registers AMD64 provides, this routine still needs to be written as self-
modifying code for maximum performance. The additional registers do allow
for further optimization over the x86 version by allowing all four pixels
to be in flight at the same time. The end result is that AMD64 ASM is about
2.18 times faster than AMD64 C and about 1.06 times faster than x86 ASM.
(For further comparison, AMD64 C and x86 C are practically the same for
this function.) Should I port any more assembly to AMD64, mvlineasm4 is the
most likely candidate, but it's not used enough at this point to bother.
Also, this may or may not work with Linux at the moment, since it doesn't
have the eh_handler metadata. Win64 is easier, since I just need to
structure the function prologue and epilogue properly and use some
assembler directives/macros to automatically generate the metadata. And
that brings up another point: You need YASM to assemble the AMD64 code,
because NASM doesn't support the Win64 metadata directives.
- Added an SSE version of DoBlending. This is strictly C intrinsics.
VC++ still throws around unneccessary register moves. GCC seems to be
pretty close to optimal, requiring only about 2 cycles/color. They're
both faster than my hand-written MMX routine, so I don't need to feel
bad about not hand-optimizing this for x64 builds.
- Removed an extra instruction from DoBlending_MMX, transposed two
instructions, and unrolled it once, shaving off about 80 cycles from the
time required to blend 256 palette entries. Why? Because I tried writing
a C version of the routine using compiler intrinsics and was appalled by
all the extra movq's VC++ added to the code. GCC was better, but still
generated extra instructions. I only wanted a C version because I can't
use inline assembly with VC++'s x64 compiler, and x64 assembly is a bit
of a pain. (It's a pain because Linux and Windows have different calling
conventions, and you need to maintain extra metadata for functions.) So,
the assembly version stays and the C version stays out.
- Removed all the pixel doubling r_detail modes, since the one platform they
were intended to assist (486) actually sees very little benefit from them.
- Rewrote CheckMMX in C and renamed it to CheckCPU.
- Fixed: CPUID function 0x80000005 is specified to return detailed L1 cache
only for AMD processors, so we must not use it on other architectures, or
we end up overwriting the L1 cache line size with 0 or some other number
we don't actually understand.
SVN r1134 (trunk)
surprised if this doesn't build in Linux right now. The CMakeLists.txt
were checked with MinGW and NMake, but how they fair under Linux is an
unknown to me at this time.
- Converted most sprintf (and all wsprintf) calls to either mysnprintf or
FStrings, depending on the situation.
- Changed the strings in the wbstartstruct to be FStrings.
- Changed myvsnprintf() to output nothing if count is greater than INT_MAX.
This is so that I can use a series of mysnprintf() calls and advance the
pointer for each one. Once the pointer goes beyond the end of the buffer,
the count will go negative, but since it's an unsigned type it will be
seen as excessively huge instead. This should not be a problem, as there's
no reason for ZDoom to be using text buffers larger than 2 GB anywhere.
- Ripped out the disabled bit from FGameConfigFile::MigrateOldConfig().
- Changed CalcMapName() to return an FString instead of a pointer to a static
buffer.
- Changed startmap in d_main.cpp into an FString.
- Changed CheckWarpTransMap() to take an FString& as the first argument.
- Changed d_mapname in g_level.cpp into an FString.
- Changed DoSubstitution() in ct_chat.cpp to place the substitutions in an
FString.
- Fixed: The MAPINFO parser wrote into the string buffer to construct a map
name when given a Hexen map number. This was fine with the old scanner
code, but only a happy coincidence prevents it from crashing with the new
code
- Added the 'B' conversion specifier to StringFormat::VWorker() for printing
binary numbers.
- Added CMake support for building with MinGW, MSYS, and NMake. Linux support
is probably broken until I get around to booting into Linux again. Niceties
provided over the existing Makefiles they're replacing:
* All command-line builds can use the same build system, rather than having
a separate one for MinGW and another for Linux.
* Microsoft's NMake tool is supported as a target.
* Progress meters.
* Parallel makes work from a fresh checkout without needing to be primed
first with a single-threaded make.
* Porting to other architectures should be simplified, whenever that day
comes.
- Replaced the makewad tool with zipdir. This handles the dependency tracking
itself instead of generating an external makefile to do it, since I couldn't
figure out how to generate a makefile with an external tool and include it
with a CMake-generated makefile. Where makewad used a master list of files
to generate the package file, zipdir just zips the entire contents of one or
more directories.
- Added the gdtoa package from netlib's fp library so that ZDoom's printf-style
formatting can be entirely independant of the CRT.
SVN r1082 (trunk)
- Added a check for PUFFGETSOWNER to A_BFGSpray.
