[Soft-Float] - Initial Interpreter Implementation of Ps2's floating point unit specification #12001

GitHubProUser67 · 2024-11-12T20:12:27Z

This Pull Request implements the first take ever on real Soft-Float support in PCSX2.

This work is a combination or several efforts and researches done prior.

Credits:

https://www.gregorygaines.com/blog/emulating-ps2-floating-point-nums-ieee-754-diffs-part-1/
https://github.com/assumenothing/Tommunism.SoftFloat/tree/main
https://github.com/GitHubProUser67/MultiServer3/blob/main/BackendServices/CastleLibrary/EmotionEngine.Emulator/Ps2Float.cs
https://github.com/Goatman13/pcsx2/tree/accurate_int_add_sub
PCSX2 Team for their help and support in this massive journey.

This pull request should be tested with every games requiring a clamping/rounding mode/float patches (cf: GameDatabase).

Currently, this PR fixes on the interpreters:

Any other games using the NegDiv hack or any other FPU rounding mode.

This sets the floor for Soft-Float in PCSX2, a long awaited contribution.

github-actions

Thank you for submitting a contribution to PCSX2

As this is your first pull request, please be aware of the contributing guidelines.

Additionally, as per recent changes in GitHub Actions, your pull request will need to be approved by a maintainer before GitHub Actions can run against it. You can find more information about this change here.

Please be patient until this happens. In the meantime if you'd like to confirm the builds are passing, you have the option of opening a PR on your own fork, just make sure your fork's master branch is up to date!

seta-san · 2024-11-12T20:46:56Z

Does this work on the recompilers or just interpreter?

refractionpcsx2 · 2024-11-12T20:48:09Z

You should try reading the title.

seta-san · 2024-11-12T20:51:02Z

You should try reading the title.

I don’t know why my brain just skimmed over that.

MrCK1 · 2024-11-12T23:33:42Z

I ran a bunch of tests for "Test Drive Unlimited" AI during the demo scene after sitting idle on the menu. No combination of settings/interpreters seems to have any effect on behavior. There must be something else going on

AmandaRoseChaqueta · 2024-11-13T04:55:20Z

I can confirm with the multiplication setting that MHG and Dos not longer have bugged bounces! (Plus the build works on linux well! [as well as running interpreter can do, of course :'D])

Shoegzer · 2024-11-13T05:01:11Z

Shouldn't this help with #2990 as well?

AmandaRoseChaqueta · 2024-11-13T05:12:18Z

Shouldn't this help with #2990 as well?

I remember seeing on the public dev channel that stuntman not longer had AI issues with this and not longer the car AI failed? it's been a while

Blackbird88 · 2024-11-13T08:37:52Z

Tourist Trophy NTSC works now too! For the first time License Tests in it work!
Requires: EE MUL/DIV

Some of the tests still don't work however.

Works

Hangs

Tokman5 · 2024-11-13T08:40:25Z

I tested the demo replays of Tokyo Xtreme Racer Zero (see issue #5597 ) and noticed that the car movement in interpreter mode is now closer to the movement in recompiler mode.
However, there are still slight differences between the two, and both are far from matching the console playback.

comp.mp4

Goatman13 · 2024-11-13T09:51:24Z

Nice work!

I tested only 2 games for now. Hype Time Quest which if i remember correctly needs accurate fpu mul, and it works fine without patch.

Second game is THPS4. While here accurate VU1 add/sub fix issue which is normally fixed by VU1 rounding in database, accurate mul/div breaks game graphics. Of course, game don't need accurate VU1 mul/div to work correctly, but something seems to be off when enabled.

Accurate vu1 add/sub/sqrt:

Accurate vu1 mul/div (same place in game):

Edit: Maybe THPS4 is just passing some big float value to GS with accurate mul/div. That may not necessarily be an issue with the float operation itself.
Additionally, I tested Burnout 2 and accurate VU1 add/sub fixed issue that was previously fixed by VU1 rounding mode (white car parts).

Edit2: Freaky Flyers is fixed with accurate mul/div on VU1. Awesome! We weren't even sure if that's floating points issue until now.

