Mario Golf: Toadstool Tour crashes randomly

[Haliinen]

Okay, sorry if this has been posted before, I did a search on this in the forums but didn't find a thread similiar to this I'm going to post, but hey, I just want to report an issue I experienced with Dolphin when playing this game.

I'll try to keep this short, this happens randomly, game runs smooth as silk with minor graphical glitches in-game, major graphical glitches in the menus but you can still navigate through it fine. I know that this has nothing to do with my PC at all, I'm running the 310.33 driver by Nvidia which is a Beta driver, but has not yield any issues at all for me. I tried both the DX11- and OpenGL backend with the same issue popping up so don't know what's going on there. I'll post the dumps below.

Crash message

Spoiler: (Show Spoiler)

---------------------------
Warning
---------------------------
Failed to compile pixel shader!
This usually happens when trying to use Dolphin with an outdated GPU or integrated GPU like the Intel GMA series.

If you're sure this is Dolphin's error anyway, post the contents of ./User/Dump/bad_ps_0000.txt along with this error message at the forums.

Debug info (ps_5_0):
(41,27): error X3004: undeclared identifier 'Normal'

---------------------------
OK ---------------------------

Text file dump

Spoiler: (Show Spoiler)

//Pixel Shader for TEV stages
//1 TEV stages, 8 texgens, XXX IND stages
sampler samp0 : register(s0), samp1 : register(s1), samp2 : register(s2), samp3 : register(s3), samp4 : register(s4), samp5 : register(s5), samp6 : register(s6), samp7 : register(s7);
Texture2D Tex0 : register(t0), Tex1 : register(t1), Tex2 : register(t2), Tex3 : register(t3), Tex4 : register(t4), Tex5 : register(t5), Tex6 : register(t6), Tex7 : register(t7);

uniform float4 color[4] : register(c0);
uniform float4 k[4] : register(c4);
uniform float4 alphaRef[1] : register(c8);
uniform float4 texdim[8] : register(c9);
uniform float4 czbias[2] : register(c17);
uniform float4 cindscale[2] : register(c19);
uniform float4 cindmtx[6] : register(c21);
uniform float4 cfog[3] : register(c27);
typedef struct { float4 col; float4 cosatt; float4 distatt; float4 pos; float4 dir; } Light;
typedef struct { Light lights[8]; } s_cLights;
uniform s_cLights cLights : register(c30);
typedef struct { float4 C0, C1, C2, C3; } s_cmtrl;
uniform s_cmtrl cmtrl : register(c70);
void main(
out float4 ocol0 : SV_Target0,
in float4 rawpos : SV_Position,
in float4 colors_0 : COLOR0,
in float4 colors_1 : COLOR1,
in float4 uv0 : TEXCOORD0,
in float4 uv1 : TEXCOORD1,
in float4 uv2 : TEXCOORD2,
in float4 uv3 : TEXCOORD3,
in float4 uv4 : TEXCOORD4,
in float4 uv5 : TEXCOORD5,
in float4 uv6 : TEXCOORD6,
in float4 uv7 : TEXCOORD7 ) {
float4 c0 = color[1], c1 = color[2], c2 = color[3], prev = float4(0.0f, 0.0f, 0.0f, 0.0f), textemp = float4(0.0f, 0.0f, 0.0f, 0.0f), rastemp = float4(0.0f, 0.0f, 0.0f, 0.0f), konsttemp = float4(0.0f, 0.0f, 0.0f, 0.0f);
float3 comp16 = float3(1.0f, 255.0f, 0.0f), comp24 = float3(1.0f, 255.0f, 255.0f*255.0f);
float4 alphabump=float4(0.0f,0.0f,0.0f,0.0f);
float3 tevcoord=float3(0.0f, 0.0f, 0.0f);
float2 wrappedcoord=float2(0.0f,0.0f), tempcoord=float2(0.0f,0.0f);
float4 cc0=float4(0.0f,0.0f,0.0f,0.0f), cc1=float4(0.0f,0.0f,0.0f,0.0f);
float4 cc2=float4(0.0f,0.0f,0.0f,0.0f), cprev=float4(0.0f,0.0f,0.0f,0.0f);
float4 crastemp=float4(0.0f,0.0f,0.0f,0.0f),ckonsttemp=float4(0.0f,0.0f,0.0f,0.0f);

