From 25e05907f0292310eaae27a032db0ee274413874 Mon Sep 17 00:00:00 2001
From: "Claudius \"keldu\" Holeksa" <mail@keldu.de>
Date: Tue, 2 Jul 2024 19:46:02 +0200
Subject: Preparing benchmark work

---
 modules/codec/c++/data.hpp                         |  47 +++--
 modules/codec/c++/rpc.hpp                          |   5 +-
 modules/codec/c++/transfer.hpp                     |   8 +-
 modules/remote-sycl/.nix/derivation.nix            |   7 +-
 modules/remote-sycl/SConstruct                     |  10 +-
 modules/remote-sycl/benchmarks/SConscript          |  39 ++++
 .../benchmarks/kernel_mixed_precision.cpp          |  42 ++++
 modules/remote-sycl/benchmarks/mixed_precision.cpp | 110 +++++++++++
 modules/remote-sycl/benchmarks/mixed_precision.hpp |  27 +++
 modules/remote-sycl/c++/common.hpp                 |   7 +-
 modules/remote-sycl/c++/data.hpp                   |   8 +-
 modules/remote-sycl/c++/device.hpp                 |   4 +-
 modules/remote-sycl/c++/transfer.hpp               |   3 +-
 modules/remote-sycl/tests/mixed_precision.cpp      | 212 +++++++++++++++++++++
 modules/remote-sycl/tests/sycl_basics.cpp          |   9 -
 15 files changed, 488 insertions(+), 50 deletions(-)
 create mode 100644 modules/remote-sycl/benchmarks/SConscript
 create mode 100644 modules/remote-sycl/benchmarks/kernel_mixed_precision.cpp
 create mode 100644 modules/remote-sycl/benchmarks/mixed_precision.cpp
 create mode 100644 modules/remote-sycl/benchmarks/mixed_precision.hpp
 create mode 100644 modules/remote-sycl/tests/mixed_precision.cpp

(limited to 'modules')

diff --git a/modules/codec/c++/data.hpp b/modules/codec/c++/data.hpp
index 5ebe579..dd70cd9 100644
--- a/modules/codec/c++/data.hpp
+++ b/modules/codec/c++/data.hpp
@@ -86,8 +86,8 @@ private:
 	static_assert(always_false<T>, "Type not supported");
 };
 
-template<typename T, size_t N>
-class data<schema::Primitive<T,N>, encode::Native, storage::Default> {
+template<typename T, size_t N, typename Storage>
+class data<schema::Primitive<T,N>, encode::Native, Storage> {
 public:
 	using Schema = schema::Primitive<T,N>;
 private:
@@ -107,29 +107,29 @@ public:
 
 	typename native_data_type<Schema>::type get() const {return value_;}
 
-	data<Schema, encode::Native, storage::Default> operator*(const data<Schema, encode::Native, storage::Default>& rhs)const{
+	data<Schema, encode::Native, Storage> operator*(const data<Schema, encode::Native, Storage>& rhs)const{
 		return {get() * rhs.get()};
 	}
 	
-	data<Schema, encode::Native, storage::Default> operator/(const data<Schema, encode::Native, storage::Default>& rhs)const{
+	data<Schema, encode::Native, Storage> operator/(const data<Schema, encode::Native, Storage>& rhs)const{
 		return {get() / rhs.get()};
 	}
 	
-	data<Schema, encode::Native, storage::Default> operator+(const data<Schema, encode::Native, storage::Default>& rhs)const{
+	data<Schema, encode::Native, Storage> operator+(const data<Schema, encode::Native, Storage>& rhs)const{
 		return {get() + rhs.get()};
 	}
 	
-	data<Schema, encode::Native, storage::Default>& operator+=(const data<Schema, encode::Native, storage::Default>& rhs)const{
+	data<Schema, encode::Native, Storage>& operator+=(const data<Schema, encode::Native, Storage>& rhs)const{
 		value_ += rhs.get();
 		return *this;
 	}
 	
-	data<Schema, encode::Native, storage::Default>& operator-=(const data<Schema, encode::Native, storage::Default>& rhs)const{
+	data<Schema, encode::Native, Storage>& operator-=(const data<Schema, encode::Native, Storage>& rhs)const{
 		value_ -= rhs.get();
 		return *this;
 	}
 
-	data<Schema, encode::Native, storage::Default> operator-(const data<Schema, encode::Native, storage::Default>& rhs)const{
+	data<Schema, encode::Native, Storage> operator-(const data<Schema, encode::Native, Storage>& rhs)const{
 		return {get() - rhs.get()};
 	}
 
