mirror of
https://gitlab.com/libeigen/eigen.git
synced 2026-04-10 11:34:33 +08:00
Converting all sycl buffers to uninitialised device only buffers; adding memcpyHostToDevice and memcpyDeviceToHost on syclDevice; modifying all examples to obey the new rules; moving sycl queue creating to the device based on Benoit suggestion; removing the sycl specefic condition for returning m_result in TensorReduction.h according to Benoit suggestion.
This commit is contained in:
Mehdi Goli
@@ -16,95 +16,93 @@
|
||||
#define EIGEN_CXX11_TENSOR_TENSOR_DEVICE_SYCL_H
|
||||
|
||||
namespace Eigen {
|
||||
/// \struct BufferT is used to specialise add_sycl_buffer function for
|
||||
// two types of buffer we have. When the MapAllocator is true, we create the
|
||||
// sycl buffer with MapAllocator.
|
||||
/// We have to const_cast the input pointer in order to work around the fact
|
||||
/// that sycl does not accept map allocator for const pointer.
|
||||
template <typename T, bool MapAllocator>
|
||||
struct BufferT {
|
||||
using Type = cl::sycl::buffer<T, 1, cl::sycl::map_allocator<T>>;
|
||||
static inline void add_sycl_buffer(const T *ptr, size_t num_bytes,std::map<const void *, std::shared_ptr<void>> &buffer_map) {
|
||||
buffer_map.insert(std::pair<const void *, std::shared_ptr<void>>(ptr, std::shared_ptr<void>(std::make_shared<Type>(Type(const_cast<T *>(ptr), cl::sycl::range<1>(num_bytes))))));
|
||||
}
|
||||
};
|
||||
|
||||
/// specialisation of the \ref BufferT when the MapAllocator is false. In this
|
||||
/// case we only create the device-only buffer.
|
||||
template <typename T>
|
||||
struct BufferT<T, false> {
|
||||
using Type = cl::sycl::buffer<T, 1>;
|
||||
static inline void add_sycl_buffer(const T *ptr, size_t num_bytes, std::map<const void *, std::shared_ptr<void>> &buffer_map) {
|
||||
buffer_map.insert(std::pair<const void *, std::shared_ptr<void>>(ptr, std::shared_ptr<void>(std::make_shared<Type>(Type(cl::sycl::range<1>(num_bytes))))));
|
||||
}
|
||||
};
|
||||
|
||||
struct SyclDevice {
|
||||
/// class members
|
||||
/// sycl queue
|
||||
cl::sycl::queue &m_queue;
|
||||
mutable cl::sycl::queue m_queue;
|
||||
/// std::map is the container used to make sure that we create only one buffer
|
||||
/// per pointer. The lifespan of the buffer
|
||||
/// now depends on the lifespan of SyclDevice. If a non-read-only pointer is
|
||||
/// needed to be accessed on the host we should manually deallocate it.
|
||||
/// per pointer. The lifespan of the buffer now depends on the lifespan of SyclDevice.
|
||||
/// If a non-read-only pointer is needed to be accessed on the host we should manually deallocate it.
|
||||
mutable std::map<const void *, std::shared_ptr<void>> buffer_map;
|
||||
|
||||
SyclDevice(cl::sycl::queue &q) : m_queue(q) {}
|
||||
/// creating device by using selector
|
||||
template<typename dev_Selector> SyclDevice(dev_Selector s)
|
||||
:m_queue(cl::sycl::queue(s, [=](cl::sycl::exception_list l) {
|
||||
for (const auto& e : l) {
|
||||
try {
|
||||
std::rethrow_exception(e);
|
||||
} catch (cl::sycl::exception e) {
|
||||
std::cout << e.what() << std::endl;
|
||||
}
|
||||
}
|
||||
})) {}
|
||||
// destructor
|
||||
~SyclDevice() { deallocate_all(); }
|
||||
|
||||
template <typename T>
|
||||
void deallocate(const T *p) const {
|
||||
template <typename T> void deallocate(T *p) const {
|
||||
auto it = buffer_map.find(p);
|
||||
if (it != buffer_map.end()) {
|
||||
buffer_map.erase(it);
|
||||
internal::aligned_free(p);
|
||||
}
|
||||
}
|
||||
void deallocate_all() const { buffer_map.clear(); }
|
||||
void deallocate_all() const {
|
||||
std::map<const void *, std::shared_ptr<void>>::iterator it=buffer_map.begin();
|
||||
while (it!=buffer_map.end()) {
|
||||
auto p=it->first;
|
||||
buffer_map.erase(it);
|
||||
internal::aligned_free(const_cast<void*>(p));
|
||||
it=buffer_map.begin();
|
||||
}
|
||||
buffer_map.clear();
|
||||
}
|
||||
|
||||
/// creation of sycl accessor for a buffer. This function first tries to find
|
||||
/// the buffer in the buffer_map.
