Skip ubershader mode works the same as hybrid ubershaders in that the
shaders are compiled asynchronously. However, instead of using the
ubershader to draw the object, it skips it entirely until the
specialized shader is made available.
This mode will likely result in broken effects where a game creates an
EFB copy, and does not redraw it every frame. Therefore, it is not a
recommended option, however, it may result in better performance on
low-end systems.
Depending on which constructor is invoked, m_id or m_compute_program_id
can end up in an uninitialized state. We should ensure that the object
is completely initialized to something deterministic regardless of the
constructor taken.
This replaces usages of the non-standard __FUNCTION__ macro with the standard
mandated __func__ identifier.
__FUNCTION__ is a preprocessor definition that is provided as an
extension by compilers. This was the only convenient option to rely on
pre-C++11. However, C++11 and greater mandate the predefined identifier
__func__, which lets us accomplish the same thing.
The difference between the two, however, is that __func__ isn't a
preprocessor macro, it's an actual identifier that exists at function
scope. The C++17 draft standard (N4659) at section [dcl.fct.def.general]
paragraph 8 states:
"
The function-local predefined variable __func__ is defined as if a
definition of the form
static const char __func__[] = "function-name ";
had been provided, where function-name is an implementation-defined
string. It is unspecified whether such
a variable has an address distinct from that of any other object in the
program.
"
Thankfully, we don't do any macro or string concatenation with __FUNCTION__
that can't be modified to use __func__.
Fixes a crash which could occur in platforms which do not support
buffer_storage, and EFB2RAM is enabled (which indirectly uses the
attributeless buffer).
Yes, this commit is only to blame OSX and Mali. Through the former supports unsynchronized mappings, the latter supports *no* way to stream dynamic data at all. Let's try to make bad news, as they ignore friendly feature requests. Maybe we just need to make more noise...
We would want to improve the granularity here in the future, but for
now, this should avoid any performance loss from switching to the
VideoCommon shader cache.
tl;dr: This PR speedups dolphin on mobiles with the Mali GPU and ES 3.2
drivers by a factor of 10 by using the method with the biggest overhead.
Please keep care not to buy this shit!
The ARM driver team seems to care very well about their customers. But
bad luck, users and open source developers are *not* their customers. So
even device-independent feature requests are just ignored for *years*:
https://community.arm.com/graphics/f/discussions/4645/gl_ext_buffer_storage-support
The bad point, they neither implement any of the other common ways to
stream dynamic content in unextented GL:
- They just ignore the GL_MAP_UNSYNCHRONIZED_BIT flag
- They don't support on-device buffer updates and just stall with
glBufferSubData
It seems like no benchmark is using any dynamic content - and like no
customer cares about anything but benchmarks, or users...
We have a flag to disable the glBufferSubData way, this PR adds the flag
to also disable the unsychronized mapping way. The second one is
available since their ES 3.2 update, but slow as hell.
So how to continue? The last remaining technical way to stream dynamic
content at all is to alloc a new buffer per draw call with glBufferData.
This is very gross, but still a factor 10 speedup compared to stalling
the GPU. Small tests shows that you can expect another 3-5 times speedup
with EXT_buffer_data, so Mali would be on pair with Adreno here. So if
you have bought such a device unfortunately, please try to make noise on
your vendor forums/support and ask for this extension. If you are going
to buy a new mobile, I'd recormend to avoid *any* mobile with a Mali GPU
in it.
We now differentiate between a resize event and surface change/destroyed
event, reducing the overhead for resizes in the Vulkan backend. It is
also now now safe to change the surface multiple times if the video thread
is lagging behind.
This could cause glReadPixels() calls which assume no buffer is bound
(e.g. CPU EFB access) to fail. The problem was limited to devices which
don't support persistent mapping, as the map path is not otherwise.
Both libusbhid (system library) and libhidapi (3rd party library)
provide a function called hid_init. Dolphin was being linked to both.
The WiimoteScannerHidapi constructor was calling hid_init without
arguments. libusbhid's hid_init expects one argument (a file path).
It was being called as if it was defined without arguments, which
resulted in a garbage path being passed in, and because of that,
the Qt GUI was failing to launch with the following error:
'dolphin-emu-qt2: @ : No such file or directory'
The console appears to behave against standard IEEE754 specification
here, in particular around how NaNs are handled. NaNs appear to have no
effect on the result, and are treated the same as positive or negative
infinity, based on the sign bit.
However, when the result would be NaN (inf - inf, or (-inf) - (-inf)),
this results in a completely fogged color, or unfogged color
respectively. We handle this by returning a constant zero for the A
varaible, and positive or negative infinity for C depending on the sign
bits of the A and C registers. This ensures that no NaN value is passed
to the GPU in the first place, and that the result of the fog
calculation cannot be NaN.
