diff --git a/Common/GPU/Vulkan/VulkanRenderManager.h b/Common/GPU/Vulkan/VulkanRenderManager.h
index 95f9658510..2469ef37b2 100644
--- a/Common/GPU/Vulkan/VulkanRenderManager.h
+++ b/Common/GPU/Vulkan/VulkanRenderManager.h
@@ -199,7 +199,7 @@ static_assert(sizeof(PackedDescriptor::buffer) == 16, "PackedDescriptor should b
 struct VKRPipelineLayout {
 	~VKRPipelineLayout();
 
-	enum { MAX_DESC_SET_BINDINGS = 10 };
+	enum { MAX_DESC_SET_BINDINGS = 6 };
 	BindingType bindingTypes[MAX_DESC_SET_BINDINGS];
 
 	uint32_t bindingTypesCount = 0;
diff --git a/Core/Config.cpp b/Core/Config.cpp
index a29d4971b6..14c66cbbc1 100644
--- a/Core/Config.cpp
+++ b/Core/Config.cpp
@@ -774,7 +774,6 @@ static const ConfigSetting graphicsSettings[] = {
 
 	// Not really a graphics setting...
 	ConfigSetting("SplineBezierQuality", SETTING(g_Config, iSplineBezierQuality), 2, CfgFlag::PER_GAME | CfgFlag::REPORT),
-	ConfigSetting("HardwareTessellation", SETTING(g_Config, bHardwareTessellation), false, CfgFlag::PER_GAME | CfgFlag::REPORT),
 	ConfigSetting("TextureShader", SETTING(g_Config, sTextureShaderName), "Off", CfgFlag::PER_GAME),
 	ConfigSetting("ShaderChainRequires60FPS", SETTING(g_Config, bShaderChainRequires60FPS), false, CfgFlag::PER_GAME),
 
diff --git a/Core/Config.h b/Core/Config.h
index 6faa70303c..0fabad5be4 100644
--- a/Core/Config.h
+++ b/Core/Config.h
@@ -362,7 +362,6 @@ public:
 	int iBloomHack; //0 = off, 1 = safe, 2 = balanced, 3 = aggressive
 	int iSkipGPUReadbackMode;  // 0 = off, 1 = skip, 2 = to texture
 	int iSplineBezierQuality; // 0 = low , 1 = Intermediate , 2 = High
-	bool bHardwareTessellation;
 	bool bShaderCache;  // Hidden ini-only setting, useful for debugging shader compile times.
 	bool bUberShaderVertex;
 	int iDefaultTab;
diff --git a/GPU/Common/DrawEngineCommon.cpp b/GPU/Common/DrawEngineCommon.cpp
index 3f239901c0..f6779f3292 100644
--- a/GPU/Common/DrawEngineCommon.cpp
+++ b/GPU/Common/DrawEngineCommon.cpp
@@ -115,7 +115,6 @@ void DrawEngineCommon::NotifyConfigChanged() {
 	decoderMap_.Clear();
 
 	useHWTransform_ = g_Config.bHardwareTransform;
-	useHWTessellation_ = UpdateUseHWTessellation(g_Config.bHardwareTessellation);
 }
 
 void DrawEngineCommon::DispatchSubmitImm(GEPrimitiveType prim, TransformedVertex *buffer, int vertexCount, int cullMode, bool continuation) {
@@ -979,35 +978,6 @@ bool DrawEngineCommon::CanUseHardwareTransform(int prim) const {
 	return !gstate.isModeThrough() && prim != GE_PRIM_RECTANGLES && prim > GE_PRIM_LINE_STRIP;
 }
 
