third_party/opencv
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Enable stateful kernels in G-API OCV Backend

Browse Source
This commit is contained in:
AsyaPronina
2020-06-04 19:55:49 +03:00
parent c3e8a82c9c
commit b083c20eb2
15 changed files with 527 additions and 44 deletions
Download Patch File Download Diff File Expand all files Collapse all files
modules/gapi/src/backends/cpu/gcpubackend.cpp
+57 -8
View File
@@ -33,13 +33,13 @@
//
// If not, we need to introduce that!
using GCPUModel = ade::TypedGraph
< cv::gimpl::Unit
< cv::gimpl::CPUUnit
, cv::gimpl::Protocol
>;
// FIXME: Same issue with Typed and ConstTyped
using GConstGCPUModel = ade::ConstTypedGraph
< cv::gimpl::Unit
< cv::gimpl::CPUUnit
, cv::gimpl::Protocol
>;
@@ -53,7 +53,7 @@ namespace
{
GCPUModel gm(graph);
auto cpu_impl = cv::util::any_cast<cv::GCPUKernel>(impl.opaque);
gm.metadata(op_node).set(cv::gimpl::Unit{cpu_impl});
gm.metadata(op_node).set(cv::gimpl::CPUUnit{cpu_impl});
}
virtual EPtr compile(const ade::Graph &graph,
@@ -78,11 +78,23 @@ cv::gimpl::GCPUExecutable::GCPUExecutable(const ade::Graph &g,
{
// Convert list of operations (which is topologically sorted already)
// into an execution script.
GConstGCPUModel gcm(m_g);
for (auto &nh : nodes)
{
switch (m_gm.metadata(nh).get<NodeType>().t)
{
case NodeType::OP: m_script.push_back({nh, GModel::collectOutputMeta(m_gm, nh)}); break;
case NodeType::OP:
{
m_script.push_back({nh, GModel::collectOutputMeta(m_gm, nh)});
// If kernel is stateful then prepare storage for its state.
GCPUKernel k = gcm.metadata(nh).get<CPUUnit>().k;
if (k.m_isStateful)
{
m_nodesToStates[nh] = GArg{ };
}
break;
}
case NodeType::DATA:
{
m_dataNodes.push_back(nh);
@@ -104,6 +116,9 @@ cv::gimpl::GCPUExecutable::GCPUExecutable(const ade::Graph &g,
default: util::throw_error(std::logic_error("Unsupported NodeType type"));
}
}
// For each stateful kernel call 'setup' user callback to initialize state.
setupKernelStates();
}
// FIXME: Document what it does
@@ -140,6 +155,26 @@ cv::GArg cv::gimpl::GCPUExecutable::packArg(const GArg &arg)
}
}
void cv::gimpl::GCPUExecutable::setupKernelStates()
{
GConstGCPUModel gcm(m_g);
for (auto& nodeToState : m_nodesToStates)
{
auto& kernelNode = nodeToState.first;
auto& kernelState = nodeToState.second;
const GCPUKernel& kernel = gcm.metadata(kernelNode).get<CPUUnit>().k;
kernel.m_setupF(GModel::collectInputMeta(m_gm, kernelNode),
m_gm.metadata(kernelNode).get<Op>().args,
kernelState);
}
}
void cv::gimpl::GCPUExecutable::handleNewStream()
{
m_newStreamStarted = true;
}
void cv::gimpl::GCPUExecutable::run(std::vector<InObj> &&input_objs,
std::vector<OutObj> &&output_objs)
{
@@ -167,6 +202,14 @@ void cv::gimpl::GCPUExecutable::run(std::vector<InObj> &&input_objs,
}
}
// In case if new video-stream happens - for each stateful kernel
// call 'setup' user callback to re-initialize state.
if (m_newStreamStarted)
{
setupKernelStates();
m_newStreamStarted = false;
}
// OpenCV backend execution is not a rocket science at all.
// Simply invoke our kernels in the proper order.
GConstGCPUModel gcm(m_g);
@@ -176,7 +219,7 @@ void cv::gimpl::GCPUExecutable::run(std::vector<InObj> &&input_objs,
// Obtain our real execution unit
// TODO: Should kernels be copyable?