GitHubProUser67 · 2024-11-13T15:10:48Z

The Tony Hawk case is fixed, the game uses an un-documented behaviour in it's 3D engine.

The PS2 has no denormals support .... except in the Mul unit apparently.

The behaviour is now emulated properly.

Shoegzer · 2024-11-13T16:33:00Z

I remember seeing on the public dev channel that stuntman not longer had AI issues with this and not longer the car AI failed? it's been a while

@AmyRoxwell Can you provide a reference to this? There's no indication of it being fixed in #2990, and I assume the devs would have closed it if it were. In any event it involves pathing in Driver 3 as well.

weirdbeardgame · 2024-11-13T17:10:03Z

2024-11-10.14-28-32.mp4

This also affects the Fatal Frame 1 issue. Meaning this + the current GameDB patch will end up being the ultimate fix

Goatman13 · 2024-11-13T18:25:53Z

#3200 is fixed when tested with b7f3806
Require accurate mul/div for FPU.
I didn't tested, but Krome studio games should profit from this pr too, ones that have patches in game db. Like Spyro, Star Wars, one of Transformers game.

ghost · 2024-11-13T19:15:33Z

This pr's EE interpreter fixes #11636 's gamedb issue.

AmandaRoseChaqueta · 2024-11-13T20:39:32Z

I remember seeing on the public dev channel that stuntman not longer had AI issues with this and not longer the car AI failed? it's been a while

@AmyRoxwell Can you provide a reference to this? There's no indication of it being fixed in #2990, and I assume the devs would have closed it if it were. In any event it involves pathing in Driver 3 as well.

I meant like, while using this PR, not that is has been fixed. Sorry if it was misunderstood. But if it's not mention on the PR maybe the thing it needs it's not here by this initial implementation.

GitHubProUser67 · 2024-11-13T20:42:35Z

Driv3r seemed fine when it was tested, Stuntman NTSC is a lot better but still can "slightly" deviate.

I suspect it is once again, the interpreter rounding/clamping values somewhere.

Shoegzer · 2024-11-13T21:34:36Z

@AmyRoxwell Ah, I understand you now. That's great news.

@GitHubProUser67 Thanks, nice to see Driv3r is looking better. Would it make sense to list these games in your OP?

SomethingUnique12 · 2024-11-14T07:30:29Z

Nice to hear of stuntman finally faring better nowadays.

Goatman13 · 2024-11-14T18:24:17Z

Game Constantine, tested US version. On level 2 there is dumpster that we need to climb to progress in game. Recompiler get it fine with positive rounding for EE. This is currently not working with soft floats, no matter what.

Reproduction steps.

Launch first level of game.
Use this cheat to skip to level 2: Press select, go to the Journal Screen, and enter RIGHT, LEFT, RIGHT, LEFT, RIGHT, LEFT, SQUARE, CIRCLE, CIRCLE, SQUARE, SQUARE, SQUARE, SQUARE, CIRCLE, CIRCLE.
Jump thru fence, kill rats, there will be now dumpster on the right side, one that do not work.

GitHubProUser67 · 2024-11-14T21:18:59Z

I managed to fix it, but the game requires accurate soft floats on the VU0 and the EE FPU, so far I tried with booth Add/Sub and Mul/Div on EE FPU/COP2 and VU0:

Edit: Positive rounding hack-fix it because it makes the float VU0 friendly. In fact the game does some weird EE FPU->VU0 communications with high floats.

GitHubProUser67 · 2024-11-14T21:31:09Z

Note to future tests:

The constantine case is a typical case of a hack fixing a problem that is related to what I call "float broadcasting".

A game might want to transfer a out of IEEE range float from the FPU to the VUs (and vice-versa).

When a game Database entry has a rounding mode, usually it covers a can of worms where the game requires accurate soft floats on more than one processor.