-bool DrawEngineCommon::CanUseHardwareTessellation(GEPatchPrimType prim) const {
-	if (useHWTessellation_) {
-		return CanUseHardwareTransform(PatchPrimToPrim(prim));
-	}
-	return false;
-}
-
-void TessellationDataTransfer::CopyControlPoints(float *pos, float *tex, float *col, int posStride, int texStride, int colStride, const SimpleVertex *const *points, int size, u32 vertType) {
-	bool hasColor = (vertType & GE_VTYPE_COL_MASK) != 0;
-	bool hasTexCoord = (vertType & GE_VTYPE_TC_MASK) != 0;
-
-	for (int i = 0; i < size; ++i) {
-		memcpy(pos, points[i]->pos.AsArray(), 3 * sizeof(float));
-		pos += posStride;
-	}
-	if (hasTexCoord) {
-		for (int i = 0; i < size; ++i) {
-			memcpy(tex, points[i]->uv, 2 * sizeof(float));
-			tex += texStride;
-		}
-	}
-	if (hasColor) {
-		for (int i = 0; i < size; ++i) {
-			memcpy(col, Vec4f::FromRGBA(points[i]->color_32).AsArray(), 4 * sizeof(float));
-			col += colStride;
-		}
-	}
-}
-
 bool DrawEngineCommon::DescribeCodePtr(const u8 *ptr, std::string &name) const {
 	if (!decJitCache_ || !decJitCache_->IsInSpace(ptr)) {
 		return false;
diff --git a/GPU/Common/DrawEngineCommon.h b/GPU/Common/DrawEngineCommon.h
index ee2bd9e8c8..bdeca83a83 100644
--- a/GPU/Common/DrawEngineCommon.h
+++ b/GPU/Common/DrawEngineCommon.h
@@ -59,13 +59,6 @@ enum FBOTexState {
 struct SimpleVertex;
 namespace Spline { struct Weight2D; }
 
-class TessellationDataTransfer {
-public:
-	virtual ~TessellationDataTransfer() {}
-	static void CopyControlPoints(float *pos, float *tex, float *col, int posStride, int texStride, int colStride, const SimpleVertex *const *points, int size, u32 vertType);
-	virtual void SendDataToShader(const SimpleVertex *const *points, int size_u, int size_v, u32 vertType, const Spline::Weight2D &weights) = 0;
-};
-
 // Culling plane, group of 8.
 struct alignas(16) Plane8 {
 	float x[8], y[8], z[8], w[8];
@@ -129,7 +122,6 @@ public:
 	static void ClearSplineBezierWeights();
 
 	bool CanUseHardwareTransform(int prim) const;
-	bool CanUseHardwareTessellation(GEPatchPrimType prim) const;
 
 	std::vector<std::string> DebugGetVertexLoaderIDs();
 	std::string DebugGetVertexLoaderString(std::string_view id, DebugShaderStringType stringType);
@@ -164,8 +156,6 @@ public:
 	void FlushQueuedDepth();
 
 protected:
-	virtual bool UpdateUseHWTessellation(bool enabled) const { return enabled; }
-
 	bool CheckClipFlags(bool useHwTransform) const;
 
 	void DecodeVerts(const VertexDecoder *dec, u8 *dest);
@@ -273,7 +263,6 @@ protected:
 	}
 
 	bool useHWTransform_ = false;
-	bool useHWTessellation_ = false;
 	// Used to prevent unnecessary flushing in softgpu.
 	bool flushOnParams_ = true;
 
@@ -346,9 +335,6 @@ protected:
 
 	ComputedPipelineState pipelineState_{};
 
-	// Hardware tessellation
-	TessellationDataTransfer *tessDataTransfer = nullptr;
-
 	GPUCommon *gpuCommon_ = nullptr;
 
 	// Software depth raster
diff --git a/GPU/Common/ShaderCommon.h b/GPU/Common/ShaderCommon.h
index 7cc848a0c1..3e5fe40d8a 100644
--- a/GPU/Common/ShaderCommon.h
+++ b/GPU/Common/ShaderCommon.h
@@ -68,10 +68,10 @@ enum : uint64_t {
 	DIRTY_SHADERBLEND = 1ULL << 17,  // Used only for in-shader blending.
 
 	DIRTY_UVSCALEOFFSET = 1ULL << 18,
-	// Free bit 19!
+	// Free uniform bit 19!
 
 	DIRTY_VIEWPORT_UNIFORMS = 1ULL << 20,
-	// Free bit 20!
+	// Free uniform bit 20!
 
 	DIRTY_WORLDMATRIX = 1ULL << 21,
 	DIRTY_VIEWMATRIX = 1ULL << 22,
@@ -85,9 +85,9 @@ enum : uint64_t {
 	DIRTY_BONEMATRIX6 = 1ULL << 30,
 	DIRTY_BONEMATRIX7 = 1ULL << 31,
 
-	DIRTY_BEZIERSPLINE = 1ULL << 32,
+	// Free uniform bit 32!,
 	DIRTY_TEXCLAMP = 1ULL << 33,
-	// Free bit 34!
+	// Free uniform  bit 34!
 