@@ -147,7 +147,7 @@ public:
 	 * Casts 
 	 */
 	template<typename Target>
-	data<Target, encode::Native, storage::Default> cast_to(){
+	data<Target, encode::Native, Storage> cast_to(){
 		auto raw_to = static_cast<typename saw::native_data_type<Target>::type>(value_);
 		return {raw_to};
 	}
@@ -156,12 +156,12 @@ public:
 /**
  * Mixed precision class for native formats
  */
-template<typename TA, uint64_t NA, typename TB, uint64_t NB> 
-class data<schema::MixedPrecision<schema::Primitive<TA,NA>, schema::Primitive<TB,NB>>, encode::Native, storage::Default>{
+template<typename TA, uint64_t NA, typename TB, uint64_t NB, typename Storage> 
+class data<schema::MixedPrecision<schema::Primitive<TA,NA>, schema::Primitive<TB,NB>>, encode::Native, Storage>{
 public:
 	using Schema = schema::MixedPrecision<schema::Primitive<TA,NA>, schema::Primitive<TB,NB>>;
 private:
-	data<typename Schema::StorageSchema, encode::Native, storage::Default> value_;
+	data<typename Schema::StorageSchema, encode::Native, Storage> value_;
 public:
 	data():value_{}{}
 
@@ -175,51 +175,58 @@ public:
 		value_.set(val.template cast_to<typename Schema::StorageSchema>().get());
 	}
 	
-	data<Schema, encode::Native, storage::Default> operator*(const data<Schema, encode::Native, storage::Default>& rhs) const {
+	data<Schema, encode::Native, Storage> operator*(const data<Schema, encode::Native, Storage>& rhs) const {
 		using CalcType = typename native_data_type<typename Schema::InterfaceSchema>::type;
 		CalcType left = static_cast<CalcType>(value_.get());
 		CalcType right = static_cast<CalcType>(rhs.get());
 		return {left * right};
 	}
 	
-	data<Schema, encode::Native, storage::Default> operator/(const data<Schema, encode::Native, storage::Default>& rhs)const{
+	data<Schema, encode::Native, Storage> operator*(const data<typename Schema::InterfaceSchema, encode::Native, Storage>& rhs) const {
+		using CalcType = typename native_data_type<typename Schema::InterfaceSchema>::type;
+		CalcType left = static_cast<CalcType>(value_.get());
+		CalcType right = rhs.get();
+		return {left * right};
+	}
+	
+	data<Schema, encode::Native, Storage> operator/(const data<Schema, encode::Native, Storage>& rhs)const{
 		using CalcType = typename native_data_type<typename Schema::InterfaceSchema>::type;
 		CalcType left = static_cast<CalcType>(value_.get());
 		CalcType right = static_cast<CalcType>(rhs.get());
 		return {left / right};
 	}
 	
-	data<Schema, encode::Native, storage::Default> operator+(const data<Schema, encode::Native, storage::Default>& rhs)const{
+	data<Schema, encode::Native, Storage> operator+(const data<Schema, encode::Native, Storage>& rhs)const{
 		using CalcType = typename native_data_type<typename Schema::InterfaceSchema>::type;
 		CalcType left = static_cast<CalcType>(value_.get());
 		CalcType right = static_cast<CalcType>(rhs.get());
 		return {left + right};
 	}
 	
-	data<Schema, encode::Native, storage::Default>& operator+=(const data<Schema, encode::Native, storage::Default>& rhs)const{
+	data<Schema, encode::Native, Storage>& operator+=(const data<Schema, encode::Native, Storage>& rhs)const{
 		*this = *this + rhs.get();
 		return *this;
 	}
 
-	data<Schema, encode::Native, storage::Default> operator-(const data<Schema, encode::Native, storage::Default>& rhs)const{
+	data<Schema, encode::Native, Storage> operator-(const data<Schema, encode::Native, Storage>& rhs)const{
 		using CalcType = typename native_data_type<typename Schema::InterfaceSchema>::type;
 		CalcType left = static_cast<CalcType>(value_.get());
 		CalcType right = static_cast<CalcType>(rhs.get());
 		return {left - right};
 	}
 	
