izzy/UnrealEngineUWP
0
0
Fork 0
mirror of https://github.com/izzy2lost/UnrealEngineUWP.git synced 2026-03-26 18:15:20 -07:00
Code Issues Packages Projects Releases Wiki Activity

Addressing a number of DXC shader validation errors and warnings

Browse Source 
#rb David.Harvey
#jira none
#rnx
#ushell-cherrypick of 12330720 by eric.mcdaniel

#ROBOMERGE-SOURCE: CL 12726817 in //UE4/Release-4.25Plus/... via CL 12781740
#ROBOMERGE-BOT: RELEASE (Release-Engine-Staging -> Main) (v681-12776863)

[CL 12782554 by eric mcdaniel in Main branch]
This commit is contained in:
eric mcdaniel
2020-04-14 13:40:58 -04:00
parent 0395fb0d3d
commit e167288eb5
25 changed files with 567 additions and 482 deletions
Download Patch File Download Diff File Expand all files Collapse all files
Engine/Shaders/Private/ComposeSeparateTranslucency.usf
+1 -1
View File
@@ -15,7 +15,7 @@ void MainPS(
float4 SvPosition : SV_POSITION,
out float4 OutColor : SV_Target0)
{
float2 BufferUV = SvPosition * View.BufferSizeAndInvSize.zw;
float2 BufferUV = SvPosition.xy * View.BufferSizeAndInvSize.zw;
float4 SceneColorSample = SceneColor.SampleLevel(SceneColorSampler, BufferUV, 0);
float4 SeparateTranslucencySample = SeparateTranslucency.SampleLevel(SeparateTranslucencySampler, BufferUV, 0);
Engine/Shaders/Private/CompositeUIPixelShader.usf
+1 -1
View File
@@ -100,7 +100,7 @@ void Main(
)
{
#if PLATFORM_SUPPORTS_RENDERTARGET_WRITE_MASK
uint CompositeUIMask = DecodeRTWriteMask(Input.Position, UIWriteMaskTexture, 1);
uint CompositeUIMask = DecodeRTWriteMask(Input.Position.xy, UIWriteMaskTexture, 1);
BRANCH
if (CompositeUIMask == 0)
{
Engine/Shaders/Private/DistanceFieldAOShared.ush
+2 -2
View File
@@ -124,8 +124,8 @@ bool SphereIntersectConeWithDepthRanges(float4 SphereCenterAndRadius, float3 Con
float ConeAxisDistance = dot(SphereCenterAndRadius.xyz - ConeVertex, ConeAxis);
float2 ConeAxisDistanceMinMax = float2(ConeAxisDistance + SphereCenterAndRadius.w, ConeAxisDistance - SphereCenterAndRadius.w);
if (ConeAxisDistanceMinMax.x > ConeAxisDepthRanges.x && ConeAxisDistanceMinMax.y < ConeAxisDepthRanges.y
|| ConeAxisDistanceMinMax.x > ConeAxisDepthRanges.z && ConeAxisDistanceMinMax.y < ConeAxisDepthRanges.w)
if ((ConeAxisDistanceMinMax.x > ConeAxisDepthRanges.x && ConeAxisDistanceMinMax.y < ConeAxisDepthRanges.y)
|| (ConeAxisDistanceMinMax.x > ConeAxisDepthRanges.z && ConeAxisDistanceMinMax.y < ConeAxisDepthRanges.w))
{
return true;
}
Engine/Shaders/Private/DistanceFieldObjectCulling.usf
+30 -18
View File
@@ -41,16 +41,20 @@ void UploadObjectsToBufferCS(
{
uint DestIndex = UploadOperationIndices[UploadOperationIndex];
UNROLL
for (uint VectorIndex = 0; VectorIndex < OBJECT_BOUNDS_STRIDE; ++VectorIndex)
{
UploadBoundsFloat4(DestIndex * OBJECT_BOUNDS_STRIDE + VectorIndex, UploadOperationIndex * UPLOAD_DATA_STRIDE + VectorIndex);
UNROLL
for (uint VectorIndex = 0; VectorIndex < OBJECT_BOUNDS_STRIDE; ++VectorIndex)
{
UploadBoundsFloat4(DestIndex * OBJECT_BOUNDS_STRIDE + VectorIndex, UploadOperationIndex * UPLOAD_DATA_STRIDE + VectorIndex);
}
}
UNROLL
for (uint VectorIndex = 0; VectorIndex < OBJECT_DATA_STRIDE; ++VectorIndex)
{
UploadDataFloat4(DestIndex * OBJECT_DATA_STRIDE + VectorIndex, UploadOperationIndex * UPLOAD_DATA_STRIDE + OBJECT_BOUNDS_STRIDE + VectorIndex);
UNROLL
for (uint VectorIndex = 0; VectorIndex < OBJECT_DATA_STRIDE; ++VectorIndex)
{
UploadDataFloat4(DestIndex * OBJECT_DATA_STRIDE + VectorIndex, UploadOperationIndex * UPLOAD_DATA_STRIDE + OBJECT_BOUNDS_STRIDE + VectorIndex);
}
}
}
}
@@ -86,16 +90,20 @@ void CopyObjectBufferCS(
uint SourceIndex = CopyOperationIndex;
uint DestIndex = CopyOperationIndex;
UNROLL
for (uint VectorIndex = 0; VectorIndex < OBJECT_BOUNDS_STRIDE; ++VectorIndex)
{
CopyBoundsFloat4(DestIndex * OBJECT_BOUNDS_STRIDE + VectorIndex, SourceIndex * OBJECT_BOUNDS_STRIDE + VectorIndex);
UNROLL
for (uint VectorIndex = 0; VectorIndex < OBJECT_BOUNDS_STRIDE; ++VectorIndex)
{
CopyBoundsFloat4(DestIndex * OBJECT_BOUNDS_STRIDE + VectorIndex, SourceIndex * OBJECT_BOUNDS_STRIDE + VectorIndex);
}
}
UNROLL
for (uint VectorIndex = 0; VectorIndex < OBJECT_DATA_STRIDE; VectorIndex++)
{
CopyDataFloat4(DestIndex * OBJECT_DATA_STRIDE + VectorIndex, SourceIndex * OBJECT_DATA_STRIDE + VectorIndex);
UNROLL
for (uint VectorIndex = 0; VectorIndex < OBJECT_DATA_STRIDE; VectorIndex++)
{
CopyDataFloat4(DestIndex * OBJECT_DATA_STRIDE + VectorIndex, SourceIndex * OBJECT_DATA_STRIDE + VectorIndex);