 	DIRTY_DEPAL = 1ULL << 35,
 	DIRTY_COLORWRITEMASK = 1ULL << 36,
diff --git a/GPU/Common/ShaderId.cpp b/GPU/Common/ShaderId.cpp
index 929a85146e..7affca5b5f 100644
--- a/GPU/Common/ShaderId.cpp
+++ b/GPU/Common/ShaderId.cpp
@@ -55,12 +55,6 @@ std::string VertexShaderDesc(const VShaderID &id) {
 	if (id.Bits(VS_BIT_WEIGHT_FMTSCALE, 2)) desc << "WScale " << id.Bits(VS_BIT_WEIGHT_FMTSCALE, 2) << " ";
 	if (id.Bit(VS_BIT_FLATSHADE)) desc << "Flat ";
 
-	if (id.Bit(VS_BIT_BEZIER)) desc << "Bezier ";
-	if (id.Bit(VS_BIT_SPLINE)) desc << "Spline ";
-	if (id.Bit(VS_BIT_HAS_COLOR_TESS)) desc << "TessC ";
-	if (id.Bit(VS_BIT_HAS_TEXCOORD_TESS)) desc << "TessT ";
-	if (id.Bit(VS_BIT_HAS_NORMAL_TESS)) desc << "TessN ";
-	if (id.Bit(VS_BIT_NORM_REVERSE_TESS)) desc << "TessRevN ";
 	if (id.Bit(VS_BIT_VERTEX_RANGE_CULLING)) desc << "RangeCull ";
 
 	if (id.Bit(VS_BIT_SIMPLE_STEREO)) desc << "SimpleStereo ";
@@ -70,7 +64,7 @@ std::string VertexShaderDesc(const VShaderID &id) {
 	return desc.str();
 }
 
-void ComputeVertexShaderID(VShaderID *id_out, u32 vertType, bool useHWTransform, bool useHWTessellation, bool weightsAsFloat, bool useSkinInDecode, ClipInfoFlags clipInfoFlags) {
+void ComputeVertexShaderID(VShaderID *id_out, u32 vertType, bool useHWTransform, bool weightsAsFloat, bool useSkinInDecode, ClipInfoFlags clipInfoFlags) {
 	const bool isModeThrough = (vertType & GE_VTYPE_THROUGH) != 0;
 	const bool isSoftwareFallback = !isModeThrough && !useHWTransform && g_Config.bHardwareTransform;
 	bool doTexture = gstate.isTextureMapEnabled() && !gstate.isModeClear();
@@ -151,18 +145,6 @@ void ComputeVertexShaderID(VShaderID *id_out, u32 vertType, bool useHWTransform,
 
 		id.SetBit(VS_BIT_NORM_REVERSE, gstate.areNormalsReversed());
 		id.SetBit(VS_BIT_HAS_TEXCOORD, vtypeHasTexcoord);
-
-		if (useHWTessellation) {
-			id.SetBit(VS_BIT_BEZIER, doBezier);
-			id.SetBit(VS_BIT_SPLINE, doSpline);
-			if (doBezier || doSpline) {
-				// These are the original vertType's values (normalized will always have colors, etc.)
-				id.SetBit(VS_BIT_HAS_COLOR_TESS, (gstate.vertType & GE_VTYPE_COL_MASK) != 0);
-				id.SetBit(VS_BIT_HAS_TEXCOORD_TESS, (gstate.vertType & GE_VTYPE_TC_MASK) != 0);
-				id.SetBit(VS_BIT_HAS_NORMAL_TESS, (gstate.vertType & GE_VTYPE_NRM_MASK) != 0 || gstate.isLightingEnabled());
-			}
-			id.SetBit(VS_BIT_NORM_REVERSE_TESS, gstate.isPatchNormalsReversed());
-		}
 	}
 
