Files
test/source/blender/gpu/GPU_vertex_buffer.hh

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

297 lines
8.6 KiB
C++
Raw Normal View History

/* SPDX-FileCopyrightText: 2016 by Mike Erwin. All rights reserved.
*
* SPDX-License-Identifier: GPL-2.0-or-later */
/** \file
* \ingroup gpu
*
* GPU vertex buffer
*/
#pragma once
#include "BLI_span.hh"
#include "BLI_utildefines.h"
#include "GPU_vertex_format.hh"
enum GPUVertBufStatus {
/** Initial state. */
GPU_VERTBUF_INVALID = 0,
/** Was init with a vertex format. */
GPU_VERTBUF_INIT = (1 << 0),
2021-02-05 16:23:34 +11:00
/** Data has been touched and need to be re-uploaded. */
GPU_VERTBUF_DATA_DIRTY = (1 << 1),
/** The buffer has been created inside GPU memory. */
GPU_VERTBUF_DATA_UPLOADED = (1 << 2),
};
ENUM_OPERATORS(GPUVertBufStatus, GPU_VERTBUF_DATA_UPLOADED)
/**
* How to create a #VertBuf:
* 1) verts = GPU_vertbuf_calloc()
* 2) GPU_vertformat_attr_add(verts->format, ...)
* 3) GPU_vertbuf_data_alloc(verts, vertex_len) <-- finalizes/packs vertex format
* 4) GPU_vertbuf_attr_fill(verts, pos, application_pos_buffer)
*/
enum GPUUsageType {
/* can be extended to support more types */
GPU_USAGE_STREAM = 0,
GPU_USAGE_STATIC = 1, /* do not keep data in memory */
GPU_USAGE_DYNAMIC = 2,
GPU_USAGE_DEVICE_ONLY = 3, /* Do not do host->device data transfers. */
/** Extended usage flags. */
/* Flag for vertex buffers used for textures. Skips additional padding/compaction to ensure
* format matches the texture exactly. Can be masked with other properties, and is stripped
* during VertBuf::init. */
GPU_USAGE_FLAG_BUFFER_TEXTURE_ONLY = 1 << 3,
};
ENUM_OPERATORS(GPUUsageType, GPU_USAGE_FLAG_BUFFER_TEXTURE_ONLY);
namespace blender::gpu {
/**
* Implementation of Vertex Buffers.
* Base class which is then specialized for each implementation (GL, VK, ...).
*/
class VertBuf {
public:
static size_t memory_usage;
GPUVertFormat format = {};
/** Number of verts we want to draw. */
uint vertex_len = 0;
/** Number of verts data. */
uint vertex_alloc = 0;
/** Status flag. */
GPUVertBufStatus flag = GPU_VERTBUF_INVALID;
#ifndef NDEBUG
/** Usage including extended usage flags. */
GPUUsageType extended_usage_ = GPU_USAGE_STATIC;
#endif
protected:
/** NULL indicates data in VRAM (unmapped) */
uchar *data_ = nullptr;
/** Usage hint for GL optimization. */
GPUUsageType usage_ = GPU_USAGE_STATIC;
private:
/** This counter will only avoid freeing the #VertBuf, not the data. */
int handle_refcount_ = 1;
public:
VertBuf();
virtual ~VertBuf();
void init(const GPUVertFormat &format, GPUUsageType usage);
void clear();
/* Data management. */
void allocate(uint vert_len);
void resize(uint vert_len);
void upload();
virtual void bind_as_ssbo(uint binding) = 0;
virtual void bind_as_texture(uint binding) = 0;
virtual void wrap_handle(uint64_t handle) = 0;
VertBuf *duplicate();
/* Size of the data allocated. */
size_t size_alloc_get() const
{
BLI_assert(this->format.packed);
return size_t(this->vertex_alloc) * this->format.stride;
}
/* Size of the data uploaded to the GPU. */
size_t size_used_get() const
{
BLI_assert(format.packed);
return size_t(this->vertex_len) * this->format.stride;
}
void reference_add()
{
handle_refcount_++;
}
void reference_remove()
{
BLI_assert(handle_refcount_ > 0);
handle_refcount_--;
if (handle_refcount_ == 0) {
delete this;
}
}
GPUUsageType get_usage_type() const
{
return usage_;
}
/**
* Returns access to the data allocated for the vertex buffer. The size of the data type must
* match the data type used on the GPU.