LoStraniero91 · 2025-03-22T01:24:18Z

Pretty much every replay for any racing game (including Stuntman and Driv3r) will require tight FPU accuracy.
So far I've tested different games with a replay saved on PS2, then playing it side-by-side with PCSX2 and real HW:
Driv3r (All floats enabled for VU0)
Gran Turismo 4 (All floats enabled for EE, or it will even fail to load the replay)
WRC Rally Evolved (All floats enabled for EE and VU0)

Edit: 23/03/2025
Toca Race Driver 3 (All floats enabled for EE and VU0)

voorunde · 2025-03-25T14:28:15Z

Fixes flickering on opponents' cars in Namco's R: Racing Evolution in bumper cam view. EE addition/subtraction and multiplication/division have to be enabled.

seta-san · 2025-03-28T13:37:53Z

this PR probably renders this obsolete and closed too. #1110

Dynamo128 · 2025-04-16T22:31:20Z

Out of curiosity, to any testers: what is the current state of Stuntman with this PR? Has anyone done thorough testing on both the NTSC and PAL versions? If so, can you point out which missions break?

IzhenPS2 · 2025-04-18T00:18:44Z

what should i enable to fix Need for Speed Undercover #9831 bug?
i tried enabling all floats for EE and VU0 and still hasn't found a solution, am i missing something?

GitHubProUser67 · 2025-04-20T09:46:15Z

This pull request will be closed in favor of a newer one since the rebasing process cannot be done.

Mrlinkwii · 2025-04-20T10:14:34Z

just so we can link it the new pr is #12550

F0bes · 2025-04-20T11:44:47Z

git reset --hard origin/master

Creating a new PR causes information to now be split.

TheLastRar · 2025-04-26T16:21:07Z

I forgot to mention at the time, but this PR was updated with the changes from #12550 and rebased.

Meaning this PR is the updated one.

…terpreters.nit specification. This work is a combination or several efforts and researches done prior. Credits: - https://www.gregorygaines.com/blog/emulating-ps2-floating-point-nums-ieee-754-diffs-part-1/ - https://github.com/GitHubProUser67/MultiServer3/tree/main/BackendServices/PS2FloatLibrary - https://github.com/Goatman13/pcsx2/tree/accurate_int_add_sub - PCSX2 Team for their help and support in this massive journey. Fixes codacity warnings.

GovanifY · 2025-05-17T14:53:28Z

Hey, I rebased your PR/removed your merge commit to make it apply cleanly against master :)

GitHubProUser67 · 2025-05-17T16:01:51Z

What would I do without you? :)

TellowKrinkle

In Discord, you claimed this was ready for review, but you haven't addressed any of the following things I pointed out in the last review or written anything detailing why you didn't address them.

BitScanReverse8 is still there
I mentioned switching option checks to a bitfield last time, and you haven't done that or given a reason why you don't think that's a good idea.
I asked you to keep PS2Float simple, and leave out things that the emulator already implements accurately. Last time, I requested that you keep it to just add, sub, mul, madd, msub, div, and sqrt, but I'll add the transcendentals into that. But we don't need things like ftoi/itof or comparisons. We also don't need a ToString.
I questioned having all the functions take a PS2Float, which now includes a bunch of output-only flags. Either make PS2Float not contain output flags, or implement the functions as static methods that take u32s, with madd/msub taking an extra boolean indicating incoming overflow. This not only improves performance, but helps make clear what is and isn't needed to people looking at the method signatures. e.g. someone could easily see that the output of add depends only on the two incoming numbers, but madd actually cares about whether the previous number overflowed or not.

In addition, I posted this test elf on Discord before the rewrite, you replied that your C# code correctly handles the broken add result, yet it's still broken after you rewrote the C++ code (along with a lot of the flag tests):