|
||||
/// If found it gets the accessor from it, if not, the function then adds an
|
||||
/// entry by creating a sycl buffer
|
||||
/// for that particular pointer.
|
||||
template <cl::sycl::access::mode AcMd, bool MapAllocator, typename T>
|
||||
inline cl::sycl::accessor<T, 1, AcMd, cl::sycl::access::target::global_buffer>
|
||||
/// the buffer in the buffer_map. If found it gets the accessor from it, if not,
|
||||
///the function then adds an entry by creating a sycl buffer for that particular pointer.
|
||||
template <cl::sycl::access::mode AcMd, typename T> inline cl::sycl::accessor<T, 1, AcMd, cl::sycl::access::target::global_buffer>
|
||||
get_sycl_accessor(size_t num_bytes, cl::sycl::handler &cgh, const T * ptr) const {
|
||||
return (get_sycl_buffer<MapAllocator,T>(num_bytes, ptr).template get_access<AcMd, cl::sycl::access::target::global_buffer>(cgh));
|
||||
return (get_sycl_buffer<T>(num_bytes, ptr)->template get_access<AcMd, cl::sycl::access::target::global_buffer>(cgh));
|
||||
}
|
||||
|
||||
template <bool MapAllocator, typename T>
|
||||
inline typename BufferT<T, MapAllocator>::Type
|
||||
get_sycl_buffer(size_t num_bytes,const T * ptr) const {
|
||||
if(MapAllocator && !ptr){
|
||||
eigen_assert("pointer with map_Allocator cannot be null. Please initialise the input pointer"); }
|
||||
auto it = buffer_map.find(ptr);
|
||||
if (it == buffer_map.end()) {
|
||||
BufferT<T, MapAllocator>::add_sycl_buffer(ptr, num_bytes, buffer_map);
|
||||
}
|
||||
return (*((typename BufferT<T, MapAllocator>::Type*)((buffer_map.at(ptr).get()))));
|
||||
template<typename T> inline std::pair<std::map<const void *, std::shared_ptr<void>>::iterator,bool> add_sycl_buffer(const T *ptr, size_t num_bytes) const {
|
||||
using Type = cl::sycl::buffer<T, 1>;
|
||||
std::pair<std::map<const void *, std::shared_ptr<void>>::iterator,bool> ret = buffer_map.insert(std::pair<const void *, std::shared_ptr<void>>(ptr, std::shared_ptr<void>(new Type(cl::sycl::range<1>(num_bytes)),
|
||||
[](void *dataMem) { delete static_cast<Type*>(dataMem); })));
|
||||
(static_cast<Type*>(buffer_map.at(ptr).get()))->set_final_data(nullptr);
|
||||
return ret;
|
||||
}
|
||||
|
||||
template <typename T> inline cl::sycl::buffer<T, 1>* get_sycl_buffer(size_t num_bytes,const T * ptr) const {
|
||||
return static_cast<cl::sycl::buffer<T, 1>*>(add_sycl_buffer(ptr, num_bytes).first->second.get());
|
||||
}
|
||||
|
||||
/// allocating memory on the cpu
|
||||
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void *allocate(size_t num_bytes) const {
|
||||
return internal::aligned_malloc(num_bytes);
|
||||
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void *allocate(size_t) const {
|
||||
return internal::aligned_malloc(8);
|
||||
}
|
||||
|
||||
// some runtime conditions that can be applied here
|
||||
bool isDeviceSuitable() const { return true; }
|
||||
|
||||
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void deallocate(void *buffer) const {
|
||||
internal::aligned_free(buffer);
|
||||
}
|
||||
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void memcpy(void *dst, const void *src, size_t n) const {
|
||||
::memcpy(dst, src, n);
|
||||
}
|
||||
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void memcpyHostToDevice(void *dst, const void *src, size_t n) const {
|
||||
memcpy(dst, src, n);
|
||||
|
||||
template<typename T> EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void memcpyHostToDevice(T *dst, const T *src, size_t n) const {
|
||||
auto host_acc= (static_cast<cl::sycl::buffer<T, 1>*>(add_sycl_buffer(dst, n).first->second.get()))-> template get_access<cl::sycl::access::mode::discard_write, cl::sycl::access::target::host_buffer>();
|
||||
memcpy(host_acc.get_pointer(), src, n);
|
||||
}
|
||||
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void memcpyDeviceToHost(void *dst, const void *src, size_t n) const {
|
||||
memcpy(dst, src, n);
|
||||
/// whith the current implementation of sycl, the data is copied twice from device to host. This will be fixed soon.