- Moved the PUFFGETSOWNER check into P_SpawnPuff and removed the limitation
to players only.
- Fixed: P_SpawnMapThing still checked FMapThing::flags for the class bits
instead of FMapThing::ClassFilter.
- Fixed: A_CustomMissile must not let P_SpawnMissile call P_CheckMissileSpawn.
It must do this itself after setting the proper owner.
- Fixed: CCMD(give) increased the total item count.
- Fixed: A_Stop didn't set the player specific variables to 0.
SVN r1066 (trunk)
during it.
- UI sounds are now omitted from savegames.
- Fixed: Menu sounds had been restricted to one at a time again.
- Moved the P_SerializeSounds() call to the end of G_SerializeLevel() so that
it will occur after the players are loaded.
- Added fixes from FreeBSD for 0-length and very large string buffers
passed to myvsnprintf.
SVN r1063 (trunk)
run into the adjacent columns.
- Added a NullSoundRenderer so that most of the checks against a NULL GSnd
can be removed.
- Fixed: Looping sounds must always successfully allocate a channel, even if
it's only a pre-evicted channel.
SVN r1058 (trunk)
code to process their messages. This was necessary to handle the %zu format
option used in some memory allocation failure messages.
- Fixed: The flat texture scaling action specials were completely broken.
SVN r1056 (trunk)
the rest of the game is concerned, these sounds will never stop once they
have been started until they are explicitly stopped. If they are evicted
from their channels, the sound code will restart them as soon as possible.
This means that instead of this:
if (!S_IsActorPlayingSomething(actor, CHAN_WEAPON, -1))
{
S_Sound(actor, CHAN_WEAPON|CHAN_LOOP, soundid, 1, ATTN_NORM);
}
The following is now just as effective:
S_Sound(actor, CHAN_WEAPON|CHAN_LOOP, soundid, 1, ATTN_NORM);
There are also a couple of other ramifications presented by this change:
* The full state of the sound system (sans music) is now stored in save
games. Any sounds that were playing when you saved will still be
playing when you load. (Try saving while Korax is making a speech in
Hexen to hear it.)
* Using snd_reset will also preserve any playing sounds.
* Movie playback is disabled, probably forever. I did not want to
update the MovieDisable/ResumeSound stuff for the new eviction
tracking code. A properly updated movie player will use the VMR,
which doesn't need these functions, since it would pipe the sound
straight through the sound system like everything else, so I decided
to dump them now, which leaves the movie player in a totally unworkable
state.
June 26, 2008
- Changed S_Sound() to take the same floating point attenuation that the
internal S_StartSound() uses. Now ambient sounds can use the public
S_Sound() interface.
- Fixed: S_RelinkSound() compared the points of the channels against the
from actor's point, rather than checking the channels' mover.
- Changed Strife's animated doors so that their sounds originate from the
interior of the sector making them and not from the entire vertical height
of the map.
SVN r1055 (trunk)
won't be left on the stack at exit.
- Fixed: PSymbol needs a virtual destructor so that PSymbolActionFunction can
free its Arguments.
- Symbols for native classes are now freed on exit.
SVN r1022 (trunk)
capturing the mouse.
- Changed: Telefragging should not thrust the victim if it isn't in precisely the
same position as the killer.
- fixed: A_SpawnItemEx must call P_TeleportMove before checking the spawned
object's position.
SVN r1014 (trunk)
define any new ones of its own, I'm pre-emptively renaming the SNDEAX lump
to REVERBS to remove any possible misunderstanding that this is something
that requires EAX hardware support. (Ideally, it would have been REVERBDEF,
but that's 10 characters long.) The eaxedit console command has also been
renamed to reverbedit for the same reason.
- Fixed: The Palette part of FRemapTable was not initialized with alpha values
other than 0. I'm not sure if it would be better to fix this in the game
palette that it copies from or not, but right now, they get set
unconditionally to 255.
SVN r983 (trunk)
optional freeverb unit.
- Removed ResetEnvironment hack, since with software reverb, losing the
existing reverb when focus is lost isn't a problem.
- Commented out the TiMidity FIXME messages.
SVN r973 (trunk)
and not properly zero-terminated when all 8 characters were used.
- Fixed: Local sound sequence definitions caused a crash because a proper
NULL check was missing.
SVN r933 (trunk)
- Also changed all true color texture creation functions to use proper alpha
values instead of inverted ones.
- Changed FRemapTable so that all palette entries must contain proper alpha
values.
- Fixed: The F1 screen check in m_menu.cpp was missing a NULL pointer check.