 	if (clipInfoFlags & ClipInfoFlags::DepthClampFragment) {
diff --git a/GPU/Common/ShaderId.h b/GPU/Common/ShaderId.h
index c5a97ccff0..e8815e1bf2 100644
--- a/GPU/Common/ShaderId.h
+++ b/GPU/Common/ShaderId.h
@@ -36,10 +36,7 @@ enum VShaderBit : uint8_t {
 	VS_BIT_HAS_NORMAL = 9,  // conditioned on hw transform
 	VS_BIT_NORM_REVERSE = 10,
 	VS_BIT_HAS_TEXCOORD = 11,
-	VS_BIT_HAS_COLOR_TESS = 12,  // 1 bit
-	VS_BIT_HAS_TEXCOORD_TESS = 13,  // 1 bit
-	VS_BIT_NORM_REVERSE_TESS = 14, // 1 bit
-	VS_BIT_HAS_NORMAL_TESS = 15, // 1 bit
+	// 4 bits free: 12-15
 	VS_BIT_UVGEN_MODE = 16,
 	VS_BIT_UVPROJ_MODE = 18,  // 2, can overlap with LS0
 	VS_BIT_LS0 = 18,  // 2
@@ -61,7 +58,7 @@ enum VShaderBit : uint8_t {
 	VS_BIT_LIGHT3_COMP = 44,  // 2 bits
 	VS_BIT_LIGHT3_TYPE = 46,  // 2 bits
 	VS_BIT_MATERIAL_UPDATE = 48,  // 3 bits
-	VS_BIT_SPLINE = 51, // 1 bit
+	// Bit 51 is free.
 	VS_BIT_LIGHT0_ENABLE = 52,
 	VS_BIT_LIGHT1_ENABLE = 53,
 	VS_BIT_LIGHT2_ENABLE = 54,
@@ -72,8 +69,7 @@ enum VShaderBit : uint8_t {
 	VS_BIT_FS_MINMAX_DISCARD = 59, // Do min/max and/or depth clamp in the fragment shader. It just means we need to forward Z and W to the fragment shader.
 	VS_BIT_FS_DEPTH_CLAMP = 60, // Do depth clamp in the fragment shader.
 	VS_BIT_FLATSHADE = 62, // 1 bit
-	VS_BIT_BEZIER = 63, // 1 bit
-	// No more free
+	// Bit 63 is free.
 };
 
 static inline VShaderBit operator +(VShaderBit bit, int i) {
@@ -259,7 +255,7 @@ namespace Draw {
 class Bugs;
 }
 
-void ComputeVertexShaderID(VShaderID *id, u32 vertType, bool useHWTransform, bool useHWTessellation, bool weightsAsFloat, bool useSkinInDecode, ClipInfoFlags clipInfoFlags);
+void ComputeVertexShaderID(VShaderID *id, u32 vertType, bool useHWTransform, bool weightsAsFloat, bool useSkinInDecode, ClipInfoFlags clipInfoFlags);
 // Generates a compact string that describes the shader. Useful in a list to get an overview
 // of the current flora of shaders.
 std::string VertexShaderDesc(const VShaderID &id);
diff --git a/GPU/Common/ShaderUniforms.cpp b/GPU/Common/ShaderUniforms.cpp
index ddb505ad60..89a61fed3d 100644
--- a/GPU/Common/ShaderUniforms.cpp
+++ b/GPU/Common/ShaderUniforms.cpp
@@ -148,10 +148,6 @@ void BaseUpdateUniforms(UB_VS_FS_Base *ub, uint64_t dirtyUniforms, bool useBuffe
 		UpdateUVScaleOff(gstate, ub->uvScaleOffset);
 	}
 
