fabi/Xenia-Canary
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fixed a bug with readback_resolve and readback_memexport that was responsible for a large portion of their overhead. readback_memexport and resolve are now usable for games, depending on your hardware. in my case games that were slideshows now run at like 20-30 fps, and my hardware isnt the best for xenia.

Browse Source 
add split_map class for mapping keys to values in a way that optimizes for frequent searches and infrequent insertions/removals
remove jump table implementation of GetColorRenderTargetFormatComponentCount, it was appearing relatively high in profiles. instead pack the component counts into a single 32 bit word, which is indexed by shifting
Add cvar to align all basic blocks to a boundary
Add mmio aware load paths
liberally apply XE_RESTRICT in ringbuffer related code
Removed the IS_TRUE and IS_FALSE opcodes, they were pointless duplicates of COMPARE_EQ/COMPARE_NE and i want to simplify our set of opcodes for future backends
More work on LVSR/LVSL/STVR/STVL opcodes
Optimized X64 translated code emission, now only compute instrkey once
Add code for pre-computing integer division magic numbers
Optimized GetHostViewportInfo a little
Move args for GetHostViewportInfo into a class, cache the result and compare for future queries. moved GetHostViewportInfo far lower on the profile
Add (currently not functional, and very racy) asynchronous memcpy code. will improve it and actually use it in future commits.
Add non-temporal memcpy function for huge page-aligned allocations. Used for copying to shared memory/readback
hoist are_accumulated_render_targets_valid_ check out of loop in render_target_cache already bound check.
Add stosb/movsb code for small constant memcpys/memsets that arent worth the overhead of memcpy/memset
This commit is contained in:
chss95cs@gmail.com
2022-08-28 14:24:25 -07:00
parent 335a390d43
commit f31869092c
32 changed files with 1576 additions and 507 deletions
Download Patch File Download Diff File Expand all files Collapse all files

96
src/xenia/gpu/draw_util.cc
View File

@@ -167,17 +167,17 @@ bool IsPixelShaderNeededWithRasterization(const Shader& shader,
return false;
}
void GetHostViewportInfo(const RegisterFile& regs,
uint32_t draw_resolution_scale_x,
uint32_t draw_resolution_scale_y,
bool origin_bottom_left, uint32_t x_max,
uint32_t y_max, bool allow_reverse_z,
reg::RB_DEPTHCONTROL normalized_depth_control,
bool convert_z_to_float24, bool full_float24_in_0_to_1,
bool pixel_shader_writes_depth,
static float ViewportRecip2_0(float f) {
float f1 = ArchReciprocalRefined(f);
return f1 + f1;
}
// chrispy: todo, the int/float divides and the nan-checked mins show up
// relatively high on uprof when i uc to 1.7ghz
void GetHostViewportInfo(GetViewportInfoArgs* XE_RESTRICT args,
ViewportInfo& viewport_info_out) {
assert_not_zero(draw_resolution_scale_x);
assert_not_zero(draw_resolution_scale_y);
assert_not_zero(args->draw_resolution_scale_x);
assert_not_zero(args->draw_resolution_scale_y);
// A vertex position goes the following path:
//
@@ -304,38 +304,32 @@ void GetHostViewportInfo(const RegisterFile& regs,
// TODO(Triang3l): Investigate the need for clamping of oDepth to 0...1 for
// D24FS8 as well.
auto pa_cl_clip_cntl = regs.Get<reg::PA_CL_CLIP_CNTL>();
auto pa_cl_vte_cntl = regs.Get<reg::PA_CL_VTE_CNTL>();
auto pa_su_sc_mode_cntl = regs.Get<reg::PA_SU_SC_MODE_CNTL>();
auto pa_su_vtx_cntl = regs.Get<reg::PA_SU_VTX_CNTL>();
auto pa_cl_clip_cntl = args->pa_cl_clip_cntl;
auto pa_cl_vte_cntl = args->pa_cl_vte_cntl;
auto pa_su_sc_mode_cntl = args->pa_su_sc_mode_cntl;
auto pa_su_vtx_cntl = args->pa_su_vtx_cntl;
// Obtain the original viewport values in a normalized way.