Test Results

[    1.6582] ADD.S 00000000 + 807FFFFF => 80000000 != 00000000
[    1.6588] VADD 00000000 + 807FFFFF => 80000000 != 00000000
[    1.6591] VADD 00000000 + 807FFFFF MAC FLAGS ZS-- != Z---
[    1.6594] VADD 00800001 + 80800000 MAC FLAGS Z--- != Z-U-
[    1.6597] VADD {3F800000 7FFFFFFF 00800000 00800001} + {BF800000 FFFFFFFF 80800000 80800000} STATUS FLAGS ZS----------Z----- != ZS--US------Z-U---
[    1.6599] VADD 80800001 + 00800000 MAC FLAGS ZS-- != ZSU-
[    1.6603] VADD {BF800000 FFFFFFFF 80800000 80800001} + {3F800000 7FFFFFFF 00800000 00800000} STATUS FLAGS ZSSS--------ZS---- != ZSSSUS------ZSU---
[    1.6605] VADD 00900001 + 80900000 MAC FLAGS Z--- != Z-U-
[    1.6607] VADD 00900000 + 808FFFFF MAC FLAGS Z--- != Z-U-
[    1.6609] VADD 80900001 + 00900000 MAC FLAGS ZS-- != ZSU-
[    1.6611] VADD 80900000 + 008FFFFF MAC FLAGS ZS-- != ZSU-
[    1.6615] VADD {00900001 00900000 80900001 80900000} + {80900000 808FFFFF 00900000 008FFFFF} STATUS FLAGS ZSSS--------ZS---- != ZSSSUS------ZSU---
[    1.6617] SUB.S 00000000 - 007FFFFF => 80000000 != 00000000
[    1.6620] VSUB 00000000 - 007FFFFF => 80000000 != 00000000
[    1.6623] VSUB 00000000 - 007FFFFF MAC FLAGS ZS-- != Z---
[    1.6625] VSUB 00800001 - 00800000 MAC FLAGS Z--- != Z-U-
[    1.6628] VSUB {3F800000 7FFFFFFF 00800000 00800001} - {3F800000 7FFFFFFF 00800000 00800000} STATUS FLAGS ZS----------Z----- != ZS--US------Z-U---
[    1.6630] VSUB 80800001 - 80800000 MAC FLAGS ZS-- != ZSU-
[    1.6633] VSUB {BF800000 FFFFFFFF 80800000 80800001} - {BF800000 FFFFFFFF 80800000 80800000} STATUS FLAGS ZSSS--------ZS---- != ZSSSUS------ZSU---
[    1.6635] VSUB 00900001 - 00900000 MAC FLAGS Z--- != Z-U-
[    1.6638] VSUB 00900000 - 008FFFFF MAC FLAGS Z--- != Z-U-
[    1.6640] VSUB 80900001 - 80900000 MAC FLAGS ZS-- != ZSU-
[    1.6642] VSUB 80900000 - 808FFFFF MAC FLAGS ZS-- != ZSU-
[    1.6645] VSUB {00900001 00900000 80900001 80900000} - {00900000 008FFFFF 80900000 808FFFFF} STATUS FLAGS ZSSS--------ZS---- != ZSSSUS------ZSU---
[    1.6651] VMUL 00800000 * 3F000000 MAC FLAGS Z--- != Z-U-
[    1.6654] VMUL 80800000 * 3F000000 MAC FLAGS ZS-- != ZSU-
[    1.6657] VMUL 00800000 * BF000000 MAC FLAGS ZS-- != ZSU-
[    1.6659] VMUL 80800000 * BF000000 MAC FLAGS Z--- != Z-U-
[    1.6663] VMUL {00800000 80800000 00800000 80800000} * {3F000000 3F000000 BF000000 BF000000} STATUS FLAGS ZSSS--------ZS---- != ZSSSUS------ZSU---
[    1.6665] VMUL 20000000 * 1F800000 MAC FLAGS Z--- != Z-U-
[    1.6667] VMUL A0000000 * 1F800000 MAC FLAGS ZS-- != ZSU-
[    1.6669] VMUL 20000000 * 9F800000 MAC FLAGS ZS-- != ZSU-
[    1.6672] VMUL {20000000 A0000000 20000000 A0400000} * {1F800000 1F800000 9F800000 9FC00000} STATUS FLAGS ZSSS--------ZS---- != ZSSSUS------ZSU---
[    1.6674] VMUL 3F080000 * 00C80000 MAC FLAGS Z--- != Z-U-
[    1.6677] VMUL {408F0000 40CF0000 3F480000 3F080000} * {7E8FFFFF 7ECFFFFF 00C80000 00C80000} STATUS FLAGS ZS----OS----Z--O-- != ZS--USOS----Z-UO--
[    1.6680] MADD.S 007FFFFF + 80800000 * 80800000 FLAGS ------------ != SU----------
[    1.6682] MADD.S BF800000 + 80800000 * 80800000 FLAGS ------------ != SU----------
[    1.6686] VMADD {00000000 00000000 00000000 00000000} + {00000000 00000000 80000000 80000000} * {00000000 80000000 00000000 80000000} STATUS FLAGS ZS----------Z----- != ZSSS--------Z-----
