A GpuTask dispatcher. More...
#include <PxGpuDispatcher.h>
Public Member Functions | |
| virtual void | startSimulation ()=0 |
| Record the start of a simulation step. More... | |
| virtual void | startGroup ()=0 |
| Record the start of a GpuTask batch submission. More... | |
| virtual void | submitTask (PxTask &task)=0 |
| Submit a GpuTask for execution. More... | |
| virtual void | finishGroup ()=0 |
| Record the end of a GpuTask batch submission. More... | |
| virtual void | addCompletionPrereq (PxBaseTask &task)=0 |
| Add a CUDA completion prerequisite dependency to a task. More... | |
| virtual PxCudaContextManager * | getCudaContextManager ()=0 |
| Retrieve the PxCudaContextManager associated with this PxGpuDispatcher. More... | |
| virtual void | stopSimulation ()=0 |
| Record the end of a simulation frame. More... | |
| virtual bool | failureDetected () const =0 |
| Returns true if a CUDA call has returned a non-recoverable error. More... | |
| virtual void | forceFailureMode ()=0 |
| Force the PxGpuDispatcher into failure mode. More... | |
| virtual void * | getCurrentProfileBuffer () const =0 |
| Returns a pointer to the current in-use profile buffer. More... | |
| virtual PxU16 | registerKernelNames (const char **, PxU16 count)=0 |
| Register kernel names with PlatformAnalyzer. More... | |
| virtual void | launchCopyKernel (PxGpuCopyDesc *desc, PxU32 count, CUstream stream)=0 |
| Launch a copy kernel with arbitrary number of copy commands. More... | |
| virtual PxBaseTask & | getPreLaunchTask ()=0 |
| Query pre launch task that runs before launching gpu kernels. More... | |
| virtual void | addPreLaunchDependent (PxBaseTask &dependent)=0 |
| Adds a gpu launch task that gets executed after the pre launch task. More... | |
| virtual PxBaseTask & | getPostLaunchTask ()=0 |
| Query post launch task that runs after the gpu is done. More... | |
| virtual void | addPostLaunchDependent (PxBaseTask &dependent)=0 |
| Adds a task that gets executed after the post launch task. More... | |
Protected Member Functions | |
| virtual | ~PxGpuDispatcher () |
| protected destructor More... | |
A GpuTask dispatcher.
A PxGpuDispatcher executes GpuTasks submitted by one or more TaskManagers (one or more scenes). It maintains a CPU worker thread which waits on GpuTask "groups" to be submitted. The submission API is explicitly sessioned so that GpuTasks are dispatched together as a group whenever possible to improve parallelism on the GPU.
A PxGpuDispatcher cannot be allocated ad-hoc, they are created as a result of creating a PxCudaContextManager. Every PxCudaContextManager has a PxGpuDispatcher instance that can be queried. In this way, each PxGpuDispatcher is tied to exactly one CUDA context.
A scene will use CPU fallback Tasks for GpuTasks if the PxTaskManager provided to it does not have a PxGpuDispatcher. For this reason, the PxGpuDispatcher must be assigned to the PxTaskManager before the PxTaskManager is given to a scene.
Multiple TaskManagers may safely share a single PxGpuDispatcher instance, thus enabling scenes to share a CUDA context.
Only failureDetected() is intended for use by the user. The rest of the PxGpuDispatcher public methods are reserved for internal use by only both TaskManagers and GpuTasks.
|
inlineprotectedvirtual |
protected destructor
GpuDispatchers are allocated and freed by their PxCudaContextManager.
|
pure virtual |
Add a CUDA completion prerequisite dependency to a task.
A GpuTask calls this function to add a prerequisite dependency on another task (usually a CpuTask) preventing that task from starting until all of the CUDA kernels and copies already launched have been completed. The PxGpuDispatcher will increment that task's reference count, blocking its execution, until the CUDA work is complete.
This is generally only required when a CPU task is expecting the results of the CUDA kernels to have been copied into host memory.
This mechanism is not at all not required to ensure CUDA kernels and copies are issued in the correct order. Kernel issue order is determined by normal task dependencies. The rule of thumb is to only use a blocking completion prerequisite if the task in question depends on a completed GPU->Host DMA.
The PxGpuDispatcher issues a blocking event record to CUDA for the purposes of tracking the already submitted CUDA work. When this event is resolved, the PxGpuDispatcher manually decrements the reference count of the specified task, allowing it to execute (assuming it does not have other pending prerequisites).
|
pure virtual |
Adds a task that gets executed after the post launch task.
