CLENQUEUERELEASEDX9M(3clc)

📦
version 1.0~svn33624 · 3 versions
, in ubuntu jammy, ubuntu noble

✨ CLENQUEUERELEASEDX9M ✨

okay so bestie, we're in the OpenCL Manual, section 3clc, and the page is CLENQUEUERELEASEDX9M(3clc). yes that's a whole mouthful, i know. say it with me. it's giving "someone sneezed on a keyboard" energy but it's actually iconic, i promise 💅

🌸 NAME 🌸

clEnqueueReleaseDX9MediaSurfacesKHR — releases OpenCL memory objects that were made from media surfaces. that's it, that's the job 🎀

🦋 SYNOPSIS 🦋

cl_int
clEnqueueReleaseDX9MediaSurfacesKHR(cl_command_queue command_queue,
                                    cl_uint num_objects,
                                    const cl_mem *mem_objects,
                                    cl_uint num_events_in_wait_list,
                                    const cl_event *event_wait_list,
                                    cl_event *event);

💖 PARAMETERS 💖

  • command_queue — a valid command-queue. gotta be legit, no fake ones 🫶
  • num_objects — how many memory objects you're releasing in mem_objects. count them, bestie, fr fr.
  • mem_objects — a pointer to a list of OpenCL memory objects that were created from media surfaces. these are the ones getting freed back to the wild 🌈
  • num_events_in_wait_list, event_wait_list — these say which events need to finish before this specific command can run. so like, the vibe check before entry ✋ if event_wait_list is NULL, then this command doesn't wait on any event at all. and if event_wait_list is NULL, then num_events_in_wait_list must be 0. if event_wait_list is not NULL, then the list of events that event_wait_list points to has to be valid and num_events_in_wait_list must be greater than 0. the events specified in event returns an event object that identifies this particular command and can be used to query or queue a wait for this particular command to complete. event can be NULL, and in that case it will not be possible for the application to query the status of this command or queue a wait for this command to complete 😭 and if event_wait_list and the event arguments are not NULL, then the event argument should not refer to an element of the event_wait_list array. no crossover, we don't do that here 💀

👑 NOTES 👑

so this is the one that releases OpenCL memory objects that were created from media surfaces. the media surfaces get released by the OpenCL context that's associated with command_queue.

here's the important part, bestie, listen up 🚨 OpenCL memory objects created from media surfaces that have been acquired by OpenCL must be released by OpenCL before they may be accessed by the media adapter API. accessing a media surface while its corresponding OpenCL memory object is acquired is in error and will result in undefined behavior — including but not limited to possible OpenCL errors, data corruption, and program termination. so like... don't. that's the whole warning. that's not a vibe, that's a crash 🔥💀

now, if CL_CONTEXT_INTEROP_USER_SYNC is not specified as CL_TRUE during context creation, then clEnqueueReleaseDX9MediaSurfacesKHR provides the synchronization guarantee that any calls to media adapter APIs involving the interop device(s) used in the OpenCL context made after the call to clEnqueueReleaseDX9MediaSurfacesKHR will not start executing until after all events in event_wait_list are complete and all work already submitted to command_queue completes execution. ✨

but — if the context was created with properties specifying CL_CONTEXT_INTEROP_USER_SYNC as CL_TRUE, then the user is responsible for guaranteeing that any media adapter API calls involving the interop device(s) used in the OpenCL context made after clEnqueueReleaseDX9MediaSurfacesKHR will not start executing until after event returned by clEnqueueReleaseDX9MediaSurfacesKHR reports completion. so yeah, you're the main character now, you handle the sync 🫡

the goal of this extension is to let applications use media surfaces as OpenCL memory objects. this allows efficient sharing of data between OpenCL and selected adapter APIs (only DX9 for now, ngl). if this extension is supported, an OpenCL image object can be created from a media surface and the OpenCL API can be used to execute kernels that read and/or write memory objects that are media surfaces. 📌 note that OpenCL memory objects may be created from the adapter media surface if and only if the OpenCL context has been created from that adapter. if-and-only-if, bestie, that's a whole math commitment 💜

if this extension is supported by an implementation, the string "cl_khr_media_sharing" will be present in the CL_PLATFORM_EXTENSIONS or CL_DEVICE_EXTENSIONS string described in the table of allowed values for param_name for clGetDeviceInfo(3clc). so go check, it'll be right there ⭐

as currently proposed the interfaces for this extension would be provided in cl_dx9_media_sharing.h. 📁