|
||||
template<typename T> EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void memcpyDeviceToHost(T *dst, const T *src, size_t n) const {
|
||||
auto it = buffer_map.find(src);
|
||||
if (it != buffer_map.end()) {
|
||||
auto host_acc= (static_cast<cl::sycl::buffer<T, 1>*>(it->second.get()))-> template get_access<cl::sycl::access::mode::read, cl::sycl::access::target::host_buffer>();
|
||||
memcpy(dst,host_acc.get_pointer(), n);
|
||||
} else{
|
||||
eigen_assert("no device memory found. The memory might be destroyed before creation");
|
||||
}
|
||||
}
|
||||
|
||||
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void memset(void *buffer, int c, size_t n) const {
|
||||
::memset(buffer, c, n);
|
||||
}
|
||||
@@ -112,6 +110,7 @@ template <bool MapAllocator, typename T>
|
||||
return 1;
|
||||
}
|
||||
};
|
||||
|
||||
} // end namespace Eigen
|
||||
|
||||
#endif // EIGEN_CXX11_TENSOR_TENSOR_DEVICE_SYCL_H
|
||||
|
||||
@@ -662,13 +662,7 @@ struct TensorEvaluator<const TensorReductionOp<Op, Dims, ArgType, MakePointer_>,
|
||||
}
|
||||
}
|
||||
|
||||
/// required by sycl in order to extract the output accessor
|
||||
#ifndef EIGEN_USE_SYCL
|
||||
EIGEN_DEVICE_FUNC typename MakePointer_<Scalar>::Type data() const { return NULL; }
|
||||
#else
|
||||
EIGEN_DEVICE_FUNC typename MakePointer_<Scalar>::Type data() const {
|
||||
return m_result; }
|
||||
#endif
|
||||
EIGEN_DEVICE_FUNC typename MakePointer_<Scalar>::Type data() const { return m_result; }
|
||||
/// required by sycl in order to extract the accessor
|
||||
const TensorEvaluator<ArgType, Device>& impl() const { return m_impl; }
|
||||
/// added for sycl in order to construct the buffer from the sycl device
|
||||
|
||||
@@ -27,9 +27,9 @@ namespace internal {
|
||||
|
||||
template<typename CoeffReturnType, typename KernelName> struct syclGenericBufferReducer{
|
||||
template<typename BufferTOut, typename BufferTIn>
|
||||
static void run(BufferTOut& bufOut, BufferTIn& bufI, const Eigen::SyclDevice& dev, size_t length, size_t local){
|
||||
static void run(BufferTOut* bufOut, BufferTIn& bufI, const Eigen::SyclDevice& dev, size_t length, size_t local){
|
||||
do {
|
||||
auto f = [length, local, &bufOut, &bufI](cl::sycl::handler& h) mutable {
|
||||
auto f = [length, local, bufOut, &bufI](cl::sycl::handler& h) mutable {
|
||||
cl::sycl::nd_range<1> r{cl::sycl::range<1>{std::max(length, local)},
|
||||
cl::sycl::range<1>{std::min(length, local)}};
|
||||
/* Two accessors are used: one to the buffer that is being reduced,
|
||||
@@ -37,7 +37,7 @@ static void run(BufferTOut& bufOut, BufferTIn& bufI, const Eigen::SyclDevice& de
|
||||
auto aI =
|
||||
bufI.template get_access<cl::sycl::access::mode::read_write>(h);
|
||||
auto aOut =
|
||||
bufOut.template get_access<cl::sycl::access::mode::discard_write>(h);
|
||||
bufOut->template get_access<cl::sycl::access::mode::discard_write>(h);
|
||||
cl::sycl::accessor<CoeffReturnType, 1, cl::sycl::access::mode::read_write,
|
||||
cl::sycl::access::target::local>
|
||||
scratch(cl::sycl::range<1>(local), h);
|
||||
@@ -134,7 +134,7 @@ struct FullReducer<Self, Op, const Eigen::SyclDevice, Vectorizable> {
|
||||
/// if the shared memory is less than the GRange, we set shared_mem size to the TotalSize and in this case one kernel would be created for recursion to reduce all to one.