float scale_xy[] = {
pa_cl_vte_cntl.vport_x_scale_ena ? regs[XE_GPU_REG_PA_CL_VPORT_XSCALE].f32
: 1.0f,
pa_cl_vte_cntl.vport_y_scale_ena ? regs[XE_GPU_REG_PA_CL_VPORT_YSCALE].f32
: 1.0f,
pa_cl_vte_cntl.vport_x_scale_ena ? args->PA_CL_VPORT_XSCALE : 1.0f,
pa_cl_vte_cntl.vport_y_scale_ena ? args->PA_CL_VPORT_YSCALE : 1.0f,
};
float scale_z = pa_cl_vte_cntl.vport_z_scale_ena
? regs[XE_GPU_REG_PA_CL_VPORT_ZSCALE].f32
: 1.0f;
float scale_z =
pa_cl_vte_cntl.vport_z_scale_ena ? args->PA_CL_VPORT_ZSCALE : 1.0f;
float offset_base_xy[] = {
pa_cl_vte_cntl.vport_x_offset_ena
? regs[XE_GPU_REG_PA_CL_VPORT_XOFFSET].f32
: 0.0f,
pa_cl_vte_cntl.vport_y_offset_ena
? regs[XE_GPU_REG_PA_CL_VPORT_YOFFSET].f32
: 0.0f,
pa_cl_vte_cntl.vport_x_offset_ena ? args->PA_CL_VPORT_XOFFSET : 0.0f,
pa_cl_vte_cntl.vport_y_offset_ena ? args->PA_CL_VPORT_YOFFSET : 0.0f,
};
float offset_z = pa_cl_vte_cntl.vport_z_offset_ena
? regs[XE_GPU_REG_PA_CL_VPORT_ZOFFSET].f32
: 0.0f;
float offset_z =
pa_cl_vte_cntl.vport_z_offset_ena ? args->PA_CL_VPORT_ZOFFSET : 0.0f;
// Calculate all the integer.0 or integer.5 offsetting exactly at full
// precision, separately so it can be used in other integer calculations
// without double rounding if needed.
float offset_add_xy[2] = {};
if (pa_su_sc_mode_cntl.vtx_window_offset_enable) {
auto pa_sc_window_offset = regs.Get<reg::PA_SC_WINDOW_OFFSET>();
auto pa_sc_window_offset = args->pa_sc_window_offset;
offset_add_xy[0] += float(pa_sc_window_offset.window_x_offset);
offset_add_xy[1] += float(pa_sc_window_offset.window_y_offset);
}
@@ -346,8 +340,11 @@ void GetHostViewportInfo(const RegisterFile& regs,
// The maximum value is at least the maximum host render target size anyway -
// and a guest pixel is always treated as a whole with resolution scaling.
uint32_t xy_max_unscaled[] = {x_max / draw_resolution_scale_x,
y_max / draw_resolution_scale_y};
// cbrispy: todo, this integer divides show up high on the profiler somehow
// (it was a very long session, too)
uint32_t xy_max_unscaled[] = {
args->draw_resolution_scale_x_divisor.Apply(args->x_max),
args->draw_resolution_scale_y_divisor.Apply(args->y_max)};
assert_not_zero(xy_max_unscaled[0]);
assert_not_zero(xy_max_unscaled[1]);
@@ -367,9 +364,11 @@ void GetHostViewportInfo(const RegisterFile& regs,
std::min(xenos::kTexture2DCubeMaxWidthHeight, xy_max_unscaled[i]);
viewport_info_out.xy_extent[i] =
extent_axis_unscaled *
(i ? draw_resolution_scale_y : draw_resolution_scale_x);
(i ? args->draw_resolution_scale_y : args->draw_resolution_scale_x);
float extent_axis_unscaled_float = float(extent_axis_unscaled);
float pixels_to_ndc_axis = 2.0f / extent_axis_unscaled_float;
float pixels_to_ndc_axis = ViewportRecip2_0(extent_axis_unscaled_float);
ndc_scale[i] = scale_xy[i] * pixels_to_ndc_axis;
ndc_offset[i] = (offset_base_xy[i] - extent_axis_unscaled_float * 0.5f +
offset_add_xy[i]) *
@@ -394,7 +393,7 @@ void GetHostViewportInfo(const RegisterFile& regs,
// doing truncation for simplicity - since maxing with 0 is done anyway
// (we only return viewports in the positive quarter-plane).
uint32_t axis_resolution_scale =
i ? draw_resolution_scale_y : draw_resolution_scale_x;
i ? args->draw_resolution_scale_y : args->draw_resolution_scale_x;
float offset_axis = offset_base_xy[i] + offset_add_xy[i];
float scale_axis = scale_xy[i];
float scale_axis_abs = std::abs(scale_xy[i]);
@@ -422,11 +421,14 @@ void GetHostViewportInfo(const RegisterFile& regs,
// space, a region previously outside -W...W should end up within it, so
// the scale should be < 1.