}
}
}
}
@@ -144,16 +152,20 @@ void RemoveObjectsFromBufferCS(
uint SourceIndex = RemoveOperationIndices[RemoveOperationIndex].y;
uint DestIndex = RemoveOperationIndices[RemoveOperationIndex].x;
UNROLL
for (uint VectorIndex = 0; VectorIndex < OBJECT_BOUNDS_STRIDE; ++VectorIndex)
{
WriteBoundsFloat4(DestIndex * OBJECT_BOUNDS_STRIDE + VectorIndex, SourceIndex * OBJECT_BOUNDS_STRIDE + VectorIndex);
UNROLL
for (uint VectorIndex = 0; VectorIndex < OBJECT_BOUNDS_STRIDE; ++VectorIndex)
{
WriteBoundsFloat4(DestIndex * OBJECT_BOUNDS_STRIDE + VectorIndex, SourceIndex * OBJECT_BOUNDS_STRIDE + VectorIndex);
}
}
UNROLL
for (uint VectorIndex = 0; VectorIndex < OBJECT_DATA_STRIDE; VectorIndex++)
{
WriteDataFloat4(DestIndex * OBJECT_DATA_STRIDE + VectorIndex, SourceIndex * OBJECT_DATA_STRIDE + VectorIndex);
UNROLL
for (uint VectorIndex = 0; VectorIndex < OBJECT_DATA_STRIDE; VectorIndex++)
{
WriteDataFloat4(DestIndex * OBJECT_DATA_STRIDE + VectorIndex, SourceIndex * OBJECT_DATA_STRIDE + VectorIndex);
}
}
}
}
Engine/Shaders/Private/DistanceFieldShadowing.usf
+2 -2
View File
@@ -146,8 +146,8 @@ void CullObjectsForShadowCS(
#if (DISTANCEFIELD_PRIMITIVE_TYPE == DFPT_SignedDistanceField)
float4 ObjectBoundingSphere = LoadFloat4(SceneObjectBounds, SourceIndex);
if (ShadowBoundingSphere.w == 0 && ShadowConvexHullIntersectSphere(ObjectBoundingSphere.xyz, ObjectBoundingSphere.w)
|| ShadowBoundingSphere.w > 0 && dot(ShadowBoundingSphere.xyz - ObjectBoundingSphere.xyz, ShadowBoundingSphere.xyz - ObjectBoundingSphere.xyz) < Square(ShadowBoundingSphere.w + ObjectBoundingSphere.w))
if ((ShadowBoundingSphere.w == 0 && ShadowConvexHullIntersectSphere(ObjectBoundingSphere.xyz, ObjectBoundingSphere.w))
|| (ShadowBoundingSphere.w > 0 && dot(ShadowBoundingSphere.xyz - ObjectBoundingSphere.xyz, ShadowBoundingSphere.xyz - ObjectBoundingSphere.xyz) < Square(ShadowBoundingSphere.w + ObjectBoundingSphere.w)))
{
// Assume ObjectBoundingSphere is located at (0, 0, 0) in local space
float ViewDist2 = length2(ObjectBoundingSphere.xyz - View.WorldViewOrigin);
Engine/Shaders/Private/FXAAShader.usf
-1
View File
@@ -66,7 +66,6 @@ void FxaaVS(
#endif
#define DECLARE_PARAMETER(a, b, c)
float4 fxaaConstDir = float4(1.0, -1.0, 0.25, -0.25);
// see Fxaa.. header file for more details
float4 fxaaConsoleRcpFrameOpt;
Engine/Shaders/Private/GPUFastFourierTransform.usf
+12 -9
View File
@@ -489,9 +489,8 @@ void ComplexMultTexture(bool bIsHorizontal, bool bUseAlpha, bool bIsGAGroup, in
//float4 FilterTint;
void ApplyTint(in Complex Tint, inout Complex LocalBuffer[2][RADIX])
{
for (uint r = 0; r < RADIX; ++r) LocalBuffer[0][r] *= Tint.x;
for (uint r = 0; r < RADIX; ++r) LocalBuffer[1][r] *= Tint.y;
{ for (uint r = 0; r < RADIX; ++r) LocalBuffer[0][r] *= Tint.x; }
{ for (uint r = 0; r < RADIX; ++r) LocalBuffer[1][r] *= Tint.y; }
}
@@ -632,19 +631,23 @@ void GSConvolutionWithTextureCS(uint3 GroupID : SV_GroupID, uint3 GroupThreadID
// Normalize R & G
for (uint r = 0; r < RADIX; ++r)
{
// This is the R or G channel
LocalBuffer[0][r] /= NormMax;
for (uint r = 0; r < RADIX; ++r)
{
// This is the R or G channel
LocalBuffer[0][r] /= NormMax;
}
}
float AorBNorm = (bIsGAGroup) ? AlphaSum : NormMax;
// Normalize B & A
for (uint r = 0; r < RADIX; ++r)
{
// This is the B or A channel
LocalBuffer[1][r] /= AorBNorm;
for (uint r = 0; r < RADIX; ++r)
{
// This is the B or A channel
LocalBuffer[1][r] /= AorBNorm;
}
}
}
Engine/Shaders/Private/GPUFastFourierTransform2DCore.ush
+4 -4
View File
@@ -114,8 +114,8 @@ void CopyDataLocalToDst(in Complex LocalBuffer[2][RADIX], bool bIsHorizontal, ui
void CopyDataSrcWindowToLocal(inout Complex LocalBuffer[2][RADIX], bool bIsHorizontal, in uint ScanIdx, uint Loc, uint Stride, uint4 Window)
{
for (uint i = 0; i < RADIX; ++i) LocalBuffer[0][ i ] = float2(0.f, 0.f);
for (uint i = 0; i < RADIX; ++i) LocalBuffer[1][ i ] = float2(0.f, 0.f);
{ for (uint i = 0; i < RADIX; ++i) LocalBuffer[0][ i ] = float2(0.f, 0.f); }
{ for (uint i = 0; i < RADIX; ++i) LocalBuffer[1][ i ] = float2(0.f, 0.f); }
if (bIsHorizontal)
{
@@ -158,8 +158,8 @@ void CopyDataSrcWindowToLocal(inout Complex LocalBuffer[2][RADIX], bool bIsHoriz
void CopyDataSrcWindowToLocal(inout Complex LocalBuffer[2][RADIX], bool bIsHorizontal, uint ScanIdx, uint Loc, uint Stride, uint2 WindowMin, uint2 WindowMax )
{
for (uint i = 0; i < RADIX; ++i) LocalBuffer[0][ i ] = float2(0.f, 0.f);