It seems it doesn't like modifying inout variables in place - so instead
use a temporary for ocol0/ocol1 and only write them once at the end of
the shader
This will generate one shader per copy format. For now, it is the same
shader with the colmat hard coded. So it should already improve the GPU
performance a bit, but a rewrite of the shader generator is suggested.
Half of the patch is done by linkmauve1:
VideoCommon: Reorganise the shader writes.
Also skips swapping the window system buffers in headless mode, as there
may not be a surface which can be swapped in the first place. Instead,
we call glFlush() at the end of a frame in this case.
Cel-damage uses the color from the lighting stage of the vertex pipeline
as texture coordinates, but sets numColorChans to zero.
We now calculate the colors in all cases, but override the color before
writing it from the vertex shader if numColorChans is set to a lower value.
All file scope variables are able to be made internally linked.
CD3DFont is essentially used as an extension to the utility interface, so
this is able to be made internal as well, removing a global from
external view.
Some lines of code in Dolphin just plainly grabbed the value of
g_ActiveConfig.iEFBScale, which resulted in Auto being treated as
0x rather than the actual automatically selected scale.
Drivers can return VK_ERROR_OUT_OF_DATE_KHR from vkQueuePresentKHR, and
we should resize the image in this case, as well as when getting it back
from vkAcquireNextImageKHR.
These rely on instance state, or are used within instance-based class
member functions, so they should belong to the instance itself instead
of being file statics.
Ideally Common.h wouldn't be a header in the Common library, and instead be renamed to something else, like PlatformCompatibility.h or something, but even then, there's still some things in the header that don't really fall under that label
This moves the version strings out to their own version header that doesn't dump a bunch of other unrelated things into scope, like what Common.h was doing.
This also places them into the Common namespace, as opposed to letting them sit in the global namespace.
SetFormat() is only ever used internally. ResetBuffer() is only
used to implement the VertexManagerBase class interface, so
there's no need to make it protected.
If we allocate a large amount of memory (A), commit a smaller amount,
then allocate memory smaller than allocation A, we will have already
waited for these fences in A, but not used the space. In this case,
don't set m_free_iterator to a position before that which we know is
safe to use, which would result in waiting on the same fence(s) next
time.
Calling vkCmdClearAttachments with a partial rect, or specifying a
render area in a render pass with the load op set to clear can cause the
GPU to lock up, or raise a bounds violation. This only occurs on MSAA
framebuffers, and it seems when there are multiple clears in a single
command buffer. Worked around by back to the slow path (drawing quads)
when MSAA is enabled.
Before this change, we simply fail if the device does not expose one
queue family that supports both graphics and present. Currently this is
fine, since devices tend to lay out their queues in this way. NV, for
instance, tends to have one queue family for all graphics operations and
one more for transfer only. However, it's not a hard requirement, and it
is cheap to use a separate queue, so we might as well.
Currently, this is only the logic op bit, but this will be extended to
the framebuffer fetch/blend modes. In the future, when/if we move to
VideoCommon pipelines, this state will be part of the pipeline UID
anyway, and we can mask it out in the backend by using a two-level map,
so the shaders/programs are shared.
This reverts commit d23fd17e1a.
Dynamic sampler indexing is broken in VK_NV_glsl as of 385.41. The
performance gap doesn't seem to be as wide with the updated driver, so
to save maintaining two code paths, it's easier to just drop the
extension support completely.
ImgTec's driver uses a major.minor@changeID versioning system
This is packed into a double so "1.9@4850625" becomes "109.4850625"
The next release brnach is expected to be 1.10, hence the need for 2
digits for the branch minor.
The changeID should be unique for each build, but is shared over all
branches, so only makes sense to compare withing a branch.
It's likely branch 'major' versions will be used for major hardware
revisions, and the drivers for both maintained in parallel. Thus it
may not make sense to compare versions between different major
verisons - if/when this happens we can hook up a DriverDetails::Family
as needed.
This optimisation doesn't work on PowerVR's Vulkan implementation. We
(incorrectly) disallow Framebuffer objects to be used with a different
load or store op than that which they were created with, despite the
spec allowing such.
This fixes the windwaker intro "smearing"
The class NonCopyable is, like the name says, supposed to disallow
copying. But should it allow moving?
For a long time, NonCopyable used to not allow moving. (It declared
a deleted copy constructor and assigment operator without declaring
a move constructor and assignment operator, making the compiler
implicitly delete the move constructor and assignment operator.)
That's fine if the classes that inherit from NonCopyable don't need
to be movable or if writing the move constructor and assignment
operator by hand is fine, but that's not the case for all classes,
as I discovered when I was working on the DirectoryBlob PR.
