301c875ecb
GPU: For native-sized captures, always read from native VRAM -- never from custom VRAM. Should make captures a little faster when running a custom framebuffer and/or using RGB888 color format.
2017-07-21 12:49:57 -07:00
1b483d266a
GPU: Fix bug where sprite blending would no longer work. (Regression from commit 624dd57.)
2017-07-21 09:51:13 -07:00
624dd5788c
GPU: Another significant performance improvement to GPU compositing in many games.
...
- Now composites pixels using explicit functions for simple copy,
masked copy, and masked effect.
- On SSE2 systems, pixels composited using a simple copy no longer
require the destination pixels to be loaded first, since all the pixels
are guaranteed to be overwritten anyways.
- Try and move the window test as far up the pipeline as possible so
that pixel rendering can bail as soon as possible if the window test
fails.
- Clean up and further standardize the code for compositing BG layers,
OBJ layers, and 3D layers.
2017-07-20 23:45:41 -07:00
80a4c3f071
GPU: Fix bug where 3D rendering sometimes does not appear correctly on SSE2 systems. (Regression from commit 5c6fcb0.)
2017-07-20 00:55:08 -07:00
5c6fcb0f47
GPU: Significantly improve GPU compositing performance in many games.
...
- Adds more possible conditions for compositing to take the fast path
(simple pixel copying).
- Adds SSE2 optimizations for the 2x and 4x scaling cases in
GPUEngineBase::_LineColorCopy().
2017-07-19 23:41:43 -07:00
c3bce87556
Cocoa Port: Expose the new feature to change the GPU color depth in the UI. (Related to commit 3b37058.)
2017-07-18 20:57:55 -07:00
3b37058cbc
GPU: Custom VRAM now works with 24-bit color depth. It can be captured natively in 24-bit and used as a source buffer in 24-bit. (Related to r5433. This rework is now COMPLETE!)
...
- NEW FEATURE: Clients may now call GPUSubsystem::SetColorFormat() to
choose the color format of the GPU output, which can be RGB555
(15-bit), RGB666 (18-bit), or RGB888 (24-bit).
- On a special note, the Deposterize filter for 3D textures can now
show its true smoothing effect when clients run the GPU in 24-bit color
mode.
2017-07-18 20:56:11 -07:00
e7f1737bdb
Cocoa Port: The Vertical Sync setting has been removed -- it is now automatically set.
...
- New behavior for layer-backed views (OpenGL on Mountain Lion and
later, or Metal): Vertical Sync is always enabled.
- New behavior for non-layer-backed views (OpenGL on Lion and earlier):
Vertical Sync is only enabled if frameskip is enabled or if the
execution speed is set to 1x or less while the speed limiter is engaged.
- Remove all associated UI for manually setting Vertical Sync.
- Also add a new menu option to the Tools menu for disabling Metal,
instead forcing display views to use OpenGL. (For developer builds
only.)
2017-07-18 17:10:26 -07:00
ddac7122d3
Cocoa Port: Ensure that display windows avoid drawing with uninitialized display info and possibly non-existent buffers on startup.
2017-07-18 01:26:42 -07:00
7791f60495
Cocoa Port: Avoid sending a torrent of MESSAGE_RELOAD_REPROCESS_REDRAW messages whenever a new display window is created.
2017-07-18 01:24:08 -07:00
68f201c3ce
SoftRasterizer: Fix incorrect coloring when using an output format of RGB888. (Regression from commit eaa5210.)
2017-07-15 17:02:56 -07:00
eaa5210fb0
GPU: Finishing 3D rendering and flushing the 3D framebuffers are now split into separate operations.
...
- Fixes the background appearance when receiving items in Pokemon
Black/White 2. (Fixes SF bug #1624.)
2017-07-15 16:00:26 -07:00
0f05656a17
Save States: Fix save state compatibility for files created between 32-bit and 64-bit systems. (Regression from r5363.)
2017-07-15 15:38:39 -07:00
63b232c1b8
GPU: Fix possible crash by properly initializing working capture buffers. (Regression from commit 28f5c3a.)
2017-07-08 10:56:29 -07:00
652af9ec71
Cocoa Port: Always allow client display views to process displays if the client did not request a custom framebuffer size.
...
- Also try to solve some synchronization issues with Metal display
views.
2017-07-07 13:13:29 -07:00
1f60446f3f
GPU: Fix bug on SSE2 builds where the window test mask would be undefined if the output color format is RGB666 or RGB888.
2017-07-05 23:30:31 -07:00
28f5c3a895
GPU: Don't do ad-hoc allocations of working capture lines in GPUEngineA::_RenderLine_DisplayCapture(). Allocate the lines in GPUEngineA::SetCustomFramebufferSize() instead.
...
- Also do some minor refactoring and code cleanup.
2017-07-05 23:22:49 -07:00
ee235d2abc
fix gdbstub compiling in posix frontend. didnt really test it... should fix #73
2017-07-03 09:37:51 -05:00
c37e892090
GPU: Partially reverts changes to the LCDC check, bringing back an optimization for display capture.