float axis_extent_rounded = float(axis_extent_int);
ndc_scale_axis = scale_axis * 2.0f / axis_extent_rounded;
float inv_axis_extent_rounded =
ArchReciprocalRefined(axis_extent_rounded);
ndc_scale_axis = scale_axis * 2.0f * inv_axis_extent_rounded;
// Move the origin of the snapped coordinates back to the original one.
ndc_offset_axis = (float(offset_axis) -
(float(axis_0_int) + axis_extent_rounded * 0.5f)) *
2.0f / axis_extent_rounded;
2.0f * inv_axis_extent_rounded;
} else {
// Empty viewport (everything outside the viewport scissor).
ndc_scale_axis = 1.0f;
@@ -497,7 +499,7 @@ void GetHostViewportInfo(const RegisterFile& regs,
ndc_scale[2] = 0.5f;
ndc_offset[2] = 0.5f;
}
if (pixel_shader_writes_depth) {
if (args->pixel_shader_writes_depth) {
// Allow the pixel shader to write any depth value since
// PA_SC_VPORT_ZMIN/ZMAX isn't present on the Adreno 200; guest pixel
// shaders don't have access to the original Z in the viewport space
@@ -515,7 +517,7 @@ void GetHostViewportInfo(const RegisterFile& regs,
// Direct3D 12 doesn't allow reverse depth range - on some drivers it
// works, on some drivers it doesn't, actually, but it was never
// explicitly allowed by the specification.
if (!allow_reverse_z && z_min > z_max) {
if (!args->allow_reverse_z && z_min > z_max) {
std::swap(z_min, z_max);
ndc_scale[2] = -ndc_scale[2];
ndc_offset[2] = 1.0f - ndc_offset[2];
@@ -523,10 +525,9 @@ void GetHostViewportInfo(const RegisterFile& regs,
}
}
if (normalized_depth_control.z_enable &&
regs.Get<reg::RB_DEPTH_INFO>().depth_format ==
xenos::DepthRenderTargetFormat::kD24FS8) {
if (convert_z_to_float24) {
if (args->normalized_depth_control.z_enable &&
args->depth_format == xenos::DepthRenderTargetFormat::kD24FS8) {
if (args->convert_z_to_float24) {
// Need to adjust the bounds that the resulting depth values will be
// clamped to after the pixel shader. Preferring adding some error to
// interpolated Z instead if conversion can't be done exactly, without
@@ -537,7 +538,7 @@ void GetHostViewportInfo(const RegisterFile& regs,
z_min = xenos::Float20e4To32(xenos::Float32To20e4(z_min, true));
z_max = xenos::Float20e4To32(xenos::Float32To20e4(z_max, true));
}
if (full_float24_in_0_to_1) {
if (args->full_float24_in_0_to_1) {
// Remap the full [0...2) float24 range to [0...1) support data round-trip
// during render target ownership transfer of EDRAM tiles through depth
// input without unrestricted depth range.
@@ -548,7 +549,7 @@ void GetHostViewportInfo(const RegisterFile& regs,
viewport_info_out.z_min = z_min;
viewport_info_out.z_max = z_max;
if (origin_bottom_left) {
if (args->origin_bottom_left) {
ndc_scale[1] = -ndc_scale[1];
ndc_offset[1] = -ndc_offset[1];
}
@@ -557,7 +558,6 @@ void GetHostViewportInfo(const RegisterFile& regs,
viewport_info_out.ndc_offset[i] = ndc_offset[i];
}
}
void GetScissor(const RegisterFile& regs, Scissor& scissor_out,
bool clamp_to_surface_pitch) {
auto pa_sc_window_scissor_tl = regs.Get<reg::PA_SC_WINDOW_SCISSOR_TL>();
@@ -868,7 +868,7 @@ bool GetResolveInfo(const RegisterFile& regs, const Memory& memory,
xenos::kMaxResolveSize);
y1 = y0 + int32_t(xenos::kMaxResolveSize);
}
//fails in forza horizon 1
// fails in forza horizon 1
assert_true(x0 < x1 && y0 < y1);
if (x0 >= x1 || y0 >= y1) {
XELOGE("Resolve region is empty");