[    1.6689] VMADD {00800000 80000000 80000000 007FFFFF} + {80000000 80800000 80000000 80800000} * {80000000 80000000 80800000 80800000} STATUS FLAGS ZS----------Z----- != ZS--US------Z-----
[    1.6693] VMADD {00800000 3F800000 BF800000 BF800000} + {80000000 80800000 80000000 80800000} * {80000000 80000000 80800000 80800000} STATUS FLAGS --SS---------S---- != ZSSSUS-------S----
[    1.6695] VMADD 00900001 + A0400000 * 1FC00000 MAC FLAGS Z--- != Z-U-
[    1.6697] VMADD 808FFFFF + A0400000 * 9FC00000 MAC FLAGS Z--- != Z-U-
[    1.6700] VMADD 80900000 + A0400000 * 9FC00002 MAC FLAGS Z--- != Z-U-
[    1.6703] VMADD {00900000 00900001 808FFFFF 80900000} + {A0400000 A0400000 A0400000 A0400000} * {1FC00000 1FC00000 9FC00000 9FC00002} STATUS FLAGS ZS----------Z----- != ZSSSUS------Z-U---
[    1.6705] VMADD 80900001 + A0400000 * 9FC00000 MAC FLAGS ZS-- != ZSU-
[    1.6707] VMADD 008FFFFF + A0400000 * 1FC00000 MAC FLAGS ZS-- != ZSU-
[    1.6710] VMADD 00900000 + A0400000 * 1FC00002 MAC FLAGS ZS-- != ZSU-
[    1.6713] VMADD {80900000 80900001 008FFFFF 00900000} + {A0400000 A0400000 A0400000 A0400000} * {9FC00000 9FC00000 1FC00000 1FC00002} STATUS FLAGS ZSSS--------ZS---- != ZSSSUS------ZSU---
[    1.6715] MSUB.S 007FFFFF - 80800000 * 00800000 FLAGS ------------ != SU----------
[    1.6717] MSUB.S BF800000 - 80800000 * 00800000 FLAGS ------------ != SU----------
[    1.6725] VMSUB {00000000 00000000 00000000 00000000} - {00000000 00000000 80000000 80000000} * {80000000 00000000 80000000 00000000} STATUS FLAGS ZS----------Z----- != ZSSS--------Z-----
[    1.6729] VMSUB {807FFFFF 80000000 80000000 007FFFFF} - {80000000 807FFFFF 80000000 807FFFFF} * {00000000 00000000 007FFFFF 007FFFFF} STATUS FLAGS ZS----------Z----- != ZSSS--------Z-----
[    1.6732] VMSUB {00800000 80000000 80000000 007FFFFF} - {80000000 80800000 80000000 80800000} * {00000000 00000000 00800000 00800000} STATUS FLAGS ZS----------Z----- != ZSSSUS------Z-----
[    1.6736] VMSUB {00800000 3F800000 BF800000 BF800000} - {80000000 80800000 80000000 80800000} * {00000000 00000000 00800000 00800000} STATUS FLAGS --SS---------S---- != ZSSSUS-------S----
[    1.6738] VMSUB 00900001 - A0400000 * 9FC00000 MAC FLAGS Z--- != Z-U-
[    1.6740] VMSUB 808FFFFF - A0400000 * 1FC00000 MAC FLAGS Z--- != Z-U-
[    1.6742] VMSUB 80900000 - A0400000 * 1FC00002 MAC FLAGS Z--- != Z-U-
[    1.6745] VMSUB {00900000 00900001 808FFFFF 80900000} - {A0400000 A0400000 A0400000 A0400000} * {9FC00000 9FC00000 1FC00000 1FC00002} STATUS FLAGS ZS----------Z----- != ZSSSUS------Z-U---
[    1.6747] VMSUB 80900001 - A0400000 * 1FC00000 MAC FLAGS ZS-- != ZSU-
[    1.6749] VMSUB 008FFFFF - A0400000 * 9FC00000 MAC FLAGS ZS-- != ZSU-
[    1.6751] VMSUB 00900000 - A0400000 * 9FC00002 MAC FLAGS ZS-- != ZSU-
[    1.6754] VMSUB {80900000 80900001 008FFFFF 00900000} - {A0400000 A0400000 A0400000 A0400000} * {1FC00000 1FC00000 9FC00000 9FC00002} STATUS FLAGS ZSSS--------ZS---- != ZSSSUS------ZSU---
[    1.6757] VMSUB {FFFFFFFF 00000000 FFFFFFFF FFFFFFFF} - {FFFFFFFF 7F800000 7FFFFFFF 7FFFFFFF} * {7FFFFFFF FF800000 BF800001 BF800000} STATUS FLAGS ZS----OS----Z--O-- != ZSSS--OS----Z--O--
[    1.6759] VMSUB {3F800000 00000000 73800000 74000000} - {7FFFFFFF 7FFFFFFF 7FFFFFFF 7FFFFFFF} * {BF800000 BF800000 BF800000 BF800000} STATUS FLAGS ------OS-------O-- != --SS--OS-------O--