|
||||
if (GRange < outTileSize) outTileSize=GRange;
|
||||
// getting final out buffer at the moment the created buffer is true because there is no need for assign
|
||||
auto out_buffer =dev.template get_sycl_buffer<true, typename Eigen::internal::remove_all<CoeffReturnType>::type>(self.dimensions().TotalSize(), output);
|
||||
auto out_buffer =dev.template get_sycl_buffer<typename Eigen::internal::remove_all<CoeffReturnType>::type>(self.dimensions().TotalSize(), output);
|
||||
/// creating the shared memory for calculating reduction.
|
||||
/// This one is used to collect all the reduced value of shared memory as we dont have global barrier on GPU. Once it is saved we can
|
||||
/// recursively apply reduction on it in order to reduce the whole.
|
||||
@@ -208,7 +208,7 @@ struct InnerReducer<Self, Op, const Eigen::SyclDevice> {
|
||||
dev.m_queue.submit([&](cl::sycl::handler &cgh) {
|
||||
// create a tuple of accessors from Evaluator
|
||||
auto tuple_of_accessors = TensorSycl::internal::createTupleOfAccessors(cgh, self.impl());
|
||||
auto output_accessor = dev.template get_sycl_accessor<cl::sycl::access::mode::discard_write, true>(num_coeffs_to_preserve,cgh, output);
|
||||
auto output_accessor = dev.template get_sycl_accessor<cl::sycl::access::mode::discard_write>(num_coeffs_to_preserve,cgh, output);
|
||||
|
||||
cgh.parallel_for<Self>( cl::sycl::nd_range<1>(cl::sycl::range<1>(GRange), cl::sycl::range<1>(tileSize)), [=](cl::sycl::nd_item<1> itemID) {
|
||||
typedef typename TensorSycl::internal::ConvertToDeviceExpression<const HostExpr>::Type DevExpr;
|
||||
|
||||
@@ -56,10 +56,10 @@ struct AccessorConstructor{
|
||||
-> decltype(utility::tuple::append(ExtractAccessor<Arg1>::getTuple(cgh, eval1),utility::tuple::append(ExtractAccessor<Arg2>::getTuple(cgh, eval2), ExtractAccessor<Arg3>::getTuple(cgh, eval3)))) {
|
||||
return utility::tuple::append(ExtractAccessor<Arg1>::getTuple(cgh, eval1),utility::tuple::append(ExtractAccessor<Arg2>::getTuple(cgh, eval2), ExtractAccessor<Arg3>::getTuple(cgh, eval3)));
|
||||
}
|
||||
template< cl::sycl::access::mode AcM, bool MapAllocator, typename Arg> static inline auto getAccessor(cl::sycl::handler& cgh, Arg eval)
|
||||
-> decltype(utility::tuple::make_tuple( eval.device().template get_sycl_accessor<AcM, MapAllocator,
|
||||
template< cl::sycl::access::mode AcM, typename Arg> static inline auto getAccessor(cl::sycl::handler& cgh, Arg eval)
|
||||
-> decltype(utility::tuple::make_tuple( eval.device().template get_sycl_accessor<AcM,
|
||||
typename Eigen::internal::remove_all<typename Arg::CoeffReturnType>::type>(eval.dimensions().TotalSize(), cgh,eval.data()))){
|
||||
return utility::tuple::make_tuple(eval.device().template get_sycl_accessor<AcM, MapAllocator, typename Eigen::internal::remove_all<typename Arg::CoeffReturnType>::type>(eval.dimensions().TotalSize(), cgh,eval.data()));
|
||||
return utility::tuple::make_tuple(eval.device().template get_sycl_accessor<AcM, typename Eigen::internal::remove_all<typename Arg::CoeffReturnType>::type>(eval.dimensions().TotalSize(), cgh,eval.data()));
|
||||
}
|
||||
};
|
||||
|
||||
@@ -141,8 +141,8 @@ struct ExtractAccessor<TensorEvaluator<TensorAssignOp<LHSExpr, RHSExpr>, Dev> >
|
||||
template <typename PlainObjectType, int Options_, typename Dev>\
|
||||