*/
template<typename T> MutableSpan<T> data()
{
return MutableSpan<uchar>(data_, this->size_alloc_get()).cast<T>();
}
virtual void update_sub(uint start, uint len, const void *data) = 0;
virtual void read(void *data) const = 0;
protected:
virtual void acquire_data() = 0;
virtual void resize_data() = 0;
virtual void release_data() = 0;
virtual void upload_data() = 0;
virtual void duplicate_data(VertBuf *dst) = 0;
};
} // namespace blender::gpu
blender::gpu::VertBuf *GPU_vertbuf_calloc();
blender::gpu::VertBuf *GPU_vertbuf_create_with_format_ex(const GPUVertFormat &format,
GPUUsageType usage);
#define GPU_vertbuf_create_with_format(format) \
GPU_vertbuf_create_with_format_ex(format, GPU_USAGE_STATIC)
/**
* (Download and) fill data with the data from the vertex buffer.
* NOTE: caller is responsible to reserve enough memory of the data parameter.
*/
void GPU_vertbuf_read(const blender::gpu::VertBuf *verts, void *data);
/** Same as discard but does not free. */
void GPU_vertbuf_clear(blender::gpu::VertBuf *verts);
void GPU_vertbuf_discard(blender::gpu::VertBuf *);
/**
* Avoid blender::gpu::VertBuf data-block being free but not its data.
*/
void GPU_vertbuf_handle_ref_add(blender::gpu::VertBuf *verts);
void GPU_vertbuf_handle_ref_remove(blender::gpu::VertBuf *verts);
void GPU_vertbuf_init_with_format_ex(blender::gpu::VertBuf &verts,
const GPUVertFormat &format,
GPUUsageType);
void GPU_vertbuf_init_build_on_device(blender::gpu::VertBuf &verts,
const GPUVertFormat &format,
uint v_len);
OpenSubDiv: add support for an OpenGL evaluator This evaluator is used in order to evaluate subdivision at render time, allowing for faster renders of meshes with a subdivision surface modifier placed at the last position in the modifier list. When evaluating the subsurf modifier, we detect whether we can delegate evaluation to the draw code. If so, the subdivision is first evaluated on the GPU using our own custom evaluator (only the coarse data needs to be initially sent to the GPU), then, buffers for the final `MeshBufferCache` are filled on the GPU using a set of compute shaders. However, some buffers are still filled on the CPU side, if doing so on the GPU is impractical (e.g. the line adjacency buffer used for x-ray, whose logic is hardly GPU compatible). This is done at the mesh buffer extraction level so that the result can be readily used in the various OpenGL engines, without having to write custom geometry or tesselation shaders. We use our own subdivision evaluation shaders, instead of OpenSubDiv's vanilla one, in order to control the data layout, and interpolation. For example, we store vertex colors as compressed 16-bit integers, while OpenSubDiv's default evaluator only work for float types. In order to still access the modified geometry on the CPU side, for use in modifiers or transform operators, a dedicated wrapper type is added `MESH_WRAPPER_TYPE_SUBD`. Subdivision will be lazily evaluated via `BKE_object_get_evaluated_mesh` which will create such a wrapper if possible. If the final subdivision surface is not needed on the CPU side, `BKE_object_get_evaluated_mesh_no_subsurf` should be used. Enabling or disabling GPU subdivision can be done through the user preferences (under Viewport -> Subdivision). See patch description for benchmarks. Reviewed By: campbellbarton, jbakker, fclem, brecht, #eevee_viewport Differential Revision: https://developer.blender.org/D12406
2021-12-27 16:34:47 +01:00
#define GPU_vertbuf_init_with_format(verts, format) \
GPU_vertbuf_init_with_format_ex(verts, format, GPU_USAGE_STATIC)
blender::gpu::VertBuf *GPU_vertbuf_duplicate(blender::gpu::VertBuf *verts);
/**
* Create a new allocation, discarding any existing data.