Because of that, I decided to make NonCopyable movable in c7602cc,
allowing me to use NonCopyable in DirectoryBlob.h. That was however
an unfortunate decision, because some of the classes that inherit
from NonCopyable have incorrect behavior when moved by default-
generated move constructors and assignment operators, and do not
explicitly delete the move constructors and assignment operators,
relying on NonCopyable being non-movable.
So what can we do about this? There are four solutions that I can
think of:
1. Make NonCopyable non-movable and tell DirectoryBlob to suck it.
2. Keep allowing moving NonCopyable, and expect that classes that
don't support moving will delete the move constructor and
assignment operator manually. Not only is this inconsistent
(having classes disallow copying one way and disallow moving
another way), but deleting the move constructor and assignment
operator manually is too easy to forget compared to how tricky
the resulting problems are.
3. Have one "MovableNonCopyable" and one "NonMovableNonCopyable".
It works, but it feels rather silly...
4. Don't have a NonCopyable class at all. Considering that deleting
the copy constructor and assignment operator only takes two lines
of code, I don't see much of a reason to keep NonCopyable. I
suppose that there was more of a point in having NonCopyable back
in the pre-C++11 days, when it wasn't possible to use "= delete".
I decided to go with the fourth one (like the commit title says).
The implementation of the commit is fairly straight-forward, though
I would like to point out that I skipped adding "= delete" lines
for classes whose only reason for being uncopyable is that they
contain uncopyable classes like File::IOFile and std::unique_ptr,
because the compiler makes such classes uncopyable automatically.
The casts to u32* are technically undefined behavior. The u8* cast is
left, as char/unsigned char is exempted from this rule to allow for
bvtewise inspection of objects (and this is what s8/u8 are typedefs of
on platforms we support).
Improve bookkeeping around formats. Hopefully make code less confusing.
- Rename TlutFormat -> TLUTFormat to follow conventions.
- Use enum classes to prevent using a Texture format where an EFB Copy format
is expected or vice-versa.
- Use common EFBCopyFormat names regardless of depth and YUV configurations.
This way it allows us to use surfaceless contexts in EGL/GLX. It also
ensures that the shared context shares a similar setup to the main
context's framebuffer, potentially reducing the number of variants a
driver needs to generate.
Previously we were falling back to an earlier version of the compiler.
The older version cannot compile our ubershaders without various
graphical issues.
This was mainly included for debugging, but could end up being confusing
for users, as well as polluting the GL program cache with a mix of uber
and specialized shaders if the option was changed.
pSysMem is of the type const void* -- because of this, it makes the
original delete[] call undefined behavior, as deleting a void pointer is
undefined behavior.
Also punning types into existence, like what was done for the stereo
image header is undefined behavior as well. The proper way to do this is
to either manually add all individual bytes manually, or memcpy the
struct into memory.
As we want to deallocate the memory before returning, and because
pSysMem is a const void*, we keep a unique_ptr to the data and just pass
pSysMem a raw pointer to the data.
This is because we re-use BlendingState for our internal drawing (e.g.
RasterFont) and for these shaders, we can't assume the presence of a
second color output.
This removes the need for token pasting, which isn't supported in GLSL
ES. Shouldn't cause any issues unless people are using reserved keywords
as option names.
The loop was allocating one-too-many levels, as well as incorrect sizes
for each level. Probably not an issue as mipmapped render targets aren't
used, but the logic should be correct anyway.
This was a regression from the remove-everything-static-from-renderer
PR. As the comment indicates, it would be nice to move all of this logic
out of the Renderer constructor, but this is a much larger change.
The FrameBufferManager::CreateTexture (from the OpenGL backend) method introduced by commit 69cedf41 incorrectly compares the texture variable (which contains a name provided by glGenTextures) against GL_TEXTURE_2D_MULTISAMPLE_ARRAY and GL_TEXTURE_2D_MULTISAMPLE.
It should instead use the texture_type variable for this (as done in the first branch of the if).
This commit should have zero performance effect if SSBOs are supported.
If they aren't (e.g. on all Macs), this commit alters FramebufferManager
to attach a new stencil buffer and VertexManager to draw to it when
bounding box is active. `BBoxRead` gets the pixel data from the buffer
and dumbly loops through it to find the bounding box.
This patch can run Paper Mario: The Thousand-Year Door at almost full
speed (50–60 FPS) without Dual-Core enabled for all common bounding
box-using actions I tested (going through pipes, Plane Mode, Paper
Mode, Prof. Frankly's gate, combat, walking around the overworld, etc.)
on my computer (macOS 10.12.3, 2.8 GHz Intel Core i7, 16 GB 1600 MHz
DDR3, and Intel Iris 1536 MB).