- Changed: The boss brain's explosions play weapons/rocklx which is an
unlimited sound. This can become extremely loud. Replaced with a new
sound which is just an alias to weapons/rocklx but has a limit of 4.
SVN r932 (trunk)
- Fixed: MinGW doesn't have _get_pgmptr(), so it couldn't compile i_main.cpp.
- Fixed: MOD_WAVETABLE and MOD_SWSYNTH are not defined by w32api, so MinGW
failed compiling the new MIDI code.
- Fixed: LocalSndInfo and LocalSndSeq in S_Start() need to be const char
pointers, since "" is a constant.
- Fixed: parsecontext.h was missing a newline at the end of the file.
SVN r904 (trunk)
channel when restarting the song, rather than emitting a single note off
event which only has a 1 in 127 chance of being for a note that's playing
on that channel. Then I decided it would probably be a good idea to reset
all the controllers as well.
- Increasing the size of the internal Timidity stream buffer from 1/14 sec
(copied from the OPL player) improved its sound dramatically, so apparently
Timidity has issues with short stream buffers. It's now at 1/2 sec in
length. However, there seems to be something weird going on with
corazonazul_ff6boss.mid near the beginning where it stops and immediately
restarts a guitar on the exact same note.
- Added a new sound debugging cvar: snd_drawoutput, which can show various
oscilloscopes and spectrums.
- Internal TiMidity now plays music.
- Changed the progdir global variable into an FString.
SVN r900 (trunk)
set to NULL before performing the delete. Otherwise, in some abnormal
situations, it could happen that the destroyed screen object is still
being accessed.
- Fixed: V_Shutdown didn't set OF_YesReallyDelete before deleting screen.
SVN r870 (trunk)
that the drivers will treat a BackBufferCount of 0 as a request for
double buffering.
- Fixed: Unsetting a cvar did not remove it from the list of tab
completions.
- Added "" as a synonym for "nullimage" in SBARINFO.
- Fixed: MAKESAVESIG's stringifier in version.h did not work as expected.
It stringified the passed macro name, not the value of the macro.
- Moved DCajunMaster off the DObject hierarchy.
- Changed DCajunMaster::getspawned into a TArray of FStrings. It was
mysteriously being left pointing to uninitialized memory during the
final GC at exit and crashing.
- Fixed: The code that removed hexdd.wad from the list of IWADs when
hexen.wad was not present did not work.
SVN r861 (trunk)
- Removed movie volume from the sound menu and renamed some of the other
options to give the MIDI device name more room to display itself.
- Moved the midi device selection into the main sound menu.
- Added FMOD as MIDI device -1, to replace the MIDI mapper. This is still the
default device. By default, it uses exactly the same DLS instruments as the
Microsoft GS Wavetable Synth. If you have another set DLS level 1 patch set
you want to use, set the snd_midipatchfile cvar to specify where it should
load the instruments from.
- Changed the ProduceMIDI function to store its output into a TArray<BYTE>.
An overloaded version wraps around it to continue to supply file-writing
support for external Timidity++ usage.
- Added an FMOD credits banner to comply with their non-commercial license.
- Reimplemented the snd_buffersize cvar for the FMOD Ex sound system. Rather
than a time in ms, this is now the length in samples of the DSP buffer.
Also added the snd_buffercount cvar to offer complete control over the
call to FMOD::System::setDSPBufferSize(). Note that with any snd_samplerate
below about 44kHz, you will need to set snd_buffersize to avoid long
latencies.
- Reimplemented the snd_output cvar for the FMOD Ex sound system.
- Changed snd_samplerate default to 0. This now means to use the default
sample rate.
- Made snd_output, snd_output_format, snd_speakermode, snd_resampler, and
snd_hrtf available through the menu.
- Split the HRTF effect selection into its own cvar: snd_hrtf.
- Removed 96000 Hz option from the menu. It's still available through the
cvar, if desired.
- Fixed: If Windows sound init failed, retry with DirectSound. (Apparently,
WASAPI doesn't work with more than two speakers and PCM-Float output at the
same time.)
- Fixed: Area sounds only played from the front speakers once you got within
the 2D panning area.
SVN r854 (trunk)
of status bar face graphics belonging to skins.
SBARINFO update #15
- Fixed: Monospacing fonts wasn't quite correct.
- Fixed: The new mug shot code forgot to use the first arg of drawmugshot (the
one that picks the default sprite prefix).
- Added: lowerHealthCap variable to SBarInfo, which is set to true by default.
- Added: ininventory event to SBarInfo to detect if one or two items are in (or
not in) the player's inventory.
- Added: The ability to print global vars using drawnumber. I need someone to
test it though.