GCPUKernel k = gcm.metadata(op_info.nh).get<Unit>().k;
GCPUKernel k = gcm.metadata(op_info.nh).get<CPUUnit>().k;
// Initialize kernel's execution context:
// - Input parameters
@@ -185,8 +228,8 @@ void cv::gimpl::GCPUExecutable::run(std::vector<InObj> &&input_objs,
using namespace std::placeholders;
ade::util::transform(op.args,
std::back_inserter(context.m_args),
std::bind(&GCPUExecutable::packArg, this, _1));
std::back_inserter(context.m_args),
std::bind(&GCPUExecutable::packArg, this, _1));
// - Output parameters.
// FIXME: pre-allocate internal Mats, etc, according to the known meta
@@ -198,8 +241,14 @@ void cv::gimpl::GCPUExecutable::run(std::vector<InObj> &&input_objs,
context.m_results[out_port] = magazine::getObjPtr(m_res, out_desc);
}
// For stateful kernel add state to its execution context
if (k.m_isStateful)
{
context.m_state = m_nodesToStates.at(op_info.nh);
}
// Now trigger the executable unit
k.apply(context);
k.m_runF(context);
//As Kernels are forbidden to allocate memory for (Mat) outputs,
//this code seems redundant, at least for Mats
modules/gapi/src/backends/cpu/gcpubackend.hpp
+10 -1
View File
@@ -23,7 +23,7 @@
namespace cv { namespace gimpl {
struct Unit
struct CPUUnit
{
static const char *name() { return "HostKernel"; }
GCPUKernel k;
@@ -48,6 +48,13 @@ class GCPUExecutable final: public GIslandExecutable
// Actual data of all resources in graph (both internal and external)
Mag m_res;
GArg packArg(const GArg &arg);
void setupKernelStates();
// TODO: Check that it is thread-safe
std::unordered_map<ade::NodeHandle, GArg,
ade::HandleHasher<ade::Node>> m_nodesToStates;
bool m_newStreamStarted = false;
public:
GCPUExecutable(const ade::Graph &graph,
@@ -62,6 +69,8 @@ public:
util::throw_error(std::logic_error("GCPUExecutable::reshape() should never be called"));
}
virtual void handleNewStream() override;
virtual void run(std::vector<InObj> &&input_objs,
std::vector<OutObj> &&output_objs) override;
};
modules/gapi/src/backends/cpu/gcpukernel.cpp
+2 -8
View File
@@ -45,13 +45,7 @@ cv::GCPUKernel::GCPUKernel()
{
}
cv::GCPUKernel::GCPUKernel(const GCPUKernel::F &f)
: m_f(f)
cv::GCPUKernel::GCPUKernel(const GCPUKernel::RunF &runF, const GCPUKernel::SetupF &setupF)
: m_runF(runF), m_setupF(setupF), m_isStateful(m_setupF != nullptr)
{
}
void cv::GCPUKernel::apply(GCPUContext &ctx)
{
GAPI_Assert(m_f);
m_f(ctx);
}
modules/gapi/src/backends/ocl/goclbackend.cpp
+4 -4
View File
@@ -33,13 +33,13 @@
//
// If not, we need to introduce that!
using GOCLModel = ade::TypedGraph
< cv::gimpl::Unit
< cv::gimpl::OCLUnit
, cv::gimpl::Protocol
>;
// FIXME: Same issue with Typed and ConstTyped
using GConstGOCLModel = ade::ConstTypedGraph
< cv::gimpl::Unit
< cv::gimpl::OCLUnit
, cv::gimpl::Protocol
>;
@@ -53,7 +53,7 @@ namespace
{
GOCLModel gm(graph);
auto ocl_impl = cv::util::any_cast<cv::GOCLKernel>(impl.opaque);
gm.metadata(op_node).set(cv::gimpl::Unit{ocl_impl});
gm.metadata(op_node).set(cv::gimpl::OCLUnit{ocl_impl});
}
virtual EPtr compile(const ade::Graph &graph,
@@ -198,7 +198,7 @@ void cv::gimpl::GOCLExecutable::run(std::vector<InObj> &&input_objs,
// Obtain our real execution unit
// TODO: Should kernels be copyable?