struct ExtractAccessor<TensorEvaluator<CVQual TensorMap<PlainObjectType, Options_>, Dev> > {\
|
||||
static inline auto getTuple(cl::sycl::handler& cgh,const TensorEvaluator<CVQual TensorMap<PlainObjectType, Options_>, Dev> eval)\
|
||||
-> decltype(AccessorConstructor::template getAccessor<ACCType, true>(cgh, eval)){\
|
||||
return AccessorConstructor::template getAccessor<ACCType, true>(cgh, eval);\
|
||||
-> decltype(AccessorConstructor::template getAccessor<ACCType>(cgh, eval)){\
|
||||
return AccessorConstructor::template getAccessor<ACCType>(cgh, eval);\
|
||||
}\
|
||||
};
|
||||
TENSORMAPEXPR(const, cl::sycl::access::mode::read)
|
||||
@@ -153,8 +153,8 @@ TENSORMAPEXPR(, cl::sycl::access::mode::read_write)
|
||||
template <typename Expr, typename Dev>
|
||||
struct ExtractAccessor<TensorEvaluator<const TensorForcedEvalOp<Expr>, Dev> > {
|
||||
static inline auto getTuple(cl::sycl::handler& cgh, const TensorEvaluator<const TensorForcedEvalOp<Expr>, Dev> eval)
|
||||
-> decltype(AccessorConstructor::template getAccessor<cl::sycl::access::mode::read, false>(cgh, eval)){
|
||||
return AccessorConstructor::template getAccessor<cl::sycl::access::mode::read, false>(cgh, eval);
|
||||
-> decltype(AccessorConstructor::template getAccessor<cl::sycl::access::mode::read>(cgh, eval)){
|
||||
return AccessorConstructor::template getAccessor<cl::sycl::access::mode::read>(cgh, eval);
|
||||
}
|
||||
};
|
||||
|
||||
@@ -167,8 +167,8 @@ struct ExtractAccessor<TensorEvaluator<TensorForcedEvalOp<Expr>, Dev> >
|
||||
template <typename Expr, typename Dev>
|
||||
struct ExtractAccessor<TensorEvaluator<const TensorEvalToOp<Expr>, Dev> > {
|
||||
static inline auto getTuple(cl::sycl::handler& cgh,const TensorEvaluator<const TensorEvalToOp<Expr>, Dev> eval)
|
||||
-> decltype(utility::tuple::append(AccessorConstructor::template getAccessor<cl::sycl::access::mode::write, false>(cgh, eval), AccessorConstructor::getTuple(cgh, eval.impl()))){
|
||||
return utility::tuple::append(AccessorConstructor::template getAccessor<cl::sycl::access::mode::write, false>(cgh, eval), AccessorConstructor::getTuple(cgh, eval.impl()));
|
||||
-> decltype(utility::tuple::append(AccessorConstructor::template getAccessor<cl::sycl::access::mode::write>(cgh, eval), AccessorConstructor::getTuple(cgh, eval.impl()))){
|
||||
return utility::tuple::append(AccessorConstructor::template getAccessor<cl::sycl::access::mode::write>(cgh, eval), AccessorConstructor::getTuple(cgh, eval.impl()));
|
||||
}
|
||||
};
|
||||
|
||||
@@ -181,8 +181,8 @@ struct ExtractAccessor<TensorEvaluator<TensorEvalToOp<Expr>, Dev> >
|
||||
template <typename OP, typename Dim, typename Expr, typename Dev>
|
||||
struct ExtractAccessor<TensorEvaluator<const TensorReductionOp<OP, Dim, Expr>, Dev> > {
|
||||
static inline auto getTuple(cl::sycl::handler& cgh, const TensorEvaluator<const TensorReductionOp<OP, Dim, Expr>, Dev> eval)
|
||||
-> decltype(AccessorConstructor::template getAccessor<cl::sycl::access::mode::read, false>(cgh, eval)){
|
||||
return AccessorConstructor::template getAccessor<cl::sycl::access::mode::read, false>(cgh, eval);
|
||||
-> decltype(AccessorConstructor::template getAccessor<cl::sycl::access::mode::read>(cgh, eval)){
|
||||
return AccessorConstructor::template getAccessor<cl::sycl::access::mode::read>(cgh, eval);
|
||||
}
|
||||
};
|
||||
|
||||
|
||||
Reference in New Issue
Block a user