float3 _norm0 = normalize(Normal.xyz);

float3 pos = float3(clipPos.x,clipPos.y,Normal.w);
float4 mat, lacc;
float3 ldir, h;
float dist, dist2, attn;
{
mat = cmtrl.C2;
lacc = cmtrl.C0;
lacc.w = 1.0f;
colors_0 = mat * saturate(lacc);
}
float4 clipPos = float4(rawpos.x, rawpos.y, uv2.w, uv3.w);
uv0.xy = uv0.xy * texdim[0].zw;
uv1.xy = uv1.xy * texdim[1].zw;
uv2.xy = uv2.xy * texdim[2].zw;
uv3.xy = uv3.xy * texdim[3].zw;
uv4.xy = uv4.xy * texdim[4].zw;
uv5.xy = uv5.xy * texdim[5].zw;
uv6.xy = uv6.xy * texdim[6].zw;
uv7.xy = uv7.xy * texdim[7].zw;
// TEV stage 0
rastemp = colors_0.rgba;
crastemp = frac(rastemp * (255.0f/256.0f)) * (256.0f/255.0f);
textemp = float4(1.0f, 1.0f, 1.0f, 1.0f);
konsttemp = float4(k[0].rgb, 1.0f);
ckonsttemp = frac(konsttemp * (255.0f/256.0f)) * (256.0f/255.0f);
cc0 = frac(c0 * (255.0f/256.0f)) * (256.0f/255.0f);
// color combine
prev.rgb = saturate((rastemp.rgb)+float3(1.0f, 1.0f, 1.0f)*(cc0.rgb));
// alpha combine
prev.a = saturate(konsttemp.a+float4(0.0f, 0.0f, 0.0f, 0.0f).a);

// TEV done
ocol0 = prev;
}

DX11 Backend settings (very similiar to my OpenGL backend settings)

Spoiler: (Show Spoiler)

[Hardware]
VSync = False
Adapter = 0
[Settings]
AspectRatio = 1
Crop = False
wideScreenHack = False
UseXFB = False
UseRealXFB = False
SafeTextureCacheColorSamples = 128
ShowFPS = False
LogFPSToFile = False
ShowInputDisplay = False
OverlayStats = False
OverlayProjStats = False
DLOptimize = 0
Show = 0
DumpTextures = False
HiresTextures = False
DumpEFBTarget = False
DumpFrames = False
FreeLook = False
UseFFV1 = False
AnaglyphStereo = False
AnaglyphStereoSeparation = 200
AnaglyphFocalAngle = 0
EnablePixelLighting = True
EnablePerPixelDepth = False
ShowEFBCopyRegions = False
MSAA = 53
EFBScale = 7
TexFmtOverlayEnable = False
TexFmtOverlayCenter = False
Wireframe = False
DstAlphaPass = False
DisableFog = False
EnableOpenCL = False
OMPDecoder = True
EnableShaderDebugging = False
[Enhancements]
ForceFiltering = True
MaxAnisotropy = 4
PostProcessingShader =
Enable3dVision = False
[Hacks]
EFBAccessEnable = True
DlistCachingEnable = False
EFBCopyEnable = True
EFBCopyDisableHotKey = False
EFBToTextureEnable = True
EFBScaledCopy = True
EFBCopyCacheEnable = False
EFBEmulateFormatChanges = False

I'm not asking for any workarounds, just a thread I created to report this mess. If you want to know more, check my profile to see my specs and Dolphin build.

[LordVador]

Next time try to post in the official game thread (when there is one).
Mario Golf: http://forums.dolphin-emu.org/Thread-gc-mario-golf-toadstool-tour--26004

[Haliinen]

(11-21-2012, 11:20 PM)LordVador Wrote: Next time try to post in the official game thread (when there is one).
Mario Golf: http://forums.dolphin-emu.org/Thread-gc-mario-golf-toadstool-tour--26004

Noted! Thanks for the heads up.