-	data<Schema, encode::Native, storage::Default>& operator-=(const data<Schema, encode::Native, storage::Default>& rhs) const {
+	data<Schema, encode::Native, Storage>& operator-=(const data<Schema, encode::Native, Storage>& rhs) const {
 		*this = *this - rhs.get();
 		return *this;
 	}
 
 	template<typename Enc>
-	bool operator==(const data<Schema, Enc>& rhs)const{
+	bool operator==(const data<Schema, Enc, Storage>& rhs)const{
 		return get() == rhs.get();
 	}
 
 	template<typename Enc>
-	bool operator<(const data<Schema, Enc>& rhs) const {
+	bool operator<(const data<Schema, Enc, Storage>& rhs) const {
 		return get() < rhs.get();
 	}
 };
diff --git a/modules/codec/c++/rpc.hpp b/modules/codec/c++/rpc.hpp
index 2c97d6b..d172ec4 100644
--- a/modules/codec/c++/rpc.hpp
+++ b/modules/codec/c++/rpc.hpp
@@ -9,7 +9,6 @@
 #include <variant>
 
 namespace saw {
-
 /**
  * This class acts as a helper for rpc calls and representing data on the remote.
  */
@@ -139,7 +138,9 @@ template<typename Remote>
 class remote_address {
 	static_assert(always_false<Remote>, "Type of remote not supported");
 
-
+	/**
+	 *
+	 */
 };
 
 /**
diff --git a/modules/codec/c++/transfer.hpp b/modules/codec/c++/transfer.hpp
index b6aa977..1fb297e 100644
--- a/modules/codec/c++/transfer.hpp
+++ b/modules/codec/c++/transfer.hpp
@@ -1,9 +1,9 @@
 #pragma once
 
 namespace saw {
-template<typename T>
-class data_client;
-
-template<typename T>
+template<typename Schema, typename Encoding, typename Remote>
 class data_server;
+
+template<typename Schema, typename Encoding, typename Remote>
+class data_client;
 }
diff --git a/modules/remote-sycl/.nix/derivation.nix b/modules/remote-sycl/.nix/derivation.nix
index 71c5dff..ba6d1ec 100644
--- a/modules/remote-sycl/.nix/derivation.nix
+++ b/modules/remote-sycl/.nix/derivation.nix
@@ -12,6 +12,7 @@
 , bash
 
 , build_examples ? "false"
+, build_benchmarks ? "false"
 }:
 
 let
@@ -33,8 +34,6 @@ in stdenv.mkDerivation {
     forstio.core
 		forstio.codec
 		forstio.async
-		forstio.io
-    forstio.io_codec
 		keldu.adaptivecpp-dev
 		ocl-icd
 		openmp
@@ -42,11 +41,11 @@ in stdenv.mkDerivation {
 	];
 
 	buildPhase = ''
-		scons build_examples=${build_examples}
+		scons build_benchmarks=${build_benchmarks} build_examples=${build_examples}
 	'';
 
 	installPhase = ''
-		scons prefix=$out build_examples=${build_examples} install
+		scons prefix=$out build_benchmarks=${build_benchmarks} build_examples=${build_examples} install
 	'';
 
   doCheck = true;
diff --git a/modules/remote-sycl/SConstruct b/modules/remote-sycl/SConstruct
index 96e9df9..40344e7 100644
--- a/modules/remote-sycl/SConstruct
+++ b/modules/remote-sycl/SConstruct
@@ -50,14 +50,19 @@ env_vars.Add(
     )
 );
 
+env_vars.Add(
+    BoolVariable('build_benchmarks',
+        help='Build benchmarks',
+        default=False
+    )
+);
+
 env=Environment(ENV=os.environ, variables=env_vars, CPPPATH=[],
     CPPDEFINES=['SAW_UNIX'],
     CXXFLAGS=['-std=c++20','-g','-Wall','-Wextra'],
     LIBS=['forstio-core'
         ,'forstio-codec'
         ,'forstio-async'
-        ,'forstio-io'
-        ,'forstio-io_codec'
         ,'acpp-rt'
         ,'OpenCL'
         ,'omp'
@@ -76,6 +81,7 @@ Export('env')
 SConscript('c++/SConscript')
 SConscript('tests/SConscript')
 SConscript('examples/SConscript')
+SConscript('benchmarks/SConscript')
 