-	if (dirtyUniforms & DIRTY_BEZIERSPLINE) {
-		ub->spline_counts = gstate_c.spline_num_points_u;
-	}
-
 	if (dirtyUniforms & DIRTY_DEPAL) {
 		ub->depal_mask_shift_off_fmt = PackDepalBits();
 	}
diff --git a/GPU/Common/ShaderUniforms.h b/GPU/Common/ShaderUniforms.h
index f6e072a919..9d8ded4038 100644
--- a/GPU/Common/ShaderUniforms.h
+++ b/GPU/Common/ShaderUniforms.h
@@ -13,7 +13,7 @@ enum : uint64_t {
 	DIRTY_WORLDMATRIX | DIRTY_PROJTHROUGHMATRIX | DIRTY_VIEWMATRIX | DIRTY_TEXMATRIX | DIRTY_ALPHACOLORREF |
 	DIRTY_PROJMATRIX | DIRTY_FOGCOLOR | DIRTY_FOGCOEF | DIRTY_TEXENV | DIRTY_TEX_ALPHA_MUL | DIRTY_STENCILREPLACEVALUE |
 	DIRTY_ALPHACOLORMASK | DIRTY_SHADERBLEND | DIRTY_COLORWRITEMASK | DIRTY_UVSCALEOFFSET | DIRTY_TEXCLAMP | DIRTY_MATAMBIENTALPHA |
-	DIRTY_BEZIERSPLINE | DIRTY_DEPAL | DIRTY_VIEWPORT_UNIFORMS | DIRTY_RASTER_OFFSET,
+	DIRTY_DEPAL | DIRTY_VIEWPORT_UNIFORMS | DIRTY_RASTER_OFFSET,
 	DIRTY_LIGHT_UNIFORMS =
 	DIRTY_LIGHT_CONTROL | DIRTY_LIGHT0 | DIRTY_LIGHT1 | DIRTY_LIGHT2 | DIRTY_LIGHT3 |
 	DIRTY_MATDIFFUSE | DIRTY_MATSPECULAR | DIRTY_MATEMISSIVE | DIRTY_AMBIENT,
@@ -32,7 +32,7 @@ struct alignas(16) UB_VS_FS_Base {
 	float rasterOffset[2]; float minZmaxZ[2];
 	float uvScaleOffset[4];
 	float matAmbient[4];
-	uint32_t spline_counts; uint32_t depal_mask_shift_off_fmt;  // 4 params packed into one.
+	uint32_t padding3; uint32_t depal_mask_shift_off_fmt;  // 4 params packed into one.
 	uint32_t colorWriteMask; float mipBias;
 	// Fragment data
 	float texNoAlpha; float texMul; float padding4[2];  // this vec4 will hold ubershader stuff. We won't use integer flags in the fragment shader.
diff --git a/GPU/Common/SplineCommon.cpp b/GPU/Common/SplineCommon.cpp
index 90ee730ddf..6f5e4d95d7 100644
--- a/GPU/Common/SplineCommon.cpp
+++ b/GPU/Common/SplineCommon.cpp
@@ -435,45 +435,6 @@ void SoftwareTessellation(OutputBuffers &output, const Surface &surface, u32 ori
 template void SoftwareTessellation<BezierSurface>(OutputBuffers &output, const BezierSurface &surface, u32 origVertType, const ControlPoints &points);
 template void SoftwareTessellation<SplineSurface>(OutputBuffers &output, const SplineSurface &surface, u32 origVertType, const ControlPoints &points);
 
-template<class Surface>
-static void HardwareTessellation(OutputBuffers &output, const Surface &surface, u32 origVertType,
-	const SimpleVertex *const *points, TessellationDataTransfer *tessDataTransfer) {
-	using WeightType = typename Surface::WeightType;
-	u32 key_u = WeightType::ToKey(surface.tess_u, surface.num_points_u, surface.type_u);
-	u32 key_v = WeightType::ToKey(surface.tess_v, surface.num_points_v, surface.type_v);
-	Weight2D weights(WeightType::weightsCache, key_u, key_v);
-	weights.size_u = WeightType::CalcSize(surface.tess_u, surface.num_points_u);
-	weights.size_v = WeightType::CalcSize(surface.tess_v, surface.num_points_v);
-	tessDataTransfer->SendDataToShader(points, surface.num_points_u, surface.num_points_v, origVertType, weights);
-
-	// Generating simple input vertices for the spline-computing vertex shader.
-	float inv_u = 1.0f / (float)surface.tess_u;
-	float inv_v = 1.0f / (float)surface.tess_v;
-	for (int patch_u = 0; patch_u < surface.num_patches_u; ++patch_u) {
-		const int start_u = surface.GetTessStart(patch_u);
-		for (int patch_v = 0; patch_v < surface.num_patches_v; ++patch_v) {
-			const int start_v = surface.GetTessStart(patch_v);
-			for (int tile_u = start_u; tile_u <= surface.tess_u; ++tile_u) {
-				const int index_u = surface.GetIndexU(patch_u, tile_u);
-				for (int tile_v = start_v; tile_v <= surface.tess_v; ++tile_v) {
-					const int index_v = surface.GetIndexV(patch_v, tile_v);
-					SimpleVertex &vert = output.vertices[surface.GetIndex(index_u, index_v, patch_u, patch_v)];
-					// Index for the weights
-					vert.pos.x = index_u;
-					vert.pos.y = index_v;
-					// For texcoord generation
-					vert.nrm.x = patch_u + (float)tile_u * inv_u;
-					vert.nrm.y = patch_v + (float)tile_v * inv_v;
-					// Patch position
-					vert.pos.z = patch_u;
-					vert.nrm.z = patch_v;
-				}
-			}
-		}
-	}
-	surface.BuildIndex(output.indices, output.count);
-}
-
 } // namespace Spline
 
 using namespace Spline;
@@ -551,15 +512,11 @@ void DrawEngineCommon::SubmitCurve(const void *control_points, const void *indic
 
 	surface.Init(maxVerts);
 