for (uint i = 0; i < RADIX; ++i) LocalBuffer[1][ i ] = float2(0.f, 0.f);
{ for (uint i = 0; i < RADIX; ++i) LocalBuffer[0][ i ] = float2(0.f, 0.f); }
{ for (uint i = 0; i < RADIX; ++i) LocalBuffer[1][ i ] = float2(0.f, 0.f); }
if (bIsHorizontal)
{
Engine/Shaders/Private/GPUFastFourierTransformCore.ush
+183 -149
View File
@@ -376,17 +376,19 @@ groupshared float SharedImag[ SCAN_LINE_LENGTH ];
void CopyLocalToGroupShared(in Complex Local[RADIX], in uint Head, in uint Stride)
{
for (uint r = 0, i = Head; r < RADIX; ++r, i += Stride)
{
SharedReal[ i ] = Local[ r ].x;
for (uint r = 0, i = Head; r < RADIX; ++r, i += Stride)
{
SharedReal[ i ] = Local[ r ].x;
}
}
for (uint r = 0, i = Head; r < RADIX; ++r, i += Stride)
{
SharedImag[ i ] = Local[ r ].y;
for (uint r = 0, i = Head; r < RADIX; ++r, i += Stride)
{
SharedImag[ i ] = Local[ r ].y;
}
}
}
void CopyLocalToGroupSharedWSync(in Complex Local[RADIX], in uint Head, in uint Stride)
@@ -398,15 +400,18 @@ void CopyLocalToGroupSharedWSync(in Complex Local[RADIX], in uint Head, in uint
void CopyGroupSharedToLocal(inout Complex Local[RADIX], in uint Head, in uint Stride)
{
for (uint r = 0, i = Head; r < RADIX; ++r, i += Stride)
{
Local[ r ].x = SharedReal[ i ];
for (uint r = 0, i = Head; r < RADIX; ++r, i += Stride)
{
Local[ r ].x = SharedReal[ i ];
}
}
for (uint r = 0, i = Head; r < RADIX; ++r, i += Stride)
{
Local[ r ].y = SharedImag[ i ];
for (uint r = 0, i = Head; r < RADIX; ++r, i += Stride)
{
Local[ r ].y = SharedImag[ i ];
}
}
}
@@ -616,13 +621,18 @@ void GroupSharedFFT(in const bool bIsForward, inout Complex Local[RADIX], in con
void Scale(inout Complex LocalBuffer[2][RADIX], in float ScaleValue)
{
// Scale
for (uint r = 0; r < RADIX; ++r)
{
LocalBuffer[0][r] *= ScaleValue;
{
for (uint r = 0; r < RADIX; ++r)
{
LocalBuffer[0][r] *= ScaleValue;
}
}
for (uint r = 0; r < RADIX; ++r)
{
LocalBuffer[1][r] *= ScaleValue;
{
for (uint r = 0; r < RADIX; ++r)
{
LocalBuffer[1][r] *= ScaleValue;
}
}
}
@@ -709,44 +719,50 @@ void SplitTwoForOne(inout Complex LocalBuffer[RADIX], in uint Head, in uint Stri
// Construct the transform for the two real signals in the LocalBuffer
UNROLL
for (uint i = 0, K = Head; i < RADIX; ++i, K += Stride)
{
//
// K = N/2 - abs(SrcIdx - N/2)
UNROLL
for (uint i = 0, K = Head; i < RADIX; ++i, K += Stride)
{
//
// K = N/2 - abs(SrcIdx - N/2)
// DstK = SrcIdx % Non2;
// N - k
// DstK = SrcIdx % Non2;
// N - k
FLATTEN uint NmK = (K > 0) ? ( N - K) : 0;
uint NmK = (K > 0) ? ( N - K) : 0;
// Z_k = LocalBuffer[i]
// If k < N/2 : Store F_k = 1/2 (Z_k + Z*_{N-k})
// If k > N/2 : Compute I*G_k = 1/2 (Z_k - Z*_{N-k})
// Z_k = LocalBuffer[i]
// If k < N/2 : Store F_k = 1/2 (Z_k + Z*_{N-k})
// If k > N/2 : Compute I*G_k = 1/2 (Z_k - Z*_{N-k})
// Tmp = {+,-}ComplexConjugate( Z_{N-k})
Complex Tmp = Complex( SharedReal[NmK], -SharedImag[NmK] );
// Tmp = {+,-}ComplexConjugate( Z_{N-k})
Complex Tmp = Complex( SharedReal[NmK], -SharedImag[NmK] );
Tmp *= (K > Non2)? -1 : 1;
Tmp *= (K > Non2)? -1 : 1;
LocalBuffer[i] += Tmp;
LocalBuffer[i] += Tmp;
}
}
for (uint i =0; i < RADIX; ++i) LocalBuffer[ i ] *= 0.5f;
UNROLL
for (uint i = 0, K = Head; i < RADIX; ++i, K += Stride)
{
// If k > N/2 get G_k from I*G_k: G_k = -I * (I G_k)
if (K > Non2) LocalBuffer[i] = ComplexMult(Complex(0, -1), LocalBuffer[i] );
if (K == Non2)
{
// F_N/2 + I * G_N/2
LocalBuffer[i] = Complex(SharedReal[Non2], SharedImag[Non2]);
}
for (uint i =0; i < RADIX; ++i) LocalBuffer[ i ] *= 0.5f;
}
{
UNROLL
for (uint i = 0, K = Head; i < RADIX; ++i, K += Stride)
{
// If k > N/2 get G_k from I*G_k: G_k = -I * (I G_k)
if (K > Non2) LocalBuffer[i] = ComplexMult(Complex(0, -1), LocalBuffer[i] );
if (K == Non2)
{
// F_N/2 + I * G_N/2
LocalBuffer[i] = Complex(SharedReal[Non2], SharedImag[Non2]);
}
}
}
if (Head == 0)
@@ -773,27 +789,29 @@ void SplitTwoForOne(inout Complex LocalBuffer[RADIX], in uint Head, in uint Stri
// Construct the transform for the two real signals in the LocalBuffer
UNROLL
for (uint i = 0, K = Head; i < RADIX; ++i, K += Stride)
{
//
// K = N/2 - abs(SrcIdx - N/2)
UNROLL
for (uint i = 0, K = Head; i < RADIX; ++i, K += Stride)
{
//
// K = N/2 - abs(SrcIdx - N/2)
// DstK = SrcIdx % Non2;
// N - k
// DstK = SrcIdx % Non2;
// N - k
FLATTEN uint NmK = (K > 0) ? ( N - K) : 0;
uint NmK = (K > 0) ? ( N - K) : 0;
// Z_k = LocalBuffer[i]
// If k < N/2 : Store F_k = 1/2 (Z_k + Z*_{N-k})
// If k > N/2 : Compute I*G_k = 1/2 (Z_k - Z*_{N-k})