A few more demanding scenes (e.g. the self-building bridge on the way
to Petalburg) slow to ~15% of their speed without this patch (though
they don't run quite at full speed even on master). The slowdown is
caused almost solely by `glReadPixels` in `OGL::BoundingBox::Get`.
Other implementation ideas:
- Use a stencil buffer that's separate from the depth buffer. This would
require ARB_texture_stencil8 / OpenGL 4.4, which isn't available on
macOS.
- Use `glGetTexImage` instead of `glReadPixels`. This is ~5 FPS slower
on my computer, presumably because it has to transfer the entire
combined depth-stencil buffer instead of only the stencil data.
Getting only stencil data from `glGetTexImage` requires
ARB_texture_stencil8 / OpenGL 4.4, which (again) is not available on
macOS.
- Don't use a PBO, and use `glReadPixels` synchronously. This has no
visible performance effect on my computer, and is theoretically
slower.
This stops the virtual method call from within the Renderer constructor.
The initialization here for GL had to be moved to VideoBackend, as the
Renderer constructor will not have been executed before the value is
required.
This moves all the byte swapping utilities into a header named Swap.h.
A dedicated header is much more preferable here due to the size of the
code itself. In general usage throughout the codebase, CommonFuncs.h was
generally only included for these functions anyway. These being in their
own header avoids dumping the lesser used utilities into scope. As well
as providing a localized area for more utilities related to byte
swapping in the future (should they be needed). This also makes it nicer
to identify which files depend on the byte swapping utilities in
particular.
Since this is a completely new header, moving the code uncovered a few
indirect includes, as well as making some other inclusions unnecessary.
Before #4581, an invocation of `SetBlendMode` could invoke
`glBlendEquationSeparate` and `glBlendFuncSeparate` even when it was
setting `glDisable(GL_BLEND)`. I couldn't figure out how to map the old
behavior over to the new BlendingState code, so I changed it to always
call the two blend functions.
Fixes https://bugs.dolphin-emu.org/issues/10120 : "Sonic Adventure 2
Battle: graphics crash when loading first Dark level".
We (the Microsoft C++ team) use the dolphin project as part of our "Real world code" tests.
I noticed a few issues in windows specific code when building dolphin with the MSVC compiler
in its conformance mode (/permissive-). For more information on /permissive- see our blog
https://blogs.msdn.microsoft.com/vcblog/2016/11/16/permissive-switch/.
These changes are to address 3 different types of issues:
1) Use of qualified names in member declarations
struct A {
void A::f() { } // error C4596: illegal qualified name in member declaration
// remove redundant 'A::' to fix
};
2) Binding a non-const reference to a temporary
struct S{};
// If arg is in 'in' parameter, then it should be made const.
void func(S& arg){}
int main() {
//error C2664: 'void func(S &)': cannot convert argument 1 from 'S' to 'S &'
//note: A non-const reference may only be bound to an lvalue
func( S() );
//Work around this by creating a local, and using it to call the function
S s;
func( s );
}
3) Add missing #include <intrin.h>
Because of the workaround you are using in the code you will need to include
this. This is because of changes in the libraries and not /permissive-
Keeps associated data together. It also eliminates the possibility of out
parameters not being initialized properly. For example, consider the
following example:
-- some FramebufferManager implementation --
void FBMgrImpl::GetTargetSize(u32* width, u32* height) override
{
// Do nothing
}
-- somewhere else where the function is used --
u32 width, height;
framebuffer_manager_instance->GetTargetSize(&width, &height);
if (texture_width != width) <-- Uninitialized variable usage
{
...
}
It makes it much more obvious to spot any initialization issues, because
it requires something to be returned, as opposed to allowing an
implementation to just not do anything.
Hopefully will fix the crash in vkDestroyInstance on the NV Shield TV,
and likely reduce boot times slightly for drivers that take a while
to create instances.
This happened when the geometry shader was disabled, and the uniform
buffer was grown to a larger size. The update would be skipped, leaving
the old buffer to be included in the descriptor set.
This also moves the pipeline and descriptor set layouts used for texture
conversion (texel buffers) to ObjectCache, and shares a binding location
with the SSBO set.
Should we ever introduce anything else that has to be done when a command
buffer is executed (e.g. invalidating constants from previous commit), we
don't have to update all the callers.
This could happen because the vertex memory was already committed, if a
uniform buffer allocation failed and caused a command buffer to be
executed, it would be associated with the previous command buffer rather
than the buffer containing the draw that consumed these vertices.
This was happening when a fence wait happened mid-frame. The data written
between the fence being queued and the allocation occuring was incorrectly
assumed to be consumed by the GPU.
JosJuice