[Shonumi]

Usually the "failed to compile pixel shaders" is the result of an inadequate GPU, but that's clearly not the case with the GTX 580. Seems like it could be a driver issue. Even if you don't have any issues anywhere else, that doesn't necessarily mean you won't have issues elsewhere (i.e. with Dolphin). You could try reverting to previous drivers, especially if you noticed that they did fine with Dolphin.

You might also want to give the D3D9 backend a try over the D3D11 or OGL ones.

[neobrain]

(11-22-2012, 03:17 AM)Shonumi Wrote: Usually the "failed to compile pixel shaders" is the result of an inadequate GPU, but that's clearly not the case with the GTX 580. Seems like it could be a driver issue. Even if you don't have any issues anywhere else, that doesn't necessarily mean you won't have issues elsewhere (i.e. with Dolphin). You could try reverting to previous drivers, especially if you noticed that they did fine with Dolphin.

You might also want to give the D3D9 backend a try over the D3D11 or OGL ones.

It's never a driver issue. Each and every time I've heard that suggestion here it was wrong :p Only exception might be the OS X lighting issue, but w/e

Fwiw, try clearing the contents of your User/ShaderCache folder. nevermind that one.

Try disabling per-pixel lighting, you seem to have that enabled.

[Shonumi]

(11-22-2012, 05:35 AM)neobrain Wrote: It's never a driver issue. Each and every time I've heard that suggestion here it was wrong :p

It's never been a driver issue, yet. One day though, Nvidia or AMD might screw up. Doubtful, but not impossible Anyway, good find with disabling Per-Pixel lighting. Some forum searching reveals it's like the #2 reason for failed compilation of pixel shaders, right after some of Intel's IGPs.

[neobrain]

Yeah, it's an actual bug in Dolphin code in this case, though

[Haliinen]

(11-22-2012, 05:35 AM)neobrain Wrote:

(11-22-2012, 03:17 AM)Shonumi Wrote: Usually the "failed to compile pixel shaders" is the result of an inadequate GPU, but that's clearly not the case with the GTX 580. Seems like it could be a driver issue. Even if you don't have any issues anywhere else, that doesn't necessarily mean you won't have issues elsewhere (i.e. with Dolphin). You could try reverting to previous drivers, especially if you noticed that they did fine with Dolphin.

You might also want to give the D3D9 backend a try over the D3D11 or OGL ones.

It's never a driver issue. Each and every time I've heard that suggestion here it was wrong :p Only exception might be the OS X lighting issue, but w/e

Fwiw, try clearing the contents of your User/ShaderCache folder. nevermind that one.

Try disabling per-pixel lighting, you seem to have that enabled.

I'll give it a shot! Seeing that other games I've tried works without a hitch (on Dolphin) I'd assume it's an issue related to Dolphin and not to my drivers.

EDIT0: I tried with per-pixel lightning off and still got the same problem.

Crash message

Spoiler: (Show Spoiler)

---------------------------
Warning
---------------------------
Failed to compile pixel shader!
This usually happens when trying to use Dolphin with an outdated GPU or integrated GPU like the Intel GMA series.

If you're sure this is Dolphin's error anyway, post the contents of ./User/Dump/bad_ps_0000.txt along with this error message at the forums.

Debug info (ps_5_0):
(29,45): error X3004: undeclared identifier 'uv2'

---------------------------
OK
---------------------------

Text file dump

Spoiler: (Show Spoiler)

//Pixel Shader for TEV stages
//1 TEV stages, 9 texgens, XXX IND stages
sampler samp0 : register(s0), samp1 : register(s1), samp2 : register(s2), samp3 : register(s3), samp4 : register(s4), samp5 : register(s5), samp6 : register(s6), samp7 : register(s7);
Texture2D Tex0 : register(t0), Tex1 : register(t1), Tex2 : register(t2), Tex3 : register(t3), Tex4 : register(t4), Tex5 : register(t5), Tex6 : register(t6), Tex7 : register(t7);

