Skip to content

[SYCL] Fixed sub-buffer memory allocation update #1486

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Apr 9, 2020
Merged

[SYCL] Fixed sub-buffer memory allocation update #1486

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
0af9713
[SYCL] Fixed sub-buffer memory allocation update
Apr 2, 2020
12a18fb
Simplified solution
Apr 7, 2020
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 12 additions & 0 deletions sycl/source/detail/scheduler/commands.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -694,6 +694,18 @@ void AllocaSubBufCommand::emitInstrumentationData() {
#endif
}

void *AllocaSubBufCommand::getMemAllocation() const {
// In some cases parent`s memory allocation might change (e.g., after
// map/unmap operations). If parent`s memory allocation changes, sub-buffer
// memory allocation should be changed as well.
if (MQueue->is_host()) {
return static_cast<void *>(
static_cast<char *>(MParentAlloca->getMemAllocation()) +
MRequirement.MOffsetInBytes);
}
return MMemAllocation;
}

cl_int AllocaSubBufCommand::enqueueImp() {
std::vector<EventImplPtr> EventImpls =
Command::prepareEvents(detail::getSyclObjImpl(MQueue->get_context()));
Expand Down
4 changes: 3 additions & 1 deletion sycl/source/detail/scheduler/commands.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -264,7 +264,7 @@ class AllocaCommandBase : public Command {

SYCLMemObjI *getSYCLMemObj() const { return MRequirement.MSYCLMemObj; }

void *getMemAllocation() const { return MMemAllocation; }
virtual void *getMemAllocation() const = 0;

const Requirement *getRequirement() const final { return &MRequirement; }

Expand Down Expand Up @@ -298,6 +298,7 @@ class AllocaCommand : public AllocaCommandBase {
bool InitFromUserData = true,
AllocaCommandBase *LinkedAllocaCmd = nullptr);

void *getMemAllocation() const final { return MMemAllocation; }
void printDot(std::ostream &Stream) const final;
void emitInstrumentationData();

Expand All @@ -314,6 +315,7 @@ class AllocaSubBufCommand : public AllocaCommandBase {
AllocaSubBufCommand(QueueImplPtr Queue, Requirement Req,
AllocaCommandBase *ParentAlloca);

void *getMemAllocation() const final;
void printDot(std::ostream &Stream) const final;
AllocaCommandBase *getParentAlloca() { return MParentAlloca; }
void emitInstrumentationData();
Expand Down
1 change: 1 addition & 0 deletions sycl/source/detail/scheduler/graph_builder.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -533,6 +533,7 @@ AllocaCommandBase *Scheduler::GraphBuilder::findAllocaForReq(
bool Res = sameCtx(AllocaCmd->getQueue()->getContextImplPtr(), Context);
if (IsSuitableSubReq(Req)) {
const Requirement *TmpReq = AllocaCmd->getRequirement();
Res &= AllocaCmd->getType() == Command::CommandType::ALLOCA_SUB_BUF;
Res &= TmpReq->MOffsetInBytes == Req->MOffsetInBytes;
Res &= TmpReq->MSYCLMemObj->getSize() == Req->MSYCLMemObj->getSize();
}
Expand Down
2 changes: 1 addition & 1 deletion sycl/test/basic_tests/buffer/subbuffer.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -279,7 +279,7 @@ void checkMultipleContexts() {
{
sycl::queue queue1;
sycl::buffer<int, 1> buf(a, sycl::range<1>(N));
sycl::buffer<int, 1> subbuf1(buf, sycl::id<1>(0), sycl::range<1>(N / 2));
sycl::buffer<int, 1> subbuf1(buf, sycl::id<1>(N / 2), sycl::range<1>(N / 2));
queue1.submit([&](sycl::handler &cgh) {
auto bufacc = subbuf1.get_access<sycl::access::mode::read_write>(cgh);
cgh.parallel_for<class sub_buffer_3>(
Expand Down