-	if (CanUseHardwareTessellation(surface.primType)) {
-		HardwareTessellation(output, surface, origVertType, points, tessDataTransfer);
-	} else {
-		ControlPoints cpoints(points, num_points, managedBuf);
-		if (cpoints.IsValid())
-			SoftwareTessellation(output, surface, origVertType, cpoints);
-		else
-			ERROR_LOG(Log::G3D, "Failed to allocate space for control point values, skipping curve draw");
-	}
+	ControlPoints cpoints(points, num_points, managedBuf);
+	if (cpoints.IsValid())
+		SoftwareTessellation(output, surface, origVertType, cpoints);
+	else
+		ERROR_LOG(Log::G3D, "Failed to allocate space for control point values, skipping curve draw");
 
 	u32 vertTypeWithIndex16 = (vertType & ~GE_VTYPE_IDX_MASK) | GE_VTYPE_IDX_16BIT;
 
diff --git a/GPU/Common/VertexShaderGenerator.cpp b/GPU/Common/VertexShaderGenerator.cpp
index b66ac96252..ce8581f64a 100644
--- a/GPU/Common/VertexShaderGenerator.cpp
+++ b/GPU/Common/VertexShaderGenerator.cpp
@@ -202,25 +202,6 @@ bool GenerateVertexShader(const VShaderID &id, char *buffer, const ShaderLanguag
 		return false;
 	}
 
-	// Apparently we don't support bezier/spline together with bones.
-	bool doBezier = id.Bit(VS_BIT_BEZIER) && !enableBones && useHWTransform;
-	bool doSpline = id.Bit(VS_BIT_SPLINE) && !enableBones && useHWTransform;
-	if (doBezier || doSpline) {
-		if (!hasNormal) {
-			// Bad usage.
-			*errorString = "Invalid flags - tess requires normal.";
-			return false;
-		}
-		if (compat.texelFetch == nullptr) {
-			*errorString = "Tess not supported on this shader language version";
-			return false;
-		}
-	}
-	bool hasColorTess = id.Bit(VS_BIT_HAS_COLOR_TESS);
-	bool hasTexcoordTess = id.Bit(VS_BIT_HAS_TEXCOORD_TESS);
-	bool hasNormalTess = id.Bit(VS_BIT_HAS_NORMAL_TESS);
-	bool flipNormalTess = id.Bit(VS_BIT_NORM_REVERSE_TESS);
-
 	// Should we do the min/max discard in the shader and/or or use depth clamping?
 	// In both cases we need to just forward
 	const bool fsMinmaxDiscard = id.Bit(VS_BIT_FS_MINMAX_DISCARD);
@@ -316,9 +297,6 @@ bool GenerateVertexShader(const VShaderID &id, char *buffer, const ShaderLanguag
 		// And the "varyings".
 		if (useHWTransform) {
 			WRITE(p, "struct VS_IN {                              \n");
-			if ((doSpline || doBezier) && compat.shaderLanguage == HLSL_D3D11) {
-				WRITE(p, "  uint instanceId : SV_InstanceID;\n");
-			}
 			if (enableBones) {
 				WRITE(p, "  %s", boneWeightAttrDeclHLSL[numBoneWeights]);
 			}
@@ -538,163 +516,6 @@ bool GenerateVertexShader(const VShaderID &id, char *buffer, const ShaderLanguag
 		}
 	}
 