This is part of an optional feature to schedule multiple gpu features at the same time to get kernels to run in parallel.
|
pure virtual |
Adds a gpu launch task that gets executed after the pre launch task.
This is part of an optional feature to schedule multiple gpu features at the same time to get kernels to run in parallel.
|
pure virtual |
Returns true if a CUDA call has returned a non-recoverable error.
A return value of true indicates a fatal error has occurred. To protect itself, the PxGpuDispatcher enters a fall through mode that allows GpuTasks to complete without being executed. This allows simulations to continue but leaves GPU content static or corrupted.
The user may try to recover from these failures by deleting GPU content so the visual artifacts are mimimized. But there is no way to recover the state of the GPU actors before the failure. Once a CUDA context is in this state, the only recourse is to create a new CUDA context, a new scene, and start over.
This is our "Best Effort" attempt to not turn a soft failure into a hard failure because continued use of a CUDA context after it has returned an error will usually result in a driver reset. However if the initial failure was serious enough, a reset may have already occurred by the time we learn of it.
|
pure virtual |
Record the end of a GpuTask batch submission.
A PxTaskManager calls this function to notify the PxGpuDispatcher that it is done submitting a group of GpuTasks (GpuTasks which were all make ready to run by the same prerequisite dependency becoming resolved). If no other group submissions are in progress, the PxGpuDispatcher will execute the set of ready tasks.
|
pure virtual |
Force the PxGpuDispatcher into failure mode.
This API should be used if user code detects a non-recoverable CUDA error. This ensures the PxGpuDispatcher does not launch any further CUDA work. Subsequent calls to failureDetected() will return true.
|
pure virtual |
Retrieve the PxCudaContextManager associated with this PxGpuDispatcher.
Every PxCudaContextManager has one PxGpuDispatcher, and every PxGpuDispatcher has one PxCudaContextManager.
|
pure virtual |
Returns a pointer to the current in-use profile buffer.
The returned pointer should be passed to all kernel launches to enable CTA/Warp level profiling. If a data collector is not attached, or CTA profiling is not enabled, the pointer will be zero.
|
pure virtual |
Query post launch task that runs after the gpu is done.
This is part of an optional feature to schedule multiple gpu features at the same time to get kernels to run in parallel.
|
pure virtual |
Query pre launch task that runs before launching gpu kernels.
This is part of an optional feature to schedule multiple gpu features at the same time to get kernels to run in parallel.
|
pure virtual |
Launch a copy kernel with arbitrary number of copy commands.
This method is intended to be called from Kernel GpuTasks, but it can function outside of that context as well.
If count is 1, the descriptor is passed to the kernel as arguments, so it may be declared on the stack.
If count is greater than 1, the kernel will read the descriptors out of host memory. Because of this, the descriptor array must be located in page locked (pinned) memory. The provided descriptors may be modified by this method (converting host pointers to their GPU mapped equivalents) and should be considered owned by CUDA until the current batch of work has completed, so descriptor arrays should not be freed or modified until you have received a completion notification.
If your GPU does not support mapping of page locked memory (SM>=1.1), this function degrades to calling CUDA copy methods.
Register kernel names with PlatformAnalyzer.
The returned PxU16 must be stored and used as a base offset for the ID passed to the KERNEL_START|STOP_EVENT macros.
|
pure virtual |
Record the start of a GpuTask batch submission.
A PxTaskManager calls this function to notify the PxGpuDispatcher that one or more GpuTasks are about to be submitted for execution. The PxGpuDispatcher will not read the incoming task queue until it receives one finishGroup() call for each startGroup() call. This is to ensure as many GpuTasks as possible are executed together as a group, generating optimal parallelism on the GPU.
|
pure virtual |
Record the start of a simulation step.
A PxTaskManager calls this function to record the beginning of a simulation step. The PxGpuDispatcher uses this notification to initialize the profiler state.
|
pure virtual |
Record the end of a simulation frame.
A PxTaskManager calls this function to record the completion of its dependency graph. If profiling is enabled, the PxGpuDispatcher will trigger the retrieval of profiling data from the GPU at this point.
|
pure virtual |
Submit a GpuTask for execution.
Submitted tasks are pushed onto an incoming queue. The PxGpuDispatcher will take the contents of this queue every time the pending group count reaches 0 and run the group of submitted GpuTasks as an interleaved group.