// Z_k = LocalBuffer[i]
// If k < N/2 : Store F_k = 1/2 (Z_k + Z*_{N-k})
// If k > N/2 : Compute I*G_k = 1/2 (Z_k - Z*_{N-k})
// Tmp = {+,-}ComplexConjugate( Z_{N-k})
float Tmp = SharedReal[NmK];
// Tmp = {+,-}ComplexConjugate( Z_{N-k})
float Tmp = SharedReal[NmK];
Tmp *= (K > Non2)? -1 : 1;
Tmp *= (K > Non2)? -1 : 1;
LocalBuffer[i].x += Tmp;
LocalBuffer[i].x += Tmp;
}
}
if (Head == 0 ) LocalBuffer[0].x = 2.f * SharedReal[0];
@@ -804,39 +822,45 @@ void SplitTwoForOne(inout Complex LocalBuffer[RADIX], in uint Head, in uint Stri
FFTMemoryBarrier();
UNROLL
for (uint i = 0, K = Head; i < RADIX; ++i, K += Stride)
{
//
// K = N/2 - abs(SrcIdx - N/2)
UNROLL
for (uint i = 0, K = Head; i < RADIX; ++i, K += Stride)
{
//
// K = N/2 - abs(SrcIdx - N/2)
// DstK = SrcIdx % Non2;
// N - k
// DstK = SrcIdx % Non2;
// N - k
FLATTEN uint NmK = (K > 0) ? ( N - K) : 0;
uint NmK = (K > 0) ? ( N - K) : 0;
// Z_k = LocalBuffer[i]
// If k < N/2 : Store F_k = 1/2 (Z_k + Z*_{N-k})
// If k > N/2 : Compute I*G_k = 1/2 (Z_k - Z*_{N-k})
// Z_k = LocalBuffer[i]
// If k < N/2 : Store F_k = 1/2 (Z_k + Z*_{N-k})
// If k > N/2 : Compute I*G_k = 1/2 (Z_k - Z*_{N-k})
// Tmp = {+,-}ComplexConjugate( Z_{N-k})
float Tmp = -SharedReal[NmK];
// Tmp = {+,-}ComplexConjugate( Z_{N-k})
float Tmp = -SharedReal[NmK];
Tmp *= (K < Non2)? 1 : -1;
Tmp *= (K < Non2)? 1 : -1;
LocalBuffer[i].y += Tmp;
//LocalBuffer[i] *= 0.5;
LocalBuffer[i].y += Tmp;
//LocalBuffer[i] *= 0.5;
}
}
if (Head == 0) LocalBuffer[0].y = 2.f * SharedReal[0];
UNROLL for (uint i = 0; i < RADIX; ++i) LocalBuffer[i] *= 0.5;
UNROLL
for (uint i = 0, K = Head; i < RADIX; ++i, K += Stride)
{
// If k > N/2 get G_k from I*G_k: G_k = -I * (I G_k)
if (K > Non2) LocalBuffer[i] = ComplexMult(Complex(0, -1), LocalBuffer[i] );
UNROLL for (uint i = 0; i < RADIX; ++i) LocalBuffer[i] *= 0.5;
}
{
UNROLL
for (uint i = 0, K = Head; i < RADIX; ++i, K += Stride)
{
// If k > N/2 get G_k from I*G_k: G_k = -I * (I G_k)
if (K > Non2) LocalBuffer[i] = ComplexMult(Complex(0, -1), LocalBuffer[i] );
}
}
}
@@ -880,45 +904,49 @@ void MergeTwoForOne(inout Complex LocalBuffer[RADIX], in uint Head, in uint Stri
// Compose the transform of a single f+ig signal from the two real signals in the LocalBuffer
UNROLL
for (uint i = 0, K = Head; i < RADIX; ++i, K += Stride)
{
UNROLL
for (uint i = 0, K = Head; i < RADIX; ++i, K += Stride)
{
// N - k
FLATTEN uint NmK = (K > 0) ? (N - K) : 0 ;
// N - k
uint NmK = (K > 0) ? (N - K) : 0 ;
// If k < N/2 : LocalBuffer[i] = F_k,
// Shared[N-K] = G[N-K] = Conjugate(G[k]) = (G[k]_r, -G[k]_i) = Complex(G_r, -G_i)
// want I G[k] = float(-G[k]_i, G[k]_r)
// Tmp = I * G_k
// If k > N/2 : LocalBuffer[i] = G_k
// Shared[N-K] = F[N-k] = Conjugate(F[k]) = Complex(F[k]_r, -F[k]_i)
// want -I F[k] = Complex(F[k]_i, -F[k]_r)
// Tmp = -I * F
// ComplexConjugate( Z_{N-k})
// If k < N/2 : LocalBuffer[i] = F_k,
// Shared[N-K] = G[N-K] = Conjugate(G[k]) = (G[k]_r, -G[k]_i) = Complex(G_r, -G_i)
// want I G[k] = float(-G[k]_i, G[k]_r)
// Tmp = I * G_k
// If k > N/2 : LocalBuffer[i] = G_k
// Shared[N-K] = F[N-k] = Conjugate(F[k]) = Complex(F[k]_r, -F[k]_i)
// want -I F[k] = Complex(F[k]_i, -F[k]_r)
// Tmp = -I * F
// ComplexConjugate( Z_{N-k})
// Tmp = (K < Non2) ? I * G_k : -I * F
Complex Tmp = Complex( SharedImag[ NmK ], SharedReal[ NmK ] );
Tmp *= (K > Non2) ? -1 : 1;
// Tmp = (K < Non2) ? I * G_k : -I * F
Complex Tmp = Complex( SharedImag[ NmK ], SharedReal[ NmK ] );
Tmp *= (K > Non2) ? -1 : 1;
LocalBuffer[i] += Tmp;
LocalBuffer[i] += Tmp;
}
}
// Compose the transform of a single f+ig signal from the two real signals in the LocalBuffer
UNROLL
for (uint i = 0, K = Head; i < RADIX; ++i, K += Stride)
{
// if k > N/2 we have G - I * F. Multiply by I to get F + I * G
if (K > Non2) LocalBuffer[ i ] = ComplexMult(Complex(0, 1), LocalBuffer[ i ]);
UNROLL
for (uint i = 0, K = Head; i < RADIX; ++i, K += Stride)
{
// if k > N/2 we have G - I * F. Multiply by I to get F + I * G
if (K > Non2) LocalBuffer[ i ] = ComplexMult(Complex(0, 1), LocalBuffer[ i ]);
if (K == Non2)
{
// F_N/2 + I * G_N/2
LocalBuffer[ i ] = Complex(SharedReal[ Non2 ], SharedImag[ Non2 ]);
}
if (K == Non2)
{
// F_N/2 + I * G_N/2
LocalBuffer[ i ] = Complex(SharedReal[ Non2 ], SharedImag[ Non2 ]);
}
}
if (Head == 0)
@@ -936,37 +964,41 @@ void MergeTwoForOne(inout Complex LocalBuffer[RADIX], in uint Head, in uint Stri
// Write the two FFTs into shared memory.