*/
void GPU_vertbuf_data_alloc(blender::gpu::VertBuf &verts, uint v_len);
/**
* Resize buffer keeping existing data.
*/
void GPU_vertbuf_data_resize(blender::gpu::VertBuf &verts, uint v_len);
/**
* Set vertex count but does not change allocation.
* Only this many verts will be uploaded to the GPU and rendered.
* This is useful for streaming data.
*/
void GPU_vertbuf_data_len_set(blender::gpu::VertBuf &verts, uint v_len);
/**
* The most important #set_attr variant is the untyped one. Get it right first.
* It takes a void* so the app developer is responsible for matching their app data types
* to the vertex attribute's type and component count. They're in control of both, so this
* should not be a problem.
*/
void GPU_vertbuf_attr_set(blender::gpu::VertBuf *, uint a_idx, uint v_idx, const void *data);
/** Fills a whole vertex (all attributes). Data must match packed layout. */
void GPU_vertbuf_vert_set(blender::gpu::VertBuf *verts, uint v_idx, const void *data);
Overlay Engine: Refactor & Cleanup This is the unification of all overlays into one overlay engine as described in T65347. I went over all the code making it more future proof with less hacks and removing old / not relevent parts. Goals / Acheivements: - Remove internal shader usage (only drw shaders) - Remove viewportSize and viewportSizeInv and put them in gloabl ubo - Fixed some drawing issues: Missing probe option and Missing Alt+B clipping of some shader - Remove old (legacy) shaders dependancy (not using view UBO). - Less shader variation (less compilation time at first load and less patching needed for vulkan) - removed some geom shaders when I could - Remove static e_data (except shaders storage where it is OK) - Clear the way to fix some anoying limitations (dithered transparency, background image compositing etc...) - Wireframe drawing now uses the same batching capabilities as workbench & eevee (indirect drawing). - Reduced complexity, removed ~3000 Lines of code in draw (also removed a lot of unused shader in GPU). - Post AA to avoid complexity and cost of MSAA. Remaining issues: - ~~Armature edits, overlay toggles, (... others?) are not refreshing viewport after AA is complete~~ - FXAA is not the best for wires, maybe investigate SMAA - Maybe do something more temporally stable for AA. - ~~Paint overlays are not working with AA.~~ - ~~infront objects are difficult to select.~~ - ~~the infront wires sometimes goes through they solid counterpart (missing clear maybe?) (toggle overlays on-off when using infront+wireframe overlay in solid shading)~~ Note: I made some decision to change slightly the appearance of some objects to simplify their drawing. Namely the empty arrows end (which is now hollow/wire) and distance points of the cameras/spots being done by lines. Reviewed By: jbakker Differential Revision: https://developer.blender.org/D6296
2019-12-02 01:40:58 +01:00
/**
* Tightly packed, non interleaved input data.
*/
void GPU_vertbuf_attr_fill(blender::gpu::VertBuf *, uint a_idx, const void *data);
void GPU_vertbuf_attr_fill_stride(blender::gpu::VertBuf *,
uint a_idx,
uint stride,
const void *data);
/**
* For low level access only.
*
* \note This is obsolete, use #VertBuf::data<T>() instead.