🌈 D3D Surface Formats 🌈

this section includes the D3D surface formats that are supported when the adapter type is one of the Direct 3D lineage . using a D3D surface format not listed here is an error 🚫 to extend the use of this extension to support media adapters beyond DirectX9, tables similar to the ones in this section will need to be defined for the surface formats supported by the new media adapter. all implementations that support this extension are required to support the NV12 surface format, and the other surface formats supported are the same surface formats that the adapter you are sharing with supports — as long as they are listed in the two tables below. 👇

so the first table column is the FOUR CC code, and the second one is the CL image format (channel order, channel data type). here's the whole lineup, no skips:

  • FOURCC('N', 'V', '1', '2'), Plane 0 — CL_R, CL_UNORM_INT8
  • FOURCC('N', 'V', '1', '2'), Plane 1 — CL_RG, CL_UNORM_INT8
  • FOURCC('Y', 'V', '1', '2'), Plane 0 — CL_R, CL_UNORM_INT8
  • FOURCC('Y', 'V', '1', '2'), Plane 1 — CL_R, CL_UNORM_INT8
  • FOURCC('Y', 'V', '1', '2'), Plane 2 — CL_R, CL_UNORM_INT8

those are the surface formats, periodt. 💅

In the table above, NV12 Plane 0 corresponds to the luminance (Y) channel and Plane 1 corresponds to the UV channels. The YV12 Plane 0 corresponds to the Y channel, Plane 1 to the U channel and Plane 2 to the V channel. And pls note 📌 that the YUV formats map to CL_R and CL_RG but do not perform any YUV to RGB conversion and vice-versa. like, none. zero. it's giving "we just label it and move on" 💅

So here's the whole table bestie — left column is the D3D format</>, right column is the CL image format (channel order, channel data type)</> ✨

  • D3DFMT_R32F — CL_R, CL_FLOAT
  • D3DFMT_R16F — CL_R, CL_HALF_FLOAT
  • D3DFMT_L16 — CL_R, CL_UNORM_INT16
  • D3DFMT_A8 — CL_A, CL_UNORM_INT8
  • D3DFMT_L8 — CL_R, CL_UNORM_INT8
  • D3DFMT_G32R32F — CL_RG, CL_FLOAT
  • D3DFMT_G16R16F — CL_RG, CL_HALF_FLOAT
  • D3DFMT_G16R16 — CL_RG, CL_UNORM_INT16
  • D3DFMT_A8L8 — CL_RG, CL_UNORM_INT8
  • D3DFMT_A32B32G32R32F — CL_RGBA, CL_FLOAT
  • D3DFMT_A16B16G16R16F — CL_RGBA, CL_HALF_FLOAT
  • D3DFMT_A16B16G16R16 — CL_RGBA, CL_UNORM_INT16
  • D3DFMT_A8B8G8R8 — CL_RGBA, CL_UNORM_INT8
  • D3DFMT_X8B8G8R8 — CL_RGBA, CL_UNORM_INT8
  • D3DFMT_A8R8G8B8 — CL_BGRA, CL_UNORM_INT8
  • D3DFMT_X8R8G8B8 — CL_BGRA, CL_UNORM_INT8

Okay and now the part that trips everyone up 🙈 — the D3D9 format names seem to imply that the order of the color channels are switched relative to OpenCL, but that is not the case. Not even a little. For example, layout of channels for each pixel for D3DFMT_A32FB32FG32FR32F is the same as CL_RGBA, CL_FLOAT. so don't go flipping things in your head, bestie, it's a trap 🪤

Properties of media surface objects may be queried using clGetMemObjectInfo(3clc) and clGetImageInfo(3clc) with param_name CL_MEM_MEDIA_ADAPTER_TYPE_KHR, CL_MEM_MEDIA_SURFACE_INFO_KHR and CL_IMAGE_MEDIA_SURFACE_PLANE_KHR as described for clGetImageInfo(3clc) and in the information below about accessing mapped regions of a memory object.

Accessing mapped regions of a memory object. This section describes the behavior of OpenCL commands that access mapped regions of a memory object.