TellowKrinkle · 2025-05-23T07:19:32Z

pcsx2/FPU.cpp

+	if (CHECK_FPU_SOFT_ADDSUB || CHECK_FPU_SOFT_MULDIV || CHECK_FPU_SOFT_SQRT) { _FdValUl_ = PS2Float::Itof(0, _FsValSl_).raw; }
+	else
+	{
+		_FdValf_ = (float)_FsValSl_;
+		_FdValf_ = fpuDouble(_FdValUl_);
+	}


Suggested change

if (CHECK_FPU_SOFT_ADDSUB || CHECK_FPU_SOFT_MULDIV || CHECK_FPU_SOFT_SQRT) { _FdValUl_ = PS2Float::Itof(0, _FsValSl_).raw; }

else

{

_FdValf_ = (float)_FsValSl_;

_FdValf_ = fpuDouble(_FdValUl_);

}

_FdValf_ = (float)_FsValSl_;

TellowKrinkle · 2025-05-23T07:20:15Z

pcsx2/FPU.cpp

 }

 void CVT_W() {
-	if ( ( _FsValUl_ & 0x7F800000 ) <= 0x4E800000 ) { _FdValSl_ = (s32)_FsValf_; }
+	if (CHECK_FPU_SOFT_ADDSUB || CHECK_FPU_SOFT_MULDIV || CHECK_FPU_SOFT_SQRT) { _FdValSl_ = PS2Float::Ftoi(0, _FsValUl_); }


It's already accurate

TellowKrinkle · 2025-05-23T07:22:48Z

pcsx2/FPU.cpp

+	{
+		PS2Float divres = fpuAccurateDiv(_FsValUl_, _FtValUl_);
+		_FdValUl_ = divres.raw;
+		if (checkDivideByZeroInvalidSoft(divres, FPUflagD | FPUflagSD, FPUflagI | FPUflagSI)) return;


Why reimplement the div0 check when the existing one is already accurate?

TellowKrinkle · 2025-05-23T07:30:19Z

pcsx2/PS2Float.cpp

+PS2Float::PS2Float(s32 value)
+	: raw((u32)value)
+{}
+
+PS2Float::PS2Float(u32 value)
+	: raw(value)
+{}
+
+PS2Float::PS2Float(float value)
+	: raw(std::bit_cast<u32>(value))
+{}
+
+PS2Float::PS2Float(bool sign, u8 exponent, u32 mantissa)
+	: raw((sign ? 1u : 0u) << 31 |
+		(u32)(exponent << MANTISSA_BITS) |
+		(mantissa & 0x7FFFFF))
+{}
+
+PS2Float PS2Float::Max()
+{
+	return PS2Float(MAX_FLOATING_POINT_VALUE);
+}
+
+PS2Float PS2Float::Min()
+{
+	return PS2Float(MIN_FLOATING_POINT_VALUE);
+}
+
+PS2Float PS2Float::One()
+{
+	return PS2Float(ONE);
+}
+
+PS2Float PS2Float::MinOne()
+{
+	return PS2Float(MIN_ONE);
+}


Simple methods like these can go in the header so they're inlinable even if you don't have LTO enabled. Maybe even add a __fi to them.