 env.Alias('cdb', env.cdb);
 env.Alias('all', [env.targets]);
diff --git a/modules/remote-sycl/benchmarks/SConscript b/modules/remote-sycl/benchmarks/SConscript
new file mode 100644
index 0000000..9976e0e
--- /dev/null
+++ b/modules/remote-sycl/benchmarks/SConscript
@@ -0,0 +1,39 @@
+#!/bin/false
+
+import os
+import os.path
+import glob
+
+
+Import('env')
+
+dir_path = Dir('.').abspath
+
+# Environment for base library
+benchmarks_env = env.Clone();
+
+
+benchmarks_sycl_env = benchmarks_env.Clone();
+benchmarks_sycl_env['CXX'] = 'acpp';
+benchmarks_sycl_env['CXXFLAGS'] += ['-O2'];
+
+benchmarks_env.sources = sorted(glob.glob(dir_path + "/*.cpp"))
+benchmarks_env.headers = sorted(glob.glob(dir_path + "/*.hpp"))
+
+env.sources += benchmarks_env.sources;
+env.headers += benchmarks_env.headers;
+
+sycl_objects = [];
+benchmarks_sycl_env.add_source_files(sycl_objects, ['kernel_mixed_precision.cpp'], shared=True);
+
+objects_static = []
+benchmarks_env.sycl_basic = benchmarks_env.Program('#bin/benchmark_mixed_precision', ['mixed_precision.cpp', env.library_static, sycl_objects]);
+
+# Set Alias
+env.benchmarks = [benchmarks_env.sycl_basic];
+env.Alias('benchmarks', env.benchmarks);
+
+if env["build_benchmarks"]:
+	env.targets += ['benchmarks'];
+	env.Install('$prefix/bin/', env.benchmarks);
+#endif
diff --git a/modules/remote-sycl/benchmarks/kernel_mixed_precision.cpp b/modules/remote-sycl/benchmarks/kernel_mixed_precision.cpp
new file mode 100644
index 0000000..0ac9756
--- /dev/null
+++ b/modules/remote-sycl/benchmarks/kernel_mixed_precision.cpp
@@ -0,0 +1,42 @@
+#include "mixed_precision.hpp"
+
+saw::interface<sch::MixedPrecisionBenchmarkInterface, saw::encode::Native, saw::rmt::Sycl, cl::sycl::queue*> listen_mixed_precision(cl::sycl::event& mixed_ev, cl::sycl::event& float64_ev, cl::sycl::event& float32_ev){
+	return {
+		/**
+		 * Mixed
+		 */
+		[&](saw::data<sch::MixedArray, saw::encode::Native, saw::rmt::Sycl>& in, cl::sycl::queue* cmd) -> saw::error_or<void> {
+	
+			mixed_ev = cmd->submit([&](cl::sycl::handler& h){
+				auto acc_buff = in.template access<cl::sycl::access::mode::read_write>(h);
+
+				h.parallel_for(cl::sycl::range<1>(in.size()), [=] (cl::sycl::id<1> it){
+					acc_buff[0u].at(it[0u]) = acc_buff[0u].at(it[0u]) * saw::data<sch::Float64>{2.0};
+				});
+			});
+			return saw::void_t{};
+		},
+		[&](saw::data<sch::Float64Array, saw::encode::Native, saw::rmt::Sycl>& in, cl::sycl::queue* cmd) -> saw::error_or<void> {
+	
+			float64_ev = cmd->submit([&](cl::sycl::handler& h){
+				auto acc_buff = in.template access<cl::sycl::access::mode::read_write>(h);
+
+				h.parallel_for(cl::sycl::range<1>(in.size()), [=] (cl::sycl::id<1> it){
+					acc_buff[0u].at(it[0u]) = acc_buff[0u].at(it[0u]) * saw::data<sch::Float64>{2.0};
+				});
+			});
+			return saw::void_t{};
+		},
+		[&](saw::data<sch::Float32Array, saw::encode::Native, saw::rmt::Sycl>& in, cl::sycl::queue* cmd) -> saw::error_or<void> {
+	
+			float32_ev = cmd->submit([&](cl::sycl::handler& h){
+				auto acc_buff = in.template access<cl::sycl::access::mode::read_write>(h);
+
+				h.parallel_for(cl::sycl::range<1>(in.size()), [=] (cl::sycl::id<1> it){
+					acc_buff[0u].at(it[0u]) = acc_buff[0u].at(it[0u]) * saw::data<sch::Float32>{2.0f};
+				});
+			});
+			return saw::void_t{};
+		}
+	};
+}
diff --git a/modules/remote-sycl/benchmarks/mixed_precision.cpp b/modules/remote-sycl/benchmarks/mixed_precision.cpp
new file mode 100644
index 0000000..e63a814
--- /dev/null
+++ b/modules/remote-sycl/benchmarks/mixed_precision.cpp
@@ -0,0 +1,110 @@
+#include "./mixed_precision.hpp"
+#include <forstio/codec/schema.hpp>
+
+
+int main(){
+	using namespace saw;
+
+	constexpr uint64_t max_test_size = 1024ul * 1024ul * 512ul;
+
+	std::random_device r;
+	std::default_random_engine e1{r()};
+	std::uniform_real_distribution<> dis{-1.0,1.0};
+
+
+	saw::event_loop loop;
+	saw::wait_scope wait{loop};
+
+	remote<rmt::Sycl> rmt;
+
+	own<remote_address<rmt::Sycl>> rmt_addr{};
+
+	rmt.resolve_address().then([&](auto addr){
+		rmt_addr = std::move(addr);
+	}).detach();
+
+	wait.poll();
+	if(!rmt_addr){
+		return -1;
+	}
+	