float TmpX[RADIX];
for (uint i = 0; i < RADIX; ++i) TmpX[i] = LocalBuffer[i].x;
{
for (uint i = 0; i < RADIX; ++i) TmpX[i] = LocalBuffer[i].x;
}
CopyLocalYToGroupShared(LocalBuffer, Head, Stride);
FFTMemoryBarrier();
// Compose the transform of a single f+ig signal from the two real signals in the LocalBuffer
UNROLL
for (uint i = 0, K = Head; i < RADIX; ++i, K += Stride)
{
UNROLL
for (uint i = 0, K = Head; i < RADIX; ++i, K += Stride)
{
// N - k
FLATTEN uint NmK = (K > 0) ? (N - K) : 0 ;
// N - k
uint NmK = (K > 0) ? (N - K) : 0 ;
// If k < N/2 : LocalBuffer[i] = F_k,
// Shared[N-K] = G[N-K] = Conjugate(G[k]) = (G[k]_r, -G[k]_i) = Complex(G_r, -G_i)
// want I G[k] = float(-G[k]_i, G[k]_r)
// Tmp = I * G_k
// If k > N/2 : LocalBuffer[i] = G_k
// Shared[N-K] = F[N-k] = Conjugate(F[k]) = Complex(F[k]_r, -F[k]_i)
// want -I F[k] = Complex(F[k]_i, -F[k]_r)
// Tmp = -I * F
// ComplexConjugate( Z_{N-k})
// If k < N/2 : LocalBuffer[i] = F_k,
// Shared[N-K] = G[N-K] = Conjugate(G[k]) = (G[k]_r, -G[k]_i) = Complex(G_r, -G_i)
// want I G[k] = float(-G[k]_i, G[k]_r)
// Tmp = I * G_k
// If k > N/2 : LocalBuffer[i] = G_k
// Shared[N-K] = F[N-k] = Conjugate(F[k]) = Complex(F[k]_r, -F[k]_i)
// want -I F[k] = Complex(F[k]_i, -F[k]_r)
// Tmp = -I * F
// ComplexConjugate( Z_{N-k})
// Tmp = (K < Non2) ? I * G_k : -I * F
float Tmp = SharedReal[ NmK ];
Tmp *= (K > Non2) ? -1 : 1;
// Tmp = (K < Non2) ? I * G_k : -I * F
float Tmp = SharedReal[ NmK ];
Tmp *= (K > Non2) ? -1 : 1;
LocalBuffer[i].x += Tmp;
LocalBuffer[i].x += Tmp;
}
}
float2 FirstElement = float2(0, SharedReal[0]);
@@ -985,45 +1017,47 @@ void MergeTwoForOne(inout Complex LocalBuffer[RADIX], in uint Head, in uint Stri
FirstElement.x = SharedReal[0];
MiddleElement.x = SharedReal[Non2];
UNROLL
for (uint i = 0, K = Head; i < RADIX; ++i, K += Stride)
{
UNROLL
for (uint i = 0, K = Head; i < RADIX; ++i, K += Stride)
{
// N - k
FLATTEN uint NmK = (K > 0) ? (N - K) : 0 ;
// N - k
uint NmK = (K > 0) ? (N - K) : 0 ;
// If k < N/2 : LocalBuffer[i] = F_k,
// Shared[N-K] = G[N-K] = Conjugate(G[k]) = (G[k]_r, -G[k]_i) = Complex(G_r, -G_i)
// want I G[k] = float(-G[k]_i, G[k]_r)
// Tmp = I * G_k
// If k > N/2 : LocalBuffer[i] = G_k
// Shared[N-K] = F[N-k] = Conjugate(F[k]) = Complex(F[k]_r, -F[k]_i)
// want -I F[k] = Complex(F[k]_i, -F[k]_r)
// Tmp = -I * F
// ComplexConjugate( Z_{N-k})
// If k < N/2 : LocalBuffer[i] = F_k,
// Shared[N-K] = G[N-K] = Conjugate(G[k]) = (G[k]_r, -G[k]_i) = Complex(G_r, -G_i)
// want I G[k] = float(-G[k]_i, G[k]_r)
// Tmp = I * G_k
// If k > N/2 : LocalBuffer[i] = G_k
// Shared[N-K] = F[N-k] = Conjugate(F[k]) = Complex(F[k]_r, -F[k]_i)
// want -I F[k] = Complex(F[k]_i, -F[k]_r)
// Tmp = -I * F
// ComplexConjugate( Z_{N-k})
// Tmp = (K < Non2) ? I * G_k : -I * F
float Tmp = SharedReal[ NmK ];
Tmp *= (K > Non2) ? -1 : 1;
// Tmp = (K < Non2) ? I * G_k : -I * F
float Tmp = SharedReal[ NmK ];
Tmp *= (K > Non2) ? -1 : 1;
LocalBuffer[i].y += Tmp;
LocalBuffer[i].y += Tmp;
}
}
UNROLL
for (uint i = 0, K = Head; i < RADIX; ++i, K += Stride)
{
// if k > N/2 we have G - I * F. Multiply by I to get F + I * G
if (K > Non2) LocalBuffer[ i ] = ComplexMult(Complex(0, 1), LocalBuffer[ i ]);
UNROLL
for (uint i = 0, K = Head; i < RADIX; ++i, K += Stride)
{
// if k > N/2 we have G - I * F. Multiply by I to get F + I * G
if (K > Non2) LocalBuffer[ i ] = ComplexMult(Complex(0, 1), LocalBuffer[ i ]);
if (K == Non2)
{
// F_N/2 + I * G_N/2
LocalBuffer[ i ] = MiddleElement;
if (K == Non2)
{
// F_N/2 + I * G_N/2
LocalBuffer[ i ] = MiddleElement;
}
}
}
if (Head == 0) LocalBuffer[ 0 ] = FirstElement;
Engine/Shaders/Private/GlobalDistanceField.usf
+46 -42
View File
@@ -238,26 +238,28 @@ void CompositeObjectDistanceFieldsCS(
NumCulledObjects = NumGridCulledObjects;
#endif
for (uint ListObjectIndex = ThreadIndex; ListObjectIndex < NumCulledObjects; ListObjectIndex += COMPOSITE_THREADGORUP_TOTALSIZE)
{
#if USE_CULL_GRID
uint ObjectIndex = CullGridObjectArray[CullTileIndex * MAX_GRID_CULLED_DF_OBJECTS + ListObjectIndex];
#else
uint ObjectIndex = ListObjectIndex;
#endif
float4 ObjectPositionAndRadius = LoadObjectPositionAndRadius(ObjectIndex);
//float BoxDistance = ComputeDistanceFromBoxToPoint(TileBoxCenter - TileBoxExtent, TileBoxCenter + TileBoxExtent, ObjectPositionAndRadius.xyz);
// ComputeSquaredDistanceFromBoxToPoint is correct but adds a lot to the final cost
float BoxDistanceSq = ComputeSquaredDistanceFromBoxToPoint(TileBoxCenter, TileBoxExtent, ObjectPositionAndRadius.xyz);
float CombinedDistance = ObjectPositionAndRadius.w + AOGlobalMaxSphereQueryRadius;
if (BoxDistanceSq < CombinedDistance * CombinedDistance)
for (uint ListObjectIndex = ThreadIndex; ListObjectIndex < NumCulledObjects; ListObjectIndex += COMPOSITE_THREADGORUP_TOTALSIZE)
{
uint DestIndex;
InterlockedAdd(NumTileCulledObjects, 1U, DestIndex);
SharedCulledObjectList[DestIndex] = ObjectIndex;
#if USE_CULL_GRID
uint ObjectIndex = CullGridObjectArray[CullTileIndex * MAX_GRID_CULLED_DF_OBJECTS + ListObjectIndex];
#else