GOCLKernel k = gcm.metadata(op_info.nh).get<Unit>().k;
GOCLKernel k = gcm.metadata(op_info.nh).get<OCLUnit>().k;
// Initialize kernel's execution context:
// - Input parameters
modules/gapi/src/backends/ocl/goclbackend.hpp
+1 -1
View File
@@ -23,7 +23,7 @@
namespace cv { namespace gimpl {
struct Unit
struct OCLUnit
{
static const char *name() { return "OCLKernel"; }
GOCLKernel k;
modules/gapi/src/backends/render/grenderocvbackend.cpp
+1 -1
View File
@@ -93,7 +93,7 @@ void cv::gimpl::render::ocv::GRenderExecutable::run(std::vector<InObj> &&input_
context.m_args.emplace_back(m_ftpr.get());
k.apply(context);
k.m_runF(context);
for (auto &it : output_objs) magazine::writeBack(m_res, it.first, it.second);
}
modules/gapi/src/compiler/gcompiled.cpp
+11
View File
@@ -72,6 +72,12 @@ void cv::GCompiled::Priv::reshape(const GMetaArgs& inMetas, const GCompileArgs&
m_metas = inMetas;
}
void cv::GCompiled::Priv::prepareForNewStream()
{
GAPI_Assert(m_exec);
m_exec->prepareForNewStream();
}
const cv::gimpl::GModel::Graph& cv::GCompiled::Priv::model() const
{
GAPI_Assert(nullptr != m_exec);
@@ -155,3 +161,8 @@ void cv::GCompiled::reshape(const GMetaArgs& inMetas, const GCompileArgs& args)
{
m_priv->reshape(inMetas, args);
}
void cv::GCompiled::prepareForNewStream()
{
m_priv->prepareForNewStream();
}
modules/gapi/src/compiler/gcompiled_priv.hpp
+1
View File
@@ -48,6 +48,7 @@ public:
bool canReshape() const;
void reshape(const GMetaArgs& inMetas, const GCompileArgs &args);
void prepareForNewStream();
void run(cv::gimpl::GRuntimeArgs &&args);
const GMetaArgs& metas() const;
modules/gapi/src/executor/gexecutor.cpp
+8
View File
@@ -265,3 +265,11 @@ void cv::gimpl::GExecutor::reshape(const GMetaArgs& inMetas, const GCompileArgs&
passes::inferMeta(ctx, true);
m_ops[0].isl_exec->reshape(g, args);
}
void cv::gimpl::GExecutor::prepareForNewStream()
{
for (auto &op : m_ops)
{
op.isl_exec->handleNewStream();
}
}
modules/gapi/src/executor/gexecutor.hpp
+2
View File
@@ -91,6 +91,8 @@ public:
bool canReshape() const;
void reshape(const GMetaArgs& inMetas, const GCompileArgs& args);
void prepareForNewStream();
const GModel::Graph& model() const; // FIXME: make it ConstGraph?
};
modules/gapi/src/executor/gstreamingexecutor.cpp
+12 -1
View File
@@ -800,6 +800,7 @@ void cv::gimpl::GStreamingExecutor::setSource(GRunArgs &&ins)
}
}
};
bool islandsRecompiled = false;
const auto new_meta = cv::descr_of(ins); // 0
if (gm.metadata().contains<OriginalInputMeta>()) // (1)
{
@@ -821,6 +822,8 @@ void cv::gimpl::GStreamingExecutor::setSource(GRunArgs &&ins)
}
update_int_metas(); // (7)
m_reshapable = util::make_optional(is_reshapable);
islandsRecompiled = true;
}
else // (8)
{
@@ -929,7 +932,15 @@ void cv::gimpl::GStreamingExecutor::setSource(GRunArgs &&ins)
for (auto &&out_eh : op.nh->outNodes()) {
out_queues.push_back(reader_queues(*m_island_graph, out_eh));
}
op.isl_exec->handleNewStream();
// If Island Executable is recompiled, all its stuff including internal kernel states
// are recreated and re-initialized automatically.
// But if not, we should notify Island Executable about new started stream to let it update
// its internal variables.
if (!islandsRecompiled)
{
op.isl_exec->handleNewStream();
}
m_threads.emplace_back(islandActorThread,
op.in_objects,