TellowKrinkle · 2025-05-23T07:33:03Z

pcsx2/PS2Float.cpp

+PS2Float PS2Float::Sub(PS2Float subtrahend)
+{
+	if (IsDenormalized() || subtrahend.IsDenormalized())
+	{
+		bool sign = DetermineSubtractionOperationSign(*this, subtrahend);
+
+		if (IsDenormalized() && !subtrahend.IsDenormalized())
+			return PS2Float(sign, subtrahend.Exponent(), subtrahend.Mantissa());
+		else if (!IsDenormalized() && subtrahend.IsDenormalized())
+			return PS2Float(sign, Exponent(), Mantissa());
+		else if (IsDenormalized() && subtrahend.IsDenormalized())
+			return PS2Float(sign, 0, 0);
+		else
+			Console.Error("Both numbers are not denormalized");
+
+		return PS2Float(0);
+	}
+
+	u32 a = raw;
+	u32 b = subtrahend.raw;
+
+	//exponent difference
+	s32 exp_diff = Exponent() - subtrahend.Exponent();
+
+	//diff = 1 .. 24, expt < expd
+	if (exp_diff > 0 && exp_diff < 25)
+	{
+		exp_diff = exp_diff - 1;
+		b = (MIN_FLOATING_POINT_VALUE << exp_diff) & b;
+	}
+
+	//diff = -24 .. -1 , expd < expt
+	else if (exp_diff < 0 && exp_diff > -25)
+	{
+		exp_diff = -exp_diff;
+		exp_diff = exp_diff - 1;
+		a = a & (MIN_FLOATING_POINT_VALUE << exp_diff);
+	}
+
+	return PS2Float(a).DoAdd(PS2Float(b).Negate());
+}


Move this to the header and implement it as Add(b ^ 0x80000000)

GoodMornang · 2025-05-31T02:12:20Z

how do you download this and put into pcsx2?

seta-san · 2025-05-31T14:45:29Z

how do you download this and put into pcsx2?

This isn’t complete nor is it tested. If you were inclined to want to test it you need to go to the “checks” part of this pr. Choose your OS and then download from the artifacts.

GoodMornang · 2025-05-31T15:04:10Z

how do you download this and put into pcsx2?

This isn’t complete nor is it tested. If you were inclined to want to test it you need to go to the “checks” part of this pr. Choose your OS and then download from the artifacts.

Ok. I'm just not very familiar with GitHub but thanks anyway.

GitHubProUser67 changed the title ~~[Soft-Float] - Initial Intepreter Implementation of Ps2's floating point uint specification~~ [Soft-Float] - Initial Intepreter Implementation of Ps2's floating point unit specification Nov 12, 2024

github-actions bot added Build | Project System Vector Units GUI/Qt First Time Contribution labels Nov 12, 2024

github-actions bot reviewed Nov 12, 2024

View reviewed changes

GitHubProUser67 changed the title ~~[Soft-Float] - Initial Intepreter Implementation of Ps2's floating point unit specification~~ [Soft-Float] - Initial Interpreter Implementation of Ps2's floating point unit specification Nov 12, 2024

GitHubProUser67 force-pushed the soft-float-int branch from 96414bd to de047ea Compare November 12, 2024 20:48

JordanTheToaster mentioned this pull request Mar 18, 2025

Armored Core 3: Z-Fighting with every renderer #2398

Open

JordanTheToaster mentioned this pull request Mar 23, 2025

[BUG]: Invisible Pause Menu in Transformers Revenge Of The Fallen on Software and Hardware Mode #5549