The contents of the region of a memory object and associated memory objects (sub-buffer objects or 1D image buffer objects that overlap this region) mapped for writing (i.e. CL_MAP_WRITE or CL_MAP_WRITE_INVALIDATE_REGION is set in map_flags argument to clEnqueueMapBuffer(3clc) or clEnqueueMapImage(3clc)) are considered to be undefined until this region is unmapped. so like. don't touch it. it's not real yet 😭

Multiple commands in command-queues can map a region or overlapping regions of a memory object and associated memory objects (sub-buffer objects or 1D image buffer objects that overlap this region) for reading (i.e. map_flags = CL_MAP_READ). The contents of the regions of a memory object mapped for reading can also be read by kernels and other OpenCL commands (such as clEnqueueCopyBuffer(3clc)) executing on a device(s). ✨ sharing is caring ✨

Mapping (and unmapping) overlapped regions in a memory object and/or associated memory objects (sub-buffer objects or 1D image buffer objects that overlap this region) for writing is an error and will result in CL_INVALID_OPERATION error returned by clEnqueueMapBuffer(3clc) or clEnqueueMapImage(3clc).

If a memory object is currently mapped for writing, the application must ensure that the memory object is unmapped before any enqueued kernels or commands that read from or write to this memory object or any of its associated memory objects (sub-buffer or 1D image buffer objects) or its parent object (if the memory object is a sub-buffer or 1D image buffer object) begin execution; otherwise the behavior is undefined. bestie this one is serious, no cap 🚨
If a memory object is currently mapped for reading, the application must ensure that the memory object is unmapped before any enqueued kernels or commands that write to this memory object or any of its associated memory objects (sub-buffer or 1D image buffer objects) or its parent object (if the memory object is a sub-buffer or 1D image buffer object) begin execution; otherwise the behavior is undefined. same energy, opposite direction 💀

Accessing the contents of the memory region referred to by the mapped pointer that has been unmapped is undefined. like, it's gone, it's not yours anymore 🦋

The mapped pointer returned by clEnqueueMapBuffer(3clc) or clEnqueueMapImage(3clc) can be used as ptr argument value to clEnqueueReadBuffer(3clc), clEnqueueWriteBuffer(3clc), clEnqueueReadBufferRect(3clc), clEnqueueWriteBufferRect(3clc), clEnqueueReadImage(3clc), and clEnqueueWriteImage(3clc), provided the rules described above are adhered to. iconic, honestly — one pointer, so many options 👑

💥 ERRORS 💥

Returns CL_SUCCESS returns CL_SUCCESS if the function is executed successfully. If num_objects is 0 and mem_objects is NULL then the function does nothing and returns CL_SUCCESS. Otherwise it returns one of the following errors:

  • CL_INVALID_VALUE — if num_objects is zero and mem_objects is not a NULL value or if num_objects > 0 and mem_objects is NULL.
  • CL_INVALID_MEM_OBJECT — if memory objects in mem_objects are not valid OpenCL memory objects or if memory objects in mem_objects have not been created from media surfaces.
  • CL_INVALID_COMMAND_QUEUE — if command_queue is not a valid command-queue.
  • CL_INVALID_CONTEXT — if context associated with command_queue was not created from a media object.
  • CL_DX9_MEDIA_SURFACE_NOT_ACQUIRED_KHR — if memory objects in mem_objects have not been previously been acquired using clEnqueueAcquireDX9MediaSurfacesKHR or have been released using clEnqueueReleaseDX9MediaSurfacesKHR(3clc) since the last time that they were acquired. (acquire it first, bestie, that's the whole bit 🔑)
  • CL_INVALID_EVENT_WAIT_LIST — if event_wait_list is NULL and num_event_in_wait_list > 0, or event_wait_list is not NULL and num_event_in_wait_list is 0, or if event objects in event_wait_list are not valid events.
  • CL_OUT_OF_HOST_MEMORY — if there is a failure to allocate resources required by the OpenCL implementation on the host. 😭

📖 SPECIFICATION 📖

OpenCL Specification[1]

👀 SEE ALSO 👀

like, the whole squad 🫶 go read them too, iykyk

✍️ AUTHORS ✍️

The Khronos Group

Copyright © 2007-2011 The Khronos Group Inc.

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and/or associated documentation files (the "Materials"), to deal in the Materials without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Materials, and to permit persons to whom the Materials are furnished to do so, subject to the condition that this copyright notice and permission notice shall be included in all copies or substantial portions of the Materials.

and one (1) last footnote, because of course 📎

  1. OpenCL Specification — page 93, section 9.10.7.4 - Sharing Memory Objects...

okay that's the whole page bestie, you ate it 🎀 this one's from The Khronos Group</>, page dated 01/14/2021</>, filed under CLENQUEUERELEASEDX9M(3clc) — go release those media surfaces, main character 💅✨

📦 where this page comes from