+	data<sch::MixedArray> mixed_host_data;
+	data<sch::Float64Array> float64_host_data;
+	data<sch::Float32Array> float32_host_data;
+	
+	cl::sycl::event mixed_ev;
+	cl::sycl::event float32_ev;
+	cl::sycl::event float64_ev;
+
+	auto sycl_iface = listen_mixed_precision(mixed_ev, float64_ev, float32_ev);
+	
+	auto time_eval = [](std::string_view tag, cl::sycl::event& ev){
+		auto end = ev.get_profiling_info<cl::sycl::info::event_profiling::command_end>();
+		auto start = ev.get_profiling_info<cl::sycl::info::event_profiling::command_start>();
+
+		std::cout<<"Elapsed "<<tag<<" kernel time: "<< (end-start) / 1.0e9 << " seconds\n";
+	};
+	
+	auto& device = rmt_addr->get_device();
+	
+	/**
+	 * Warmup
+	 */
+	{
+		uint64_t test_size = max_test_size;
+		mixed_host_data = {test_size};
+		float64_host_data = {test_size};
+		float32_host_data = {test_size};
+		for(uint64_t i = 0; i < test_size; ++i){
+			double gen_num = dis(e1);
+			mixed_host_data.at(i) = static_cast<double>(gen_num);
+			float64_host_data.at(i) = static_cast<double>(gen_num);
+			float32_host_data.at(i) = static_cast<float>(gen_num);
+		}
+		data<sch::MixedArray, encode::Native, rmt::Sycl> mixed_device_data{mixed_host_data};
+		data<sch::Float64Array, encode::Native, rmt::Sycl> float64_device_data{float64_host_data};
+		data<sch::Float32Array, encode::Native, rmt::Sycl> float32_device_data{float32_host_data};
+		
+		sycl_iface.template call<"float64_32">(mixed_device_data, &(device.get_handle()));
+		device.get_handle().wait();
+		sycl_iface.template call<"float64">(float64_device_data, &(device.get_handle()));
+		device.get_handle().wait();
+		sycl_iface.template call<"float32">(float32_device_data, &(device.get_handle()));
+		device.get_handle().wait();
+
+	}
+
+	/**
+	 * Benchmark
+	 */
+	for(uint64_t test_size = 1ul; test_size < max_test_size; test_size *= 2ul){
+		device.get_handle().wait();
+		mixed_host_data = {test_size};
+		float64_host_data = {test_size};
+		float32_host_data = {test_size};
+		for(uint64_t i = 0; i < test_size; ++i){
+			double gen_num = dis(e1);
+			mixed_host_data.at(i) = static_cast<double>(gen_num);
+			float64_host_data.at(i) = static_cast<double>(gen_num);
+			float32_host_data.at(i) = static_cast<float>(gen_num);
+		}
+		data<sch::MixedArray, encode::Native, rmt::Sycl> mixed_device_data{mixed_host_data};
+		data<sch::Float64Array, encode::Native, rmt::Sycl> float64_device_data{float64_host_data};
+		data<sch::Float32Array, encode::Native, rmt::Sycl> float32_device_data{float32_host_data};
+		
+		sycl_iface.template call<"float64_32">(mixed_device_data);
+		device.get_handle().wait();
+		sycl_iface.template call<"float64">(float64_device_data);
+		device.get_handle().wait();
+		sycl_iface.template call<"float32">(float32_device_data);
+		device.get_handle().wait();
+		
+		std::cout<<"\nSize: "<<test_size<<'\n';
+		time_eval("Mixed", mixed_ev);
+		time_eval("Float32", float32_ev);
+		time_eval("Float64", float64_ev);
+	}
+	std::cout<<std::endl;
+
+	return 0;
+}
diff --git a/modules/remote-sycl/benchmarks/mixed_precision.hpp b/modules/remote-sycl/benchmarks/mixed_precision.hpp
new file mode 100644
index 0000000..3462bcd
--- /dev/null
+++ b/modules/remote-sycl/benchmarks/mixed_precision.hpp
@@ -0,0 +1,27 @@
+#pragma once
+
+#include "../c++/remote.hpp"
+
+namespace sch {
+using namespace saw::schema;
+
+using MixedArray = Array<
+	MixedPrecision<Float64, Float32>
+>;
+
+using Float64Array = Array<
+	Float64
+>;
+
+using Float32Array = Array<
+	Float32
+>;
+
+using MixedPrecisionBenchmarkInterface = Interface<
+	Member<Function<MixedArray,Void>,   "float64_32">,
+	Member<Function<Float64Array,Void>, "float64">,
+	Member<Function<Float32Array,Void>, "float32">
+>;
+}
+
+saw::interface<sch::MixedPrecisionBenchmarkInterface, saw::encode::Native, saw::rmt::Sycl, cl::sycl::queue*> listen_mixed_precision(cl::sycl::event& mixed_ev, cl::sycl::event& float64_ev, cl::sycl::event& float32_ev);
diff --git a/modules/remote-sycl/c++/common.hpp b/modules/remote-sycl/c++/common.hpp
index 9bc734d..078879f 100644
--- a/modules/remote-sycl/c++/common.hpp
+++ b/modules/remote-sycl/c++/common.hpp
@@ -1,6 +1,6 @@
 #pragma once
 