uniform float4 color[4] : register(c0);
uniform float4 k[4] : register(c4);
uniform float4 alphaRef[1] : register(c8);
uniform float4 texdim[8] : register(c9);
uniform float4 czbias[2] : register(c17);
uniform float4 cindscale[2] : register(c19);
uniform float4 cindmtx[6] : register(c21);
uniform float4 cfog[3] : register(c27);
void main(
out float4 ocol0 : SV_Target0,
in float4 rawpos : SV_Position,
in float4 colors_0 : COLOR0,
in float4 colors_1 : COLOR1,
in float4 uv0 : TEXCOORD0 ) {
float4 c0 = color[1], c1 = color[2], c2 = color[3], prev = float4(0.0f, 0.0f, 0.0f, 0.0f), textemp = float4(0.0f, 0.0f, 0.0f, 0.0f), rastemp = float4(0.0f, 0.0f, 0.0f, 0.0f), konsttemp = float4(0.0f, 0.0f, 0.0f, 0.0f);
float3 comp16 = float3(1.0f, 255.0f, 0.0f), comp24 = float3(1.0f, 255.0f, 255.0f*255.0f);
float4 alphabump=float4(0.0f,0.0f,0.0f,0.0f);
float3 tevcoord=float3(0.0f, 0.0f, 0.0f);
float2 wrappedcoord=float2(0.0f,0.0f), tempcoord=float2(0.0f,0.0f);
float4 cc0=float4(0.0f,0.0f,0.0f,0.0f), cc1=float4(0.0f,0.0f,0.0f,0.0f);
float4 cc2=float4(0.0f,0.0f,0.0f,0.0f), cprev=float4(0.0f,0.0f,0.0f,0.0f);
float4 crastemp=float4(0.0f,0.0f,0.0f,0.0f),ckonsttemp=float4(0.0f,0.0f,0.0f,0.0f);

float4 clipPos = float4(rawpos.x, rawpos.y, uv2.w, uv3.w);
uv0.xy = uv0.xy * texdim[0].zw;
if (uv1.z) uv1.xy = uv1.xy / uv1.z;
uv1.xy = uv1.xy * texdim[1].zw;
if (uv2.z) uv2.xy = uv2.xy / uv2.z;
uv2.xy = uv2.xy * texdim[2].zw;
uv3.xy = uv3.xy * texdim[3].zw;
uv4.xy = uv4.xy * texdim[4].zw;
uv5.xy = uv5.xy * texdim[5].zw;
uv6.xy = uv6.xy * texdim[6].zw;
uv7.xy = uv7.xy * texdim[7].zw;
uv8.xy = uv8.xy * texdim[8].zw;
// TEV stage 0
rastemp = colors_0.rgba;
crastemp = frac(rastemp * (255.0f/256.0f)) * (256.0f/255.0f);
tevcoord.xy = uv0.xy;
textemp=Tex0.Sample(samp0,tevcoord.xy * texdim[0].xy).rgba;
// color combine
prev.rgb = saturate((rastemp.rgb)+lerp((crastemp.rgb), float3(1.0f, 1.0f, 1.0f), (textemp.aaa)));
// alpha combine
prev.a = saturate(2.0f*(crastemp.a*textemp.a));

// TEV done
if(!( (prev.a > alphaRef[0].r - (0.25f/255.0f)) || (false))) {
ocol0 = 0;
discard;
}
float zCoord = czbias[1].x + (clipPos.z / clipPos.w) * czbias[1].y;
float ze = cfog[1].x / (cfog[1].y - (zCoord / cfog[1].w));
float x_adjust = (2.0f * (clipPos.x / cfog[2].y)) - 1.0f - cfog[2].x;
x_adjust = sqrt(x_adjust * x_adjust + cfog[2].z * cfog[2].z) / cfog[2].z;
ze *= x_adjust;
float fog = saturate(ze - cfog[1].z);
prev.rgb = lerp(prev.rgb,cfog[0].rgb,fog);
ocol0 = prev;
}

So yeah, idk what's cooking there.

EDIT1: Posting this to let you guys know I'm not screwing you over with outdated hardware

Spoiler: (Show Spoiler)

[Shonumi]

Have you tried using Direct3D9 for the graphics backend yet?

[LordVador]

Tried with older builds? Tried with Dolphin 3.0?