-	// Hardware tessellation
-	if (doBezier || doSpline) {
-		*uniformMask |= DIRTY_BEZIERSPLINE;
-
-		if (compat.shaderLanguage == GLSL_VULKAN) {
-			WRITE(p, "struct TessData {\n");
-			WRITE(p, "  vec4 pos;\n");
-			WRITE(p, "  vec4 tex;\n");
-			WRITE(p, "  vec4 col;\n");
-			WRITE(p, "};\n");
-			WRITE(p, "layout (std430, set = 0, binding = %d) readonly buffer s_tess_data {\n", DRAW_BINDING_TESS_STORAGE_BUF);
-			WRITE(p, "  TessData tess_data[];\n");
-			WRITE(p, "};\n");
-
-			WRITE(p, "struct TessWeight {\n");
-			WRITE(p, "  vec4 basis;\n");
-			WRITE(p, "  vec4 deriv;\n");
-			WRITE(p, "};\n");
-			WRITE(p, "layout (std430, set = 0, binding = %d) readonly buffer s_tess_weights_u {\n", DRAW_BINDING_TESS_STORAGE_BUF_WU);
-			WRITE(p, "  TessWeight tess_weights_u[];\n");
-			WRITE(p, "};\n");
-			WRITE(p, "layout (std430, set = 0, binding = %d) readonly buffer s_tess_weights_v {\n", DRAW_BINDING_TESS_STORAGE_BUF_WV);
-			WRITE(p, "  TessWeight tess_weights_v[];\n");
-			WRITE(p, "};\n");
-		} else if (ShaderLanguageIsOpenGL(compat.shaderLanguage)) {
-			WRITE(p, "uniform sampler2D u_tess_points;\n"); // Control Points
-			WRITE(p, "uniform sampler2D u_tess_weights_u;\n");
-			WRITE(p, "uniform sampler2D u_tess_weights_v;\n");
-
-			WRITE(p, "uniform int u_spline_counts;\n");
-		} else if (compat.shaderLanguage == HLSL_D3D11) {
-			WRITE(p, "struct TessData {\n");
-			WRITE(p, "  vec3 pos; float pad1;\n");
-			WRITE(p, "  vec2 tex; vec2 pad2;\n");
-			WRITE(p, "  vec4 col;\n");
-			WRITE(p, "};\n");
-			WRITE(p, "StructuredBuffer<TessData> tess_data : register(t0);\n");
-
-			WRITE(p, "struct TessWeight {\n");
-			WRITE(p, "  vec4 basis;\n");
-			WRITE(p, "  vec4 deriv;\n");
-			WRITE(p, "};\n");
-			WRITE(p, "StructuredBuffer<TessWeight> tess_weights_u : register(t1);\n");
-			WRITE(p, "StructuredBuffer<TessWeight> tess_weights_v : register(t2);\n");
-		}
-
-		const char *init[3] = { "0.0, 0.0", "0.0, 0.0, 0.0", "0.0, 0.0, 0.0, 0.0" };
-		for (int i = 2; i <= 4; i++) {
-			// Define 3 types vec2, vec3, vec4
-			WRITE(p, "vec%d tess_sample(in vec%d points[16], mat4 weights) {\n", i, i);
-			WRITE(p, "  vec%d pos = vec%d(%s);\n", i, i, init[i - 2]);
-			for (int v = 0; v < 4; ++v) {
-				for (int u = 0; u < 4; ++u) {
-					WRITE(p, "  pos += weights[%i][%i] * points[%i];\n", v, u, v * 4 + u);
-				}
-			}
-			WRITE(p, "  return pos;\n");
-			WRITE(p, "}\n");
-		}
-
-		if (ShaderLanguageIsOpenGL(compat.shaderLanguage) && compat.glslVersionNumber < 130) { // For glsl version 1.10
-			WRITE(p, "mat4 outerProduct(vec4 u, vec4 v) {\n");
-			WRITE(p, "  return mat4(u * v[0], u * v[1], u * v[2], u * v[3]);\n");
-			WRITE(p, "}\n");
-		} else if (compat.shaderLanguage == HLSL_D3D11) {
-			WRITE(p, "mat4 outerProduct(vec4 u, vec4 v) {\n");
-			WRITE(p, "  return mul((float4x1)v, (float1x4)u);\n");
-			WRITE(p, "}\n");
-		}
-
-		WRITE(p, "struct Tess {\n");
-		WRITE(p, "  vec3 pos;\n");
-		WRITE(p, "  vec2 tex;\n");
-		WRITE(p, "  vec4 col;\n");
-		if (hasNormalTess)
-			WRITE(p, "  vec3 nrm;\n");
-		WRITE(p, "};\n");
-
-		if (compat.shaderLanguage == HLSL_D3D11) {
-			WRITE(p, "void tessellate(in VS_IN In, out Tess tess) {\n");
-			WRITE(p, "  vec3 position = In.position;\n");
-			WRITE(p, "  vec3 normal = In.normal;\n");
-		} else {
-			WRITE(p, "void tessellate(out Tess tess) {\n");