-#include <forstio/io_codec/rpc.hpp>
+#include <forstio/codec/rpc.hpp>
 #include <forstio/codec/data.hpp>
 #include <forstio/codec/id_map.hpp>
 
@@ -17,9 +17,4 @@ class remote<rmt::Sycl>;
 template<typename T>
 class device;
 
-template<typename Schema, typename Encoding, typename Remote>
-class data_server;
-
-template<typename Schema, typename Encoding, typename Remote>
-class data_client;
 }
diff --git a/modules/remote-sycl/c++/data.hpp b/modules/remote-sycl/c++/data.hpp
index e436e72..d939a53 100644
--- a/modules/remote-sycl/c++/data.hpp
+++ b/modules/remote-sycl/c++/data.hpp
@@ -12,9 +12,11 @@ template<typename Schema>
 class data<Schema, encode::Native, rmt::Sycl> {
 private:
 	cl::sycl::buffer<data<Schema, encode::Native, storage::Default>> data_;
+	uint64_t size_;
 public:
 	data(const data<Schema, encode::Native, storage::Default>& data__):
-		data_{&data__, 1u}
+		data_{&data__, 1u},
+		size_{data__.size()}
 	{}
 
 	auto& get_handle() {
@@ -25,6 +27,10 @@ public:
 		return data_;
 	}
 
+	uint64_t size() const {
+		return size_;
+	}
+
 	template<cl::sycl::access::mode AccessMode>
 	auto access(cl::sycl::handler& h){
 		return data_.template get_access<AccessMode>(h);
diff --git a/modules/remote-sycl/c++/device.hpp b/modules/remote-sycl/c++/device.hpp
index e6f3fe3..ae19524 100644
--- a/modules/remote-sycl/c++/device.hpp
+++ b/modules/remote-sycl/c++/device.hpp
@@ -11,7 +11,9 @@ class device<rmt::Sycl> final {
 private:
 	cl::sycl::queue cmd_queue_;
 public:
-	device() = default;
+	device():
+		cmd_queue_{cl::sycl::default_selector_v, cl::sycl::property_list{cl::sycl::property::queue::enable_profiling()}}
+	{}
 
 	SAW_FORBID_COPY(device);
 	SAW_FORBID_MOVE(device);
diff --git a/modules/remote-sycl/c++/transfer.hpp b/modules/remote-sycl/c++/transfer.hpp
index f535751..72b111f 100644
--- a/modules/remote-sycl/c++/transfer.hpp
+++ b/modules/remote-sycl/c++/transfer.hpp
@@ -6,6 +6,7 @@
 