- Added: aspectratio command to detect what the user's aspect ratio is.
- Added: missing spacing argument to drawstring.
- Changed the sbarinfo display routine for drawnumber to not use cmd.value to
store what it is about to display. Now it uses a new variable.
- More conversions from DrawImage to screen->DrawTexture. I think only the
inventory bar drawing functions have to be changed now.
SVN r846 (trunk)
class so they all the low-level details of MIDI streaming are kept in
one place.
- Converted the SMF MIDI playback to use the same MIDI streams as MUS
playback.
- Moved MUS playback back into its own thread so that it can continue
uninterrupted if the main thread is too busy to service it in a timely
manner.
- Fixed: The MEVT_* values are not defined shifted into their spot for a
MIDIEVENT, so I need to do it myself.
- Fixed: Pausing a MUS and the changing snd_midivolume caused the paused
notes to become audible.
SVN r784 (trunk)
of ZDoom, except now the entire song isn't prebuffered in large chunks, so
I can insert MIDI events into the playback with fairly low latency. This
should offer more precise timing than the combination of low-level MIDI and
WaitForSingleObject timeouts.
SVN r783 (trunk)
- Fixed two problems with printfs from the net controller stuff: one of them
is totally valid, and the other is GCC not being as nice as VC++.
- Fixed: The case for defining SAVESIG when SVN_REVISION_NUMBER == 0 did not
work for GCC (and presumably VC++, though I never ran into that case with
it) because it tried to stringify something that wasn't a macro argument.
SVN r762 (trunk)
- Fixed: The MusicVolumes list was allocated with M_Malloc but freed with
delete.
- Fixed: demobuffer was inconsistantly handled with new[]/delete[] and
malloc/free.
- Added used memory tracking to M_Malloc() and M_Realloc(). This
necessitated the addition of an M_Free() call to track frees.
- Removed M_Calloc since it was only used in one place, and can just as well
be done with an M_Malloc/memset pair.
- Bumped DEMOGAMEVERSION for the new net controller codes.
SVN r751 (trunk)
- Disabled console alpha when the console is fullscreen.
- Re-enabled maybedrawnow for the software renderer. This should be replaced
with a different progress bar implementation at some point that can work
with hardware graphics.
SVN r747 (trunk)
frame to avoid the remote possibility that it could change during the middle
of drawing the frame and cause strange things to happen.
- Renamed test2d to vid_hw2d.
SVN r744 (trunk)
- Added .txt files to the list of types (wad, zip, and pk3) that can be
loaded without listing them after -file.
- Fonts that are created by the ACS setfont command to wrap a texture now
support animated textures.
- FON2 fonts can now use their full palette for CR_UNTRANSLATED when drawn
with the hardware 2D path instead of being restricted to the game palette.
- Fixed: Toggling vid_vsync would reset the displayed fullscreen gamma to 1
on a Radeon 9000.
- Added back the off-by-one palette handling, but in a much more limited
scope than before. The skipped entry is assumed to always be at 248, and
it is assumed that all Shader Model 1.4 cards suffer from this. That's
because all SM1.4 cards are based on variants of the ATI R200 core, and the
RV250 in a Radeon 9000 craps up like this. I see no reason to assume that
other flavors of the R200 are any different. (Interesting note: With the
Radeon 9000, D3DTADDRESS_CLAMP is an invalid address mode when using the
debug Direct3D 9 runtime, but it works perfectly fine with the retail
Direct3D 9 runtime.) (Insight: The R200 probably uses bytes for all its
math inside pixel shaders. That would explain perfectly why I can't use
constants greater than 1 with PS1.4 and why it can't do an exact mapping to
every entry in the color palette.
- Fixed: The software shaded drawer did not work for 2D, because its selected
"color"map was replaced with the identitymap before being used.
- Fixed: I cannot use Printf to output messages before the framebuffer was
completely setup, meaning that Shader Model 1.4 cards could not change
resolution.
- I have decided to let remap palettes specify variable alpha values for
their colors. D3DFB no longer forces them to 255.
- Updated re2c to version 0.12.3.
- Fixed: A_Wander used threshold as a timer, when it should have used
reactiontime.
- Fixed: A_CustomRailgun would not fire at all for actors without a target
when the aim parameter was disabled.
- Made the warp command work in multiplayer, again courtesy of Karate Chris.
- Fixed: Trying to spawn a bot while not in a game made for a crashing time.
(Patch courtesy of Karate Chris.)
- Removed some floating point math from hu_scores.cpp that somebody's GCC
gave warnings for (not mine, though).
- Fixed: The SBarInfo drawbar command crashed if the sprite image was
unavailable.