*/
struct GPUVertBufRaw {
uint size;
uint stride;
2018-07-18 23:09:31 +10:00
unsigned char *data;
unsigned char *data_init;
#ifndef NDEBUG
/* Only for overflow check */
2018-07-18 23:09:31 +10:00
unsigned char *_data_end;
#endif
};
GPU_INLINE void *GPU_vertbuf_raw_step(GPUVertBufRaw *a)
{
2018-07-18 23:09:31 +10:00
unsigned char *data = a->data;
a->data += a->stride;
#ifndef NDEBUG
BLI_assert(data < a->_data_end);
#endif
return (void *)data;
}
GPU_INLINE uint GPU_vertbuf_raw_used(const GPUVertBufRaw *a)
{
2017-06-29 20:33:24 +10:00
return ((a->data - a->data_init) / a->stride);
}
void GPU_vertbuf_attr_get_raw_data(blender::gpu::VertBuf *, uint a_idx, GPUVertBufRaw *access);
const GPUVertFormat *GPU_vertbuf_get_format(const blender::gpu::VertBuf *verts);
uint GPU_vertbuf_get_vertex_alloc(const blender::gpu::VertBuf *verts);
uint GPU_vertbuf_get_vertex_len(const blender::gpu::VertBuf *verts);
GPUVertBufStatus GPU_vertbuf_get_status(const blender::gpu::VertBuf *verts);
void GPU_vertbuf_tag_dirty(blender::gpu::VertBuf *verts);
/**
* Should be rename to #GPU_vertbuf_data_upload.
*/
void GPU_vertbuf_use(blender::gpu::VertBuf *);
void GPU_vertbuf_bind_as_ssbo(blender::gpu::VertBuf *verts, int binding);
void GPU_vertbuf_bind_as_texture(blender::gpu::VertBuf *verts, int binding);
void GPU_vertbuf_wrap_handle(blender::gpu::VertBuf *verts, uint64_t handle);
OpenSubDiv: add support for an OpenGL evaluator This evaluator is used in order to evaluate subdivision at render time, allowing for faster renders of meshes with a subdivision surface modifier placed at the last position in the modifier list. When evaluating the subsurf modifier, we detect whether we can delegate evaluation to the draw code. If so, the subdivision is first evaluated on the GPU using our own custom evaluator (only the coarse data needs to be initially sent to the GPU), then, buffers for the final `MeshBufferCache` are filled on the GPU using a set of compute shaders. However, some buffers are still filled on the CPU side, if doing so on the GPU is impractical (e.g. the line adjacency buffer used for x-ray, whose logic is hardly GPU compatible). This is done at the mesh buffer extraction level so that the result can be readily used in the various OpenGL engines, without having to write custom geometry or tesselation shaders. We use our own subdivision evaluation shaders, instead of OpenSubDiv's vanilla one, in order to control the data layout, and interpolation. For example, we store vertex colors as compressed 16-bit integers, while OpenSubDiv's default evaluator only work for float types. In order to still access the modified geometry on the CPU side, for use in modifiers or transform operators, a dedicated wrapper type is added `MESH_WRAPPER_TYPE_SUBD`. Subdivision will be lazily evaluated via `BKE_object_get_evaluated_mesh` which will create such a wrapper if possible. If the final subdivision surface is not needed on the CPU side, `BKE_object_get_evaluated_mesh_no_subsurf` should be used. Enabling or disabling GPU subdivision can be done through the user preferences (under Viewport -> Subdivision). See patch description for benchmarks. Reviewed By: campbellbarton, jbakker, fclem, brecht, #eevee_viewport Differential Revision: https://developer.blender.org/D12406
2021-12-27 16:34:47 +01:00
/**
* XXX: do not use!
* This is just a wrapper for the use of the Hair refine workaround.
* To be used with #GPU_vertbuf_use().
*/
void GPU_vertbuf_update_sub(blender::gpu::VertBuf *verts, uint start, uint len, const void *data);
/* Metrics */
uint GPU_vertbuf_get_memory_usage();
/* Macros */
#define GPU_VERTBUF_DISCARD_SAFE(verts) \
do { \
if (verts != nullptr) { \
GPU_vertbuf_discard(verts); \
verts = nullptr; \
} \
} while (0)