...
- Revert commit abe2e61997adf682eb23
2017-07-01 13:51:15 -07:00
76063a389c
Cocoa Port: Fix bug that causes building to fail. (Regression from commit c365617.)
...
- Also do some minor code cleanup to byteSizeStringWithLargerUnit:.
2017-07-01 13:44:52 -07:00
73c8f8af37
Merge pull request #71 from kongfl888/screenshot
...
winport: fix the screenshot type is not saved.
2017-06-09 23:16:00 -05:00
65bd5b565c
winport: fix the screenshot type is not saved.
2017-06-10 10:47:23 +08:00
d6872769ae
cheats: fix bug in EXXXXXXX code eating one too many code lines
2017-05-29 10:05:22 -05:00
50390c6f4f
Revert "winport: use 666 color". Now I know how to do it... but it wasn't 100% ready anyway, and it breaks avi/screenshot, so let's wait and fix those when the 666 color is finally 100% done.
...
This reverts commit a0d42e1536
2017-05-28 20:11:45 -04:00
126db2d6de
fix null assignments to std::strings the right way
2017-05-05 14:26:34 -05:00
4707380746
Merge pull request #68 from klapeto/master
...
PosixGTK Graphics Dialog UI
2017-05-05 14:25:34 -05:00
a2527ef733
posixGTK: Additional graphic options added to Graphics Dialog
2017-05-05 21:42:51 +03:00
adf5151568
cheats: reset jit, sometimes, when it's needed
2017-05-05 00:45:28 -05:00
08e2244cea
posixGTK: Added Graphics Settings dialog and 3D core switch ui
2017-05-04 22:34:59 +03:00
b460d76698
Fixed segmentation faults caused by NULL assignments to std::strings
2017-05-04 16:16:12 +03:00
58b69b912a
winport: fix crash opening nonexistent recent rom ( fixes #67 )
2017-05-03 19:32:24 -05:00
6b92b6fb2a
Merge pull request #66 from np511/master
...
Add GUI to change JIT settings
2017-05-03 14:20:40 -05:00
ff2ebe400d
Add GUI to change JIT settings
2017-05-03 12:24:11 -04:00
e0945c362b
fix booting from firmware since I broke it adding the new slot-1 timing (failed to support it on arm7). fixes #64 (broken by 281268e)
2017-05-02 17:59:33 -05:00
09ffe3545e
try to fix bad, bad, badness caused by running nitrofs tool in another thread (hopefully fixes #65 )
2017-05-02 17:50:23 -05:00
9357fd79e3
fix bug in fsnitro FAT reading; fixes #63
2017-04-28 20:53:40 -05:00
bc2e1d99d3
fix crash in slot1 retail(debug) when it's set while no game is loaded ( fixes #63 )
2017-04-27 20:37:40 -05:00
281268e05f
emulate slot-1 read timings (fixes SF bug 1258)
2017-04-26 18:34:07 -05:00
6600a2c393
Merge pull request #61 from atsampson/master
...
fix: add missing comma in ZIPROMReader definition
2017-04-26 16:47:54 -05:00
ebde061afe
fix: add missing comma in ZIPROMReader definition
...
This was omitted in c365617ff737bdb3c06c307159078ca4e5f7e175; the other
ROMReader_structs it updated were correct.
2017-04-26 22:44:08 +01:00
669515bf6e
winport: fix OSD with prescale + single CPU
2017-04-26 15:02:46 -05:00
a0d42e1536
winport: use 666 color
2017-04-26 15:00:47 -05:00
06f0bb492b
winport: fix non-saving KillStylusOffScreen
2017-04-26 14:03:38 -05:00
751987cb90
fsnitro: fix build error from recent commit (for some compilers, anyway). fixes commit c365617ff7
2017-04-12 10:31:13 -05:00
fbad56bd14
fix bugs in lingering rom reading cursor caching code that made the winport memview capable of ruining rom reads
2017-04-11 21:14:16 -05:00
297449caa2
winport: fix mis-sorted full/cartrom memdomains
2017-04-11 21:13:17 -05:00
c365617ff7
take a stab at fixing the ram/streaming option so that there arent split codepaths everywhere. homebrew that is eligible for DLDI patching now switches entirely to load from memory, because we lack the facilities to patch the DLDI in flight (although we now have a point to hook it into). really, all this code should be torched, but it's less of a mess now.
2017-04-11 19:40:31 -05:00
e332e3df26
fix error in initialization of clear depth register - fixes #57 (cod3r colorz top screen)
2017-04-10 21:12:25 -05:00
abe2e61997
fix possibility for display capture to VRAM block not-mapped-to-LCDC doing something which I guess can clobber important memory. Mario kart does this after course demo during black screen transition if buttons are pressed (maybe programming error or emulation timing bug). The functionality that used to block this was lost during HD revisions. Fixes #54
2017-04-09 17:30:00 -05:00
e7d7c08455
Merge pull request #55 from BigBigos/rtc-override
...
Add rtc-day/hour options to override NDS clock
2017-04-09 12:51:44 -05:00
rogerman