- Fixed: FString::operator=(const char *) did not release its old buffer when
being assigned to the null string.
- The scanner no longer has an upper limit on the length of strings it
accepts, though short strings will be faster than long ones.
- Moved all the text scanning functions into a class. Mainly, this means that
multiple script scanner states can be stored without being forced to do so
recursively. I think I might be taking advantage of that in the near
future. Possibly. Maybe.
- Removed some potential buffer overflows from the decal parser.
- Applied Blzut3's SBARINFO update #9:
* Fixed: When using even length values in drawnumber it would cap to a 98
value instead of a 99 as intended.
* The SBarInfo parser can now accept negatives for coordinates. This
doesn't allow much right now, but later I plan to add better fullscreen
hud support in which the negatives will be more useful. This also cleans
up the source a bit since all calls for (x, y) coordinates are with the
function getCoordinates().
- Added support for stencilling actors.
- Added support for non-black colors specified with DTA_ColorOverlay to the
software renderer.
- Fixed: The inverse, gold, red, and green fixed colormaps each allocated
space for 32 different colormaps, even though each only used the first one.
- Added two new blending flags to make reverse subtract blending more useful:
STYLEF_InvertSource and STYLEF_InvertOverlay. These invert the color that
gets blended with the background, since that seems like a good idea for
reverse subtraction. They also work with the other two blending operations.
- Added subtract and reverse subtract blending operations to the renderer.
Since the ERenderStyle enumeration was getting rather unwieldy, I converted
it into a new FRenderStyle structure that lets each parameter of the
blending equation be set separately. This simplified the set up for the
blend quite a bit, and it means a number of new combinations are available
by setting the parameters properly.
SVN r710 (trunk)
* Fixed: Drawbar had a few memory leaks.
- Added another temporary cvar called pc to play with the multipiler used by D3DFB::SetPaletteTexture().
SVN r704 (trunk)
truncate the coordinates to integers before sending them to the hardware.
Otherwise, there can be one pixel gaps compared to the software renderer,
because the hardware is rounding to nearest but the software renderer is
simply truncating the fractional part of the coordinate. This is the real
cause of the gap above the status bar at 1152x864 (and another gap to the
left of the status bar at 800x500).
- Fixed: When D3DFB::DrawTextureV had to clip a tile, it adjusted the
texture coordinates erroneously, still using the old calculations from
before texture packing was implemented.
SVN r695 (trunk)
of the frame buffer so I can get rid of the last remaining renderer check
outside the display code in GZDoom.
- made V_SetResolution a virtual function in IVideo. It's probably not a perfect
solution but at least it allows overriding it (which I need in GZDoom.)
Note: There's a lot of redundancy between hardware.cpp/h in the SDL and Win32
folders so some cleaning up might be a good idea.
SVN r692 (trunk)
can share the same hardware texture. This greatly reduces the number of
DrawPrimitive calls that need to be made when drawing text (or any 2D
graphics in general), so now hardware text is much faster than software text
all around. (As an example, one scenario went from 315 fps to over 1635 fps
for hardware, compared to 540 fps for software.)
SVN r687 (trunk)
- Modified M_DrawFrame() and R_DrawTopBorder() so that they call FlatFill() to
draw the edges of the frames. This at least seems a bit faster for hardware
2D.
- Implemented FlatFill() for D3DFB. It seems to be exactly as fast as the
default implementation that just calls DrawTexture() to tile the pieces onto
the screen, so I'm not sure it was worth the bother.
SVN r686 (trunk)
vertex buffer, made line batching automatic, and added an index buffer for
use when batching quads. The index buffer actually offered more of a
performance boost than simply batching the quads alone did.
SVN r685 (trunk)
players are on teams.
- Set TEAM_None back to 255. Since a player's team has already been accessible
through ACS, needlessly redefining this is a bad thing to do, since it can
break existing maps. 255 different teams should still be more than enough.
- Fixed: At certain resolutions, there was a one pixel row between the status
bar and the rest of the screen, thanks to rounding error.
- Added automatic batching of quads to D3DFB. Screens with a lot of text are
ever-so-slightly faster now, though still only about half the speed of
sofware-only text. I suppose the only way to see a marked improvement is
going to be by stuffing multiple glyphs in a single texture.
- Fixed: Crosshairgrow's animation was not framerate-independent.
SVN r668 (trunk)
and into the vertex data.
- Added functions for doing line drawing with Direct3D, including a new pair
of functions to do batched line drawing so that the map can actually be
drawn faster in hardware than in software (instead of an order of magnitude
slower).
SVN r663 (trunk)
