDF_1") IN(d1) OUT(NofXd1) APPROX_TECH(MEMO_IN|MEMO_OUT)
+  NofXd1 = CNDF(d1);
+
+  //@APPROX LABEL("CNDF_2") IN(d2) OUT(NofXd2) APPROX_TECH(MEMO_IN|MEMO_OUT)
+  NofXd2 = CNDF(d2);
+
+  temp = -(rate * time);
+
+  FutureValueX = (exp(temp));
+
+  FutureValueX *= strike;
+
+  if (otype == 0) {
+    OptionPrice = (sptprice * NofXd1) - (FutureValueX * NofXd2);
+  } else {
+    NegNofXd1 = (one - NofXd1);
+    NegNofXd2 = (one - NofXd2);
+    OptionPrice = (FutureValueX * NegNofXd2) - (sptprice * NegNofXd1);
+  }
+
+  return OptionPrice;
+}
+
+
+int compute(fptype *sptprice,
+            fptype *strike,
+            fptype *rate,
+            fptype *volatility,
+            fptype *otime,
+            int *otype,
+            fptype *prices,
+            size_t numOptions)
+{
+  int i, j, k;
+  fptype priceDelta;
+  for (i = 0; i < numOptions; i++) {
+    prices[i] = BlkSchlsEqEuroNoDiv(
+        sptprice[i], strike[i], rate[i], volatility[i], otime[i], otype[i], 0);
+  }
+
+  return 0;
+}