uint ObjectIndex = ListObjectIndex;
#endif
float4 ObjectPositionAndRadius = LoadObjectPositionAndRadius(ObjectIndex);
//float BoxDistance = ComputeDistanceFromBoxToPoint(TileBoxCenter - TileBoxExtent, TileBoxCenter + TileBoxExtent, ObjectPositionAndRadius.xyz);
// ComputeSquaredDistanceFromBoxToPoint is correct but adds a lot to the final cost
float BoxDistanceSq = ComputeSquaredDistanceFromBoxToPoint(TileBoxCenter, TileBoxExtent, ObjectPositionAndRadius.xyz);
float CombinedDistance = ObjectPositionAndRadius.w + AOGlobalMaxSphereQueryRadius;
if (BoxDistanceSq < CombinedDistance * CombinedDistance)
{
uint DestIndex;
InterlockedAdd(NumTileCulledObjects, 1U, DestIndex);
SharedCulledObjectList[DestIndex] = ObjectIndex;
}
}
}
@@ -277,35 +279,37 @@ void CompositeObjectDistanceFieldsCS(
MinDistance = ParentGlobalDistanceFieldTexture[IntCoordinate].r;
#endif
LOOP
for (uint ListObjectIndex = 0; ListObjectIndex < NumCulledObjects; ListObjectIndex++)
{
#if USE_OBJECT_COMPOSITING_TILE_CULLING
uint ObjectIndex = SharedCulledObjectList[ListObjectIndex];
#else
uint ObjectIndex = ListObjectIndex;
#endif
float3 LocalPositionExtent = LoadObjectLocalPositionExtent(ObjectIndex);
float4x4 WorldToVolume = LoadObjectWorldToVolume(ObjectIndex);
bool bGeneratedAsTwoSided;
float4 UVScaleAndVolumeScale = LoadObjectUVScale(ObjectIndex, bGeneratedAsTwoSided);
float3 VolumePosition = mul(float4(WorldPosition, 1), WorldToVolume).xyz;
float BoxDistance = ComputeDistanceFromBoxToPoint(-LocalPositionExtent, LocalPositionExtent, VolumePosition) * UVScaleAndVolumeScale.w;
BRANCH
if (BoxDistance < AOGlobalMaxSphereQueryRadius)
LOOP
for (uint ListObjectIndex = 0; ListObjectIndex < NumCulledObjects; ListObjectIndex++)
{
float3 UVAdd = LoadObjectUVAddAndSelfShadowBias(ObjectIndex).xyz;
float2 DistanceFieldMAD = LoadObjectDistanceFieldMAD(ObjectIndex);
#if USE_OBJECT_COMPOSITING_TILE_CULLING
uint ObjectIndex = SharedCulledObjectList[ListObjectIndex];
#else
uint ObjectIndex = ListObjectIndex;
#endif
float3 ClampedSamplePosition = clamp(VolumePosition, -LocalPositionExtent, LocalPositionExtent);
float DistanceToClamped = length(VolumePosition - ClampedSamplePosition);
float3 LocalPositionExtent = LoadObjectLocalPositionExtent(ObjectIndex);
float4x4 WorldToVolume = LoadObjectWorldToVolume(ObjectIndex);
bool bGeneratedAsTwoSided;
float4 UVScaleAndVolumeScale = LoadObjectUVScale(ObjectIndex, bGeneratedAsTwoSided);
float3 VolumePosition = mul(float4(WorldPosition, 1), WorldToVolume).xyz;
float BoxDistance = ComputeDistanceFromBoxToPoint(-LocalPositionExtent, LocalPositionExtent, VolumePosition) * UVScaleAndVolumeScale.w;
float3 StepVolumeUV = DistanceFieldVolumePositionToUV(ClampedSamplePosition, UVScaleAndVolumeScale.xyz, UVAdd);
float DistanceToOccluder = (SampleMeshDistanceField(StepVolumeUV, DistanceFieldMAD).x + DistanceToClamped) * UVScaleAndVolumeScale.w;
BRANCH
if (BoxDistance < AOGlobalMaxSphereQueryRadius)
{
float3 UVAdd = LoadObjectUVAddAndSelfShadowBias(ObjectIndex).xyz;
float2 DistanceFieldMAD = LoadObjectDistanceFieldMAD(ObjectIndex);
MinDistance = min(MinDistance, DistanceToOccluder);
float3 ClampedSamplePosition = clamp(VolumePosition, -LocalPositionExtent, LocalPositionExtent);
float DistanceToClamped = length(VolumePosition - ClampedSamplePosition);
float3 StepVolumeUV = DistanceFieldVolumePositionToUV(ClampedSamplePosition, UVScaleAndVolumeScale.xyz, UVAdd);
float DistanceToOccluder = (SampleMeshDistanceField(StepVolumeUV, DistanceFieldMAD).x + DistanceToClamped) * UVScaleAndVolumeScale.w;
MinDistance = min(MinDistance, DistanceToOccluder);
}
}
}
Engine/Shaders/Private/GpuSkinVertexFactory.ush
+2 -1
View File
@@ -39,11 +39,13 @@ uint PerBoneMotionBlur;
#if GPUSKIN_APEX_CLOTH
/** Vertex buffer from which to read simulated positions of clothing. */
#if FEATURE_LEVEL >= FEATURE_LEVEL_SM4
Buffer<float2> ClothSimulVertsPositionsNormals;
Buffer<float2> PreviousClothSimulVertsPositionsNormals;
/** Transform to apply when positions/normals are loaded */
float4x4 ClothLocalToWorld;
float4x4 PreviousClothLocalToWorld;
#endif // #if FEATURE_LEVEL >= FEATURE_LEVEL_SM4
/** blend weight between simulated positions and original key-framed animation */
float ClothBlendWeight;
#endif// #if GPUSKIN_APEX_CLOTH
@@ -301,7 +303,6 @@ float3 GetClothSimulNormal(int Index, bool bPrevious)
}
#endif //FEATURE_LEVEL >= FEATURE_LEVEL_SM4
}
float3 ClothingPosition(FClothVertex Input, bool bPrevious)
Engine/Shaders/Private/HZB.usf
+1 -1
View File
@@ -181,7 +181,7 @@ void HZBBuildCS(
void HZBBuildPS(float4 SvPosition : SV_POSITION, out float4 OutColor : SV_Target0)
{
float2 BufferUV = SvPosition * DispatchThreadIdToBufferUV.xy + DispatchThreadIdToBufferUV.zw;
float2 BufferUV = SvPosition.xy * DispatchThreadIdToBufferUV.xy + DispatchThreadIdToBufferUV.zw;
float4 DeviceZ = Gather4(ParentTextureMip, ParentTextureMipSampler, BufferUV);
float FurthestDeviceZ = min(min(DeviceZ.x, DeviceZ.y), min(DeviceZ.z, DeviceZ.w));
Engine/Shaders/Private/HZBOcclusion.usf
+21 -17
View File