 #include <forstio/error.hpp>
 #include <forstio/reduce_templates.hpp>
+#include <forstio/codec/transfer.hpp>
 
 namespace saw {
 namespace impl {
@@ -31,7 +32,7 @@ private:
 	/**
 	 * Store for the data the server manages.
 	 */
-	impl::data_server_redux<Encoding, typename tmpl_reduce<tmpl_group<Schema...>>::type >::type values_;
+	typename impl::data_server_redux<Encoding, typename tmpl_reduce<tmpl_group<Schema...>>::type >::type values_;
 public:
 	/**
 	 * Main constructor
diff --git a/modules/remote-sycl/tests/mixed_precision.cpp b/modules/remote-sycl/tests/mixed_precision.cpp
new file mode 100644
index 0000000..4a62569
--- /dev/null
+++ b/modules/remote-sycl/tests/mixed_precision.cpp
@@ -0,0 +1,212 @@
+#include <forstio/test/suite.hpp>
+
+#include "../c++/remote.hpp"
+
+#include <random>
+
+namespace {
+namespace schema {
+using namespace saw::schema;
+
+using TestMixedArray = Array<
+	MixedPrecision<Float64,Float32>
+>;
+
+using MixedFoo = Interface<
+	Member<Function<TestMixedArray, Void>, "foo">
+>;
+
+using TestDoubleArray = Array<
+	Float64
+>;
+
+using DoubleFoo = Interface<
+	Member<Function<TestDoubleArray, Void>, "foo">
+>;
+
+using TestFloatArray = Array<
+	Float32
+>;
+
+using FloatFoo = Interface<
+	Member<Function<TestFloatArray, Void>, "foo">
+>;
+}
+
+constexpr uint64_t test_size = 1024ul;
+
+SAW_TEST("SYCL Mixed Test"){
+	using namespace saw;
+
+	std::random_device r;
+	std::default_random_engine e1{r()};
+	std::uniform_real_distribution<> dis{-1.0,1.0};
+
+	data<schema::TestMixedArray> host_data;
+	host_data = {test_size};
+	for(uint64_t i = 0; i < test_size; ++i){
+		host_data.at(i) = static_cast<double>(dis(e1));
+	}
+
+	saw::event_loop loop;
+	saw::wait_scope wait{loop};
+
+	remote<rmt::Sycl> rmt;
+
+	own<remote_address<rmt::Sycl>> rmt_addr{};
+
+	rmt.resolve_address().then([&](auto addr){
+		rmt_addr = std::move(addr);
+	}).detach();
+
+	wait.poll();
+	SAW_EXPECT(rmt_addr, "Remote address hasn't been filled");
+	
+	data<schema::TestMixedArray, encode::Native, rmt::Sycl> device_data{host_data};
+	
+	cl::sycl::event ev;
+
+	interface<schema::MixedFoo, encode::Native,rmt::Sycl, cl::sycl::queue*> cl_iface {
+[&](data<schema::TestMixedArray, encode::Native, rmt::Sycl>& in, cl::sycl::queue* cmd) -> error_or<void> {
+	
+			ev = cmd->submit([&](cl::sycl::handler& h){
+
+				auto acc_buff = in.template access<cl::sycl::access::mode::read_write>(h);
+
+				h.parallel_for(cl::sycl::range<1>(test_size), [=] (cl::sycl::id<1> it){
+					acc_buff[0u].at(it[0u]) = acc_buff[0u].at(it[0u]) * data<schema::Float64>{2.0};
+				});
+			});
+			return saw::void_t{};
+		}
+	};
+	auto& device = rmt_addr->get_device();
+	
+	cl_iface.template call <"foo">(device_data, &(device.get_handle()));
+	device.get_handle().wait();
+	
+	{
+		auto end = ev.get_profiling_info<cl::sycl::info::event_profiling::command_end>();
+		auto start = ev.get_profiling_info<cl::sycl::info::event_profiling::command_start>();
+
+		std::cout<<"Elapsed kernel time: "<< (end-start) / 1.0e9 << " seconds"<<std::endl;
+	}
+}
+
+SAW_TEST("SYCL Float Test"){
+	using namespace saw;
+
+	std::random_device r;
+	std::default_random_engine e1{r()};
+	std::uniform_real_distribution<> dis{-1.0,1.0};
+
+	data<schema::TestFloatArray> host_data;
+	host_data = {test_size};
+	for(uint64_t i = 0; i < test_size; ++i){
+		host_data.at(i) = static_cast<double>(dis(e1));
+	}
+
+	saw::event_loop loop;
+	saw::wait_scope wait{loop};
+
+	remote<rmt::Sycl> rmt;
+
+	own<remote_address<rmt::Sycl>> rmt_addr{};
+