Open

Mrlinkwii mentioned this pull request Apr 10, 2025

Manhunt (NTSC-U) [SLUS-20827] - Crane freeze only on right side of the crane #3031

Open

GitHubProUser67 closed this Apr 20, 2025

GitHubProUser67 deleted the soft-float-int branch April 20, 2025 09:46

GitHubProUser67 restored the soft-float-int branch April 20, 2025 17:05

JordanTheToaster reopened this Apr 20, 2025

TheLastRar force-pushed the soft-float-int branch 2 times, most recently from 2536b24 to e8e8d94 Compare April 20, 2025 19:17

Mrlinkwii removed the Needs Rebase label Apr 20, 2025

DfactorLongplays mentioned this pull request May 10, 2025

[BUG]: Ace Combat 04 - Inaccurate trajectory calculation #12641

Open

GitHubProUser67 force-pushed the soft-float-int branch from 46f4624 to 5c5b448 Compare May 17, 2025 13:48

JordanTheToaster added this to the Release 2.X milestone May 17, 2025

GovanifY force-pushed the soft-float-int branch from 17fe955 to 52a9243 Compare May 17, 2025 14:52

TellowKrinkle reviewed May 25, 2025

View reviewed changes

Uh oh!

[Soft-Float] - Initial Interpreter Implementation of Ps2's floating point unit specification #12001

Are you sure you want to change the base?

[Soft-Float] - Initial Interpreter Implementation of Ps2's floating point unit specification #12001

Conversation

GitHubProUser67 commented Nov 12, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions bot left a comment

Choose a reason for hiding this comment

Thank you for submitting a contribution to PCSX2

Uh oh!

seta-san commented Nov 12, 2024

Uh oh!

refractionpcsx2 commented Nov 12, 2024

Uh oh!

seta-san commented Nov 12, 2024

Uh oh!

MrCK1 commented Nov 12, 2024

Uh oh!

AmandaRoseChaqueta commented Nov 13, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Shoegzer commented Nov 13, 2024

Uh oh!

AmandaRoseChaqueta commented Nov 13, 2024

Uh oh!

Blackbird88 commented Nov 13, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Tokman5 commented Nov 13, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Goatman13 commented Nov 13, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

GitHubProUser67 commented Nov 13, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Shoegzer commented Nov 13, 2024

Uh oh!

weirdbeardgame commented Nov 13, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Goatman13 commented Nov 13, 2024

Uh oh!

ghost commented Nov 13, 2024

Uh oh!

AmandaRoseChaqueta commented Nov 13, 2024

Uh oh!

GitHubProUser67 commented Nov 13, 2024

Uh oh!

Shoegzer commented Nov 13, 2024

Uh oh!

SomethingUnique12 commented Nov 14, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Goatman13 commented Nov 14, 2024

Uh oh!

GitHubProUser67 commented Nov 14, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

GitHubProUser67 commented Nov 14, 2024

Uh oh!

LoStraniero91 commented Mar 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

voorunde commented Mar 25, 2025

Uh oh!

seta-san commented Mar 28, 2025

Uh oh!

Dynamo128 commented Apr 16, 2025

Uh oh!

IzhenPS2 commented Apr 18, 2025

Uh oh!

GitHubProUser67 commented Apr 20, 2025

Uh oh!

Mrlinkwii commented Apr 20, 2025

Uh oh!

F0bes commented Apr 20, 2025

GitHubProUser67 commented Nov 12, 2024 •

edited

Loading

AmandaRoseChaqueta commented Nov 13, 2024 •

edited

Loading

Blackbird88 commented Nov 13, 2024 •

edited

Loading

Tokman5 commented Nov 13, 2024 •

edited

Loading

Goatman13 commented Nov 13, 2024 •

edited

Loading

GitHubProUser67 commented Nov 13, 2024 •

edited

Loading

weirdbeardgame commented Nov 13, 2024 •

edited

Loading

SomethingUnique12 commented Nov 14, 2024 •

edited

Loading

GitHubProUser67 commented Nov 14, 2024 •

edited

Loading

LoStraniero91 commented Mar 22, 2025 •

edited

Loading