@@ -37,18 +37,20 @@ void HZBTestPS(float2 InUV : TEXCOORD0, out float4 OutColor : SV_Target0)
// Screen rect from bounds
float3 RectMin = float3( 1, 1, 1 );
float3 RectMax = float3( -1, -1, -1 );
UNROLL for( int i = 0; i < 8; i++ )
{
float3 PointSrc;
PointSrc.x = Bounds[ (i >> 0) & 1 ].x;
PointSrc.y = Bounds[ (i >> 1) & 1 ].y;
PointSrc.z = Bounds[ (i >> 2) & 1 ].z;
UNROLL for( int i = 0; i < 8; i++ )
{
float3 PointSrc;
PointSrc.x = Bounds[ (i >> 0) & 1 ].x;
PointSrc.y = Bounds[ (i >> 1) & 1 ].y;
PointSrc.z = Bounds[ (i >> 2) & 1 ].z;
float4 PointClip = mul( float4( PointSrc, 1 ), View.TranslatedWorldToClip );
float3 PointScreen = PointClip.xyz / PointClip.w;
float4 PointClip = mul( float4( PointSrc, 1 ), View.TranslatedWorldToClip );
float3 PointScreen = PointClip.xyz / PointClip.w;
RectMin = min( RectMin, PointScreen );
RectMax = max( RectMax, PointScreen );
RectMin = min( RectMin, PointScreen );
RectMax = max( RectMax, PointScreen );
}
}
#if 0
@@ -194,15 +196,17 @@ void HZBTestPS(float2 InUV : TEXCOORD0, out float4 OutColor : SV_Target0)
float2 Bias = HZBUvFactor.xy * Rect.xy;
float4 MinDepth = 1;
UNROLL for( int i = 0; i < 4; i++ )
{
// TODO could vectorize this
float4 Depth;
Depth.x = HZBTexture.SampleLevel( HZBSampler, float2( i, 0 ) * Scale + Bias, Level ).r;
Depth.y = HZBTexture.SampleLevel( HZBSampler, float2( i, 1 ) * Scale + Bias, Level ).r;
Depth.z = HZBTexture.SampleLevel( HZBSampler, float2( i, 2 ) * Scale + Bias, Level ).r;
Depth.w = HZBTexture.SampleLevel( HZBSampler, float2( i, 3 ) * Scale + Bias, Level ).r;
MinDepth = min( MinDepth, Depth );
UNROLL for( int i = 0; i < 4; i++ )
{
// TODO could vectorize this
float4 Depth;
Depth.x = HZBTexture.SampleLevel( HZBSampler, float2( i, 0 ) * Scale + Bias, Level ).r;
Depth.y = HZBTexture.SampleLevel( HZBSampler, float2( i, 1 ) * Scale + Bias, Level ).r;
Depth.z = HZBTexture.SampleLevel( HZBSampler, float2( i, 2 ) * Scale + Bias, Level ).r;
Depth.w = HZBTexture.SampleLevel( HZBSampler, float2( i, 3 ) * Scale + Bias, Level ).r;
MinDepth = min( MinDepth, Depth );
}
}
MinDepth.x = min( min(MinDepth.x, MinDepth.y), min(MinDepth.z, MinDepth.w) );
Engine/Shaders/Private/LPVPropagate.usf
+13 -11
View File
@@ -423,20 +423,22 @@ void CSPropagate(uint3 DTid : SV_DispatchThreadID, uint3 GTid : SV_GroupThreadID
ldsCacheOrigin -= int3(1,1,1);
#endif
for (int i=0;i<4;i++)
{
// Get the cell index,from 0-215 (6x6x6 block). There will be 4 indices per thread (with some duplication)
uint cellCacheIndex = laneID*4+i;
if ( cellCacheIndex < (6*6*6) )
for (int i=0;i<4;i++)
{
uint3 offset3;
offset3.x = ( cellCacheIndex % 6 );
offset3.y = ( cellCacheIndex / 6 ) % 6;
offset3.z = cellCacheIndex / ( 6 * 6 );
int3 cachedGridPos = ldsCacheOrigin + int3(offset3);
LPVCell CachedThreadCell = ReadLpvCell( GetGridAddress(cachedGridPos) );
// Get the cell index,from 0-215 (6x6x6 block). There will be 4 indices per thread (with some duplication)
uint cellCacheIndex = laneID*4+i;
if ( cellCacheIndex < (6*6*6) )
{
uint3 offset3;
offset3.x = ( cellCacheIndex % 6 );
offset3.y = ( cellCacheIndex / 6 ) % 6;
offset3.z = cellCacheIndex / ( 6 * 6 );
int3 cachedGridPos = ldsCacheOrigin + int3(offset3);
LPVCell CachedThreadCell = ReadLpvCell( GetGridAddress(cachedGridPos) );
LDSCachedLPV[cellCacheIndex] = PackLPVCell( CachedThreadCell );
LDSCachedLPV[cellCacheIndex] = PackLPVCell( CachedThreadCell );
}
}
}
Engine/Shaders/Private/PlanarReflectionShaders.usf
+10 -6
View File
@@ -86,16 +86,20 @@ void PrefilterPlanarReflectionPS(
float AccumulatedWeight = 0;
// Two texels at a time with bilinear
LOOP
for (float y = 0; y < KernelRadiusY; y += 2)
{
VerticalSample(BufferUV, y, KernelRadiusY, CenterDistanceToPlane, CenterReflectionDistance, false, AccumulatedColor, AccumulatedWeight);
LOOP
for (float y = 0; y < KernelRadiusY; y += 2)
{
VerticalSample(BufferUV, y, KernelRadiusY, CenterDistanceToPlane, CenterReflectionDistance, false, AccumulatedColor, AccumulatedWeight);
}
}
LOOP
for (float y = -2; y > -KernelRadiusY; y -= 2)
{
VerticalSample(BufferUV, y, KernelRadiusY, CenterDistanceToPlane, CenterReflectionDistance, true, AccumulatedColor, AccumulatedWeight);
LOOP
for (float y = -2; y > -KernelRadiusY; y -= 2)
{
VerticalSample(BufferUV, y, KernelRadiusY, CenterDistanceToPlane, CenterReflectionDistance, true, AccumulatedColor, AccumulatedWeight);
}
}
AccumulatedColor = AccumulatedColor / max(AccumulatedWeight, .0001f);
Engine/Shaders/Private/PostProcessTemporalAA.usf
+1 -1
View File
@@ -2078,7 +2078,7 @@ FTAAHistoryPayload TemporalAASample(uint2 GroupId, uint2 GroupThreadId, uint Gro
bool Dynamic7 = SceneVelocityBuffer.SampleLevel(SceneVelocityBufferSampler, InputParams.NearestBufferUV, 0, int2( 0, 1)).x > 0;
bool Dynamic = Dynamic1 || Dynamic3 || Dynamic4 || Dynamic5 || Dynamic7;
IgnoreHistory = IgnoreHistory || !Dynamic && History.Color.a > 0;
IgnoreHistory = IgnoreHistory || (!Dynamic && History.Color.a > 0);
}
#endif
Engine/Shaders/Private/RadixSortShaders.usf
+16 -12
View File
@@ -443,13 +443,15 @@ void RadixSort_Spine(
}
// Scan raking totals.
