Add instance draw counts to vertex buffers

This commit is contained in:
CamilleLaVey
2026-08-29 02:53:00 -04:00
parent 064ab33c08
commit 02fef4418d
4 changed files with 44 additions and 3 deletions
+32 -3
View File
@@ -1244,6 +1244,20 @@ void BufferCache<P>::UpdateIndexBuffer() {
template <class P>
void BufferCache<P>::UpdateVertexBuffers() {
auto& flags = maxwell3d->dirty.flags;
const u32 base_instance = maxwell3d->draw_manager.draw_state.base_instance;
if (draw_instance_count != last_draw_instance_count ||
base_instance != last_draw_base_instance) {
last_draw_instance_count = draw_instance_count;
last_draw_base_instance = base_instance;
const auto& instances = maxwell3d->regs.vertex_stream_instances;
for (u32 index = 0; index < NUM_VERTEX_BUFFERS; ++index) {
if (!instances.IsInstancingEnabled(index)) {
continue;
}
flags[Dirty::VertexBuffer0 + index] = true;
flags[Dirty::VertexBuffers] = true;
}
}
if (!maxwell3d->dirty.flags[Dirty::VertexBuffers]) {
return;
}
@@ -1270,9 +1284,24 @@ void BufferCache<P>::UpdateVertexBuffer(u32 index) {
return;
}
// TODO: Analyze stride and number of vertices
u64 address_size = (std::min)(gpu_addr_end - gpu_addr_begin,
u64{(std::numeric_limits<u32>::max)()});
if (!gpu_memory->IsWithinGPUAddressRange(gpu_addr_end) || address_size >= 64_MiB) {
constexpr u64 implausible_size = 64_MiB;
u64 address_size = gpu_addr_end - gpu_addr_begin;
if (address_size > u64{(std::numeric_limits<u32>::max)()}) {
address_size = implausible_size;
}
const bool is_instanced = maxwell3d->regs.vertex_stream_instances.IsInstancingEnabled(index);
const u64 stride = static_cast<u64>(array.stride);
if (is_instanced && stride != 0 && draw_instance_count != 0) {
const u64 base_instance = static_cast<u64>(maxwell3d->draw_manager.draw_state.base_instance);
u64 elements = base_instance + 1;
if (array.frequency != 0) {
elements = (base_instance + static_cast<u64>(draw_instance_count) - 1) /
static_cast<u64>(array.frequency) +
1;
}
address_size = (std::min)(address_size, (elements + 1) * stride);
}
if (!gpu_memory->IsWithinGPUAddressRange(gpu_addr_end) || address_size >= implausible_size) {
address_size = gpu_memory->MaxContinuousRange(gpu_addr_begin, address_size);
}
const u32 size = static_cast<u32>(address_size);
@@ -307,6 +307,10 @@ public:
current_draw_indirect = current_draw_indirect_;
}
void SetDrawInstanceCount(u32 draw_instance_count_) {
draw_instance_count = draw_instance_count_;
}
[[nodiscard]] std::pair<Buffer*, u32> GetDrawIndirectCount();
[[nodiscard]] std::pair<Buffer*, u32> GetDrawIndirectBuffer();
@@ -480,6 +484,10 @@ private:
const Tegra::Engines::Maxwell3D::DrawManager::IndirectParams* current_draw_indirect{};
u32 draw_instance_count = 0;
u32 last_draw_instance_count = 0;
u32 last_draw_base_instance = 0;
u32 last_index_count = 0;
u32 enabled_vertex_buffers_mask = 0;
@@ -259,6 +259,7 @@ void RasterizerOpenGL::PrepareDraw(bool is_indexed, Func&& draw_func) {
}
void RasterizerOpenGL::Draw(bool is_indexed, u32 instance_count) {
buffer_cache.SetDrawInstanceCount(instance_count);
PrepareDraw(is_indexed, [this, is_indexed, instance_count](GLenum primitive_mode) {
const auto& draw_state = maxwell3d->draw_manager.draw_state;
const GLuint base_instance = GLuint(draw_state.base_instance);
@@ -304,6 +305,7 @@ void RasterizerOpenGL::Draw(bool is_indexed, u32 instance_count) {
void RasterizerOpenGL::DrawIndirect() {
const auto& params = maxwell3d->draw_manager.indirect_state;
buffer_cache.SetDrawIndirect(&params);
buffer_cache.SetDrawInstanceCount(0);
PrepareDraw(params.is_indexed, [this, &params](GLenum primitive_mode) {
if (params.is_byte_count) {
const GPUVAddr tfb_object_base_addr = params.indirect_start_address - 4U;
@@ -260,6 +260,7 @@ void RasterizerVulkan::PrepareDraw(bool is_indexed, Func&& draw_func) {
}
void RasterizerVulkan::Draw(bool is_indexed, u32 instance_count) {
buffer_cache.SetDrawInstanceCount(instance_count);
PrepareDraw(is_indexed, [this, is_indexed, instance_count] {
const auto& draw_state = maxwell3d->draw_manager.draw_state;
const u32 num_instances{instance_count};
@@ -295,6 +296,7 @@ void RasterizerVulkan::Draw(bool is_indexed, u32 instance_count) {
void RasterizerVulkan::DrawIndirect() {
const auto& params = maxwell3d->draw_manager.indirect_state;
buffer_cache.SetDrawIndirect(&params);
buffer_cache.SetDrawInstanceCount(0);
PrepareDraw(params.is_indexed, [this, &params] {
const auto indirect_buffer = buffer_cache.GetDrawIndirectBuffer();
const auto& buffer = indirect_buffer.first;