+	rmt.resolve_address().then([&](auto addr){
+		rmt_addr = std::move(addr);
+	}).detach();
+
+	wait.poll();
+	SAW_EXPECT(rmt_addr, "Remote address hasn't been filled");
+	
+	data<schema::TestFloatArray, encode::Native, rmt::Sycl> device_data{host_data};
+
+	cl::sycl::event ev;
+
+	interface<schema::FloatFoo, encode::Native,rmt::Sycl, cl::sycl::queue*> cl_iface {
+[&](data<schema::TestFloatArray, encode::Native, rmt::Sycl>& in, cl::sycl::queue* cmd) -> error_or<void> {
+	
+			ev = cmd->submit([&](cl::sycl::handler& h){
+
+				auto acc_buff = in.template access<cl::sycl::access::mode::read_write>(h);
+
+				h.parallel_for(cl::sycl::range<1>(test_size), [=] (cl::sycl::id<1> it){
+					acc_buff[0u].at(it[0u]) = acc_buff[0u].at(it[0u]) * data<schema::Float32>{2.0};
+				});
+			});
+			return saw::void_t{};
+		}
+	};
+	auto& device = rmt_addr->get_device();
+	
+	cl_iface.template call <"foo">(device_data, &(device.get_handle()));
+	device.get_handle().wait();
+
+	{
+		auto end = ev.get_profiling_info<cl::sycl::info::event_profiling::command_end>();
+		auto start = ev.get_profiling_info<cl::sycl::info::event_profiling::command_start>();
+
+		std::cout<<"Elapsed kernel time: "<< (end-start) / 1.0e9 << " seconds"<<std::endl;
+	}
+}
+
+SAW_TEST("SYCL Double Test"){
+	using namespace saw;
+
+	std::random_device r;
+	std::default_random_engine e1{r()};
+	std::uniform_real_distribution<> dis{-1.0,1.0};
+
+	data<schema::TestDoubleArray> host_data;
+	host_data = {test_size};
+	for(uint64_t i = 0; i < test_size; ++i){
+		host_data.at(i) = static_cast<double>(dis(e1));
+	}
+
+	saw::event_loop loop;
+	saw::wait_scope wait{loop};
+
+	remote<rmt::Sycl> rmt;
+
+	own<remote_address<rmt::Sycl>> rmt_addr{};
+
+	rmt.resolve_address().then([&](auto addr){
+		rmt_addr = std::move(addr);
+	}).detach();
+
+	wait.poll();
+	SAW_EXPECT(rmt_addr, "Remote address hasn't been filled");
+	
+	data<schema::TestDoubleArray, encode::Native, rmt::Sycl> device_data{host_data};
+	
+	cl::sycl::event ev;
+
+	interface<schema::DoubleFoo, encode::Native,rmt::Sycl, cl::sycl::queue*> cl_iface {
+[&](data<schema::TestDoubleArray, encode::Native, rmt::Sycl>& in, cl::sycl::queue* cmd) -> error_or<void> {
+	
+			ev = cmd->submit([&](cl::sycl::handler& h){
+
+				auto acc_buff = in.template access<cl::sycl::access::mode::read_write>(h);
+
+				h.parallel_for(cl::sycl::range<1>(test_size), [=] (cl::sycl::id<1> it){
+					acc_buff[0u].at(it[0u]) = acc_buff[0u].at(it[0u]) * data<schema::Float64>{2.0};
+				});
+			});
+			return saw::void_t{};
+		}
+	};
+	auto& device = rmt_addr->get_device();
+	
+	cl_iface.template call <"foo">(device_data, &(device.get_handle()));
+	device.get_handle().wait();
+	
+	{
+		auto end = ev.get_profiling_info<cl::sycl::info::event_profiling::command_end>();
+		auto start = ev.get_profiling_info<cl::sycl::info::event_profiling::command_start>();
+
+		std::cout<<"Elapsed kernel time: "<< (end-start) / 1.0e9 << " seconds"<<std::endl;
+	}
+}
+
+}
diff --git a/modules/remote-sycl/tests/sycl_basics.cpp b/modules/remote-sycl/tests/sycl_basics.cpp
index 61e0d87..7212a77 100644
--- a/modules/remote-sycl/tests/sycl_basics.cpp
+++ b/modules/remote-sycl/tests/sycl_basics.cpp
@@ -15,15 +15,6 @@ using TestStruct = Struct<
 using Foo = Interface<
 	Member<Function<TestStruct, Void>, "foo">
 >;
-
-using Calculator = Interface<
-	Member<
-		Function<Tuple<Int64, Int64>, Int64>, "add"
-	>
-,	Member<
-		Function<Tuple<Int64, Int64>, Int64>, "multiply"
-	>
->;
 }
 SAW_TEST("SYCL Test Setup"){
 	using namespace saw;
-- 
cgit v1.2.3