From 7828b00268bf5907b34620bf03046c33b39f9bbf Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Max=20L=C3=BCbke?= <mluebke@uni-potsdam.de>
Date: Thu, 7 Mar 2024 12:27:13 +0100
Subject: [PATCH] Add OpenSSL dependency and implement MD5 hash function

---
 CMakeLists.txt                                |  1 +
 src/CMakeLists.txt                            |  1 +
 src/Chemistry/SurrogateModels/DHT_Wrapper.cpp |  3 ++-
 .../SurrogateModels/HashFunctions.cpp         | 21 +++++++++++++++++++
 .../SurrogateModels/HashFunctions.hpp         |  1 +
 5 files changed, 26 insertions(+), 1 deletion(-)

diff --git a/CMakeLists.txt b/CMakeLists.txt
index c4ad05c6e..ba6c76b08 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -21,6 +21,7 @@ get_poet_version()
 find_package(MPI REQUIRED)
 find_package(OpenMP)
 find_package(RRuntime REQUIRED)
+find_package(OpenSSL REQUIRED)
 
 add_subdirectory(src)
 add_subdirectory(bench)
diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt
index 0d3d38786..3e443034e 100644
--- a/src/CMakeLists.txt
+++ b/src/CMakeLists.txt
@@ -20,6 +20,7 @@ target_link_libraries(poetlib PUBLIC
   PhreeqcRM 
   tug
   DHT_UCX
+  OpenSSL::Crypto
 )
 
 target_compile_definitions(poetlib PUBLIC STRICT_R_HEADERS OMPI_SKIP_MPICXX)
diff --git a/src/Chemistry/SurrogateModels/DHT_Wrapper.cpp b/src/Chemistry/SurrogateModels/DHT_Wrapper.cpp
index 6b78c82af..30b3e5dd6 100644
--- a/src/Chemistry/SurrogateModels/DHT_Wrapper.cpp
+++ b/src/Chemistry/SurrogateModels/DHT_Wrapper.cpp
@@ -22,6 +22,7 @@
 
 #include "DHT_Wrapper.hpp"
 #include "DHT_ucx/UCX_bcast_functions.h"
+#include "HashFunctions.hpp"
 
 #include <algorithm>
 #include <cassert>
@@ -62,7 +63,7 @@ DHT_Wrapper::DHT_Wrapper(MPI_Comm dht_comm, std::uint64_t dht_size,
       .key_size = static_cast<int>(key_size),
       .data_size = static_cast<int>(data_size),
       .bucket_count = static_cast<unsigned int>(buckets_per_process),
-      .hash_func = &poet::Murmur2_64A,
+      .hash_func = &poet::md5_sum,
       .bcast_func = UCX_INIT_BSTRAP_MPI,
       .bcast_func_args = &ucx_bcast_mpi_args};
   dht_object = DHT_create(&dht_init);
diff --git a/src/Chemistry/SurrogateModels/HashFunctions.cpp b/src/Chemistry/SurrogateModels/HashFunctions.cpp
index d93702990..d76883f5b 100644
--- a/src/Chemistry/SurrogateModels/HashFunctions.cpp
+++ b/src/Chemistry/SurrogateModels/HashFunctions.cpp
@@ -25,6 +25,9 @@
 */
 
 #include "HashFunctions.hpp"
+#include <cstdint>
+
+#include <openssl/md5.h>
 
 #if defined(_MSC_VER)
 
@@ -93,3 +96,21 @@ uint64_t poet::Murmur2_64A(int len, const void *key) {
 
   return h;
 }
+
+uint64_t poet::md5_sum(int len, const void *key) {
+  MD5_CTX ctx;
+  unsigned char sum[MD5_DIGEST_LENGTH];
+  uint64_t retval;
+  uint64_t *v1;
+  uint64_t *v2;
+
+  MD5_Init(&ctx);
+  MD5_Update(&ctx, key, len);
+  MD5_Final(sum, &ctx);
+
+  v1 = (uint64_t *)&sum[0];
+  v2 = (uint64_t *)&sum[8];
+  retval = *v1 ^ *v2;
+
+  return retval;
+}
\ No newline at end of file
diff --git a/src/Chemistry/SurrogateModels/HashFunctions.hpp b/src/Chemistry/SurrogateModels/HashFunctions.hpp
index 49e86559a..438b86512 100644
--- a/src/Chemistry/SurrogateModels/HashFunctions.hpp
+++ b/src/Chemistry/SurrogateModels/HashFunctions.hpp
@@ -31,6 +31,7 @@ namespace poet {
 constexpr uint32_t HASH_SEED = 80 + 79 + 69 + 84;
 
 uint64_t Murmur2_64A(int len, const void *key);
+uint64_t md5_sum(int len, const void *key);
 
 } // namespace poet