[unroll]
for ( uint RakingOffset = 1; RakingOffset < RAKING_THREADS_PER_DIGIT; RakingOffset <<= 1 )
{
// Acquire LocalRakingTotals.
PPS_BARRIER();
Total = Total + LocalRakingTotals[RakingIndex - RakingOffset];
LocalRakingTotals[RakingIndex] = Total;
[unroll]
for ( uint RakingOffset = 1; RakingOffset < RAKING_THREADS_PER_DIGIT; RakingOffset <<= 1 )
{
// Acquire LocalRakingTotals.
PPS_BARRIER();
Total = Total + LocalRakingTotals[RakingIndex - RakingOffset];
LocalRakingTotals[RakingIndex] = Total;
}
}
// Acquire LocalRakingTotals.
@@ -459,13 +461,15 @@ void RadixSort_Spine(
LocalTotals[ThreadId + DIGIT_SCAN_PADDING] = DigitSum;
// Scan local totals.
[unroll]
for ( uint RakingOffset = 1; RakingOffset < DIGIT_COUNT; RakingOffset <<= 1 )
{
// Acquire LocalTotals.
PPS_BARRIER();
DigitSum += LocalTotals[ThreadId + DIGIT_SCAN_PADDING - RakingOffset];
LocalTotals[ThreadId + DIGIT_SCAN_PADDING] = DigitSum;
[unroll]
for ( uint RakingOffset = 1; RakingOffset < DIGIT_COUNT; RakingOffset <<= 1 )
{
// Acquire LocalTotals.
PPS_BARRIER();
DigitSum += LocalTotals[ThreadId + DIGIT_SCAN_PADDING - RakingOffset];
LocalTotals[ThreadId + DIGIT_SCAN_PADDING] = DigitSum;
}
}
// Acquire LocalTotals.
Engine/Shaders/Private/ReflectionEnvironmentPixelShader.usf
+1 -1
View File
@@ -84,7 +84,7 @@ void RemapClearCoatDiffuseAndSpecularColor(FGBufferData GBuffer, float2 ScreenPo
//BaseColor += Dither / 255.f;
DiffuseColor = BaseColor - BaseColor * GBuffer.Metallic;
float3 Specular = lerp(1, RefractionScale, ClearCoat);
float Specular = lerp(1, RefractionScale, ClearCoat);
SpecularColor = ComputeF0(Specular, BaseColor, GBuffer.Metallic);
}
}
Engine/Shaders/Private/ShaderPrintDraw.usf
+14 -10
View File
@@ -97,22 +97,26 @@ void AddUIntSymbols(inout float2 ScreenPos, in uint Value)
uint SymbolCount = 1;
uint Divisor = 1;
uint TestValue = Value;
for (uint i = 0; i < MaxSymbols-1; ++i)
{
TestValue /= 10;
if (TestValue > 0)
for (uint i = 0; i < MaxSymbols-1; ++i)
{
SymbolCount ++;
Divisor *= 10;
TestValue /= 10;
if (TestValue > 0)
{
SymbolCount ++;
Divisor *= 10;
}
}
}
for (uint i = 0; i < SymbolCount; ++i)
{
uint Digit = Value / Divisor;
AddSymbol(ScreenPos, GetDecimalSymbol((int)Digit));
Value = Value - (Digit * Divisor);
Divisor /= 10;
for (uint i = 0; i < SymbolCount; ++i)
{
uint Digit = Value / Divisor;
AddSymbol(ScreenPos, GetDecimalSymbol((int)Digit));
Value = Value - (Digit * Divisor);
Divisor /= 10;
}
}
}
Engine/Shaders/Private/SlateVertexShader.usf
+12 -10
View File
@@ -65,7 +65,7 @@ VertexToPixelInterpolants Main(
{
UNROLL
for( int CoordinateIndex = 0; CoordinateIndex < NUM_MATERIAL_TEXCOORDS_VERTEX; CoordinateIndex++ )
for( uint CoordinateIndex = 0; CoordinateIndex < NUM_MATERIAL_TEXCOORDS_VERTEX; CoordinateIndex++ )
{
VertexParameters.TexCoords[CoordinateIndex] = UVArrays[CoordinateIndex];
}
@@ -76,17 +76,19 @@ VertexToPixelInterpolants Main(
float2 CustomizedUVs[NUM_SLATE_TEXCOORDS];
GetMaterialCustomizedUVs(VertexParameters, CustomizedUVs);
UNROLL
for( uint CoordinateIndex = 0; CoordinateIndex < NUM_SLATE_TEXCOORDS; CoordinateIndex++ )
{
FLATTEN
if (Mod(CoordinateIndex, 2))
UNROLL
for( uint CoordinateIndex = 0; CoordinateIndex < NUM_SLATE_TEXCOORDS; CoordinateIndex++ )
{
VOut.TextureCoordinates[CoordinateIndex / 2].zw = CustomizedUVs[CoordinateIndex];
}
else
{
VOut.TextureCoordinates[CoordinateIndex / 2].xy = CustomizedUVs[CoordinateIndex];
FLATTEN
if (Mod(CoordinateIndex, 2))
{
VOut.TextureCoordinates[CoordinateIndex / 2].zw = CustomizedUVs[CoordinateIndex];
}
else
{
VOut.TextureCoordinates[CoordinateIndex / 2].xy = CustomizedUVs[CoordinateIndex];
}
}
}

Some files were not shown because too many files have changed in this diff Show More