-		}
-		WRITE(p, "  ivec2 point_pos = ivec2(position.z, normal.z)%s;\n", doBezier ? " * 3" : "");
-		WRITE(p, "  ivec2 weight_idx = ivec2(position.xy);\n");
-
-		// Load 4x4 control points
-		WRITE(p, "  vec3 _pos[16];\n");
-		WRITE(p, "  vec2 _tex[16];\n");
-		WRITE(p, "  vec4 _col[16];\n");
-		if (compat.coefsFromBuffers) {
-			WRITE(p, "  int index;\n");
-			for (int i = 0; i < 4; i++) {
-				for (int j = 0; j < 4; j++) {
-					WRITE(p, "  index = (%i + point_pos.y) * int(u_spline_counts) + (%i + point_pos.x);\n", i, j);
-					WRITE(p, "  _pos[%i] = tess_data[index].pos.xyz;\n", i * 4 + j);
-					if (hasTexcoordTess)
-						WRITE(p, "  _tex[%i] = tess_data[index].tex.xy;\n", i * 4 + j);
-					if (hasColorTess)
-						WRITE(p, "  _col[%i] = tess_data[index].col;\n", i * 4 + j);
-				}
-			}
-
-			// Basis polynomials as weight coefficients
-			WRITE(p, "  vec4 basis_u = tess_weights_u[weight_idx.x].basis;\n");
-			WRITE(p, "  vec4 basis_v = tess_weights_v[weight_idx.y].basis;\n");
-			WRITE(p, "  mat4 basis = outerProduct(basis_u, basis_v);\n");
-		} else {
-			WRITE(p, "  int index_u, index_v;\n");
-			for (int i = 0; i < 4; i++) {
-				for (int j = 0; j < 4; j++) {
-					WRITE(p, "  index_u = (%i + point_pos.x);\n", j);
-					WRITE(p, "  index_v = (%i + point_pos.y);\n", i);
-					WRITE(p, "  _pos[%i] = %s(u_tess_points, ivec2(index_u, index_v), 0).xyz;\n", i * 4 + j, compat.texelFetch);
-					if (hasTexcoordTess)
-						WRITE(p, "  _tex[%i] = %s(u_tess_points, ivec2(index_u + u_spline_counts, index_v), 0).xy;\n", i * 4 + j, compat.texelFetch);
-					if (hasColorTess)
-						WRITE(p, "  _col[%i] = %s(u_tess_points, ivec2(index_u + u_spline_counts * 2, index_v), 0).rgba;\n", i * 4 + j, compat.texelFetch);
-				}
-			}
-
-			// Basis polynomials as weight coefficients
-			WRITE(p, "  vec4 basis_u = %s(u_tess_weights_u, %s, 0);\n", compat.texelFetch, "ivec2(weight_idx.x * 2, 0)");
-			WRITE(p, "  vec4 basis_v = %s(u_tess_weights_v, %s, 0);\n", compat.texelFetch, "ivec2(weight_idx.y * 2, 0)");
-			WRITE(p, "  mat4 basis = outerProduct(basis_u, basis_v);\n");
-		}
-
-		// Tessellate
-		WRITE(p, "  tess.pos = tess_sample(_pos, basis);\n");
-		if (hasTexcoordTess)
-			WRITE(p, "  tess.tex = tess_sample(_tex, basis);\n");
-		else
-			WRITE(p, "  tess.tex = normal.xy;\n");
-		if (hasColorTess)
-			WRITE(p, "  tess.col = tess_sample(_col, basis);\n");
-		else
-			WRITE(p, "  tess.col = u_matambientalpha;\n");
-		if (hasNormalTess) {
-			if (compat.coefsFromBuffers) {
-				// Derivatives as weight coefficients
-				WRITE(p, "  vec4 deriv_u = tess_weights_u[weight_idx.x].deriv;\n");
-				WRITE(p, "  vec4 deriv_v = tess_weights_v[weight_idx.y].deriv;\n");
-			} else {
-				// Derivatives as weight coefficients
-				WRITE(p, "  vec4 deriv_u = %s(u_tess_weights_u, %s, 0);\n", compat.texelFetch, "ivec2(weight_idx.x * 2 + 1, 0)");
-				WRITE(p, "  vec4 deriv_v = %s(u_tess_weights_v, %s, 0);\n", compat.texelFetch, "ivec2(weight_idx.y * 2 + 1, 0)");
-			}
-
-			WRITE(p, "  vec3 du = tess_sample(_pos, outerProduct(deriv_u, basis_v));\n");
-			WRITE(p, "  vec3 dv = tess_sample(_pos, outerProduct(basis_u, deriv_v));\n");
-			WRITE(p, "  tess.nrm = normalize(cross(du, dv));\n");
-		}
-		WRITE(p, "}\n");
-	}
-
 	if (useHWTransform) {
 		WRITE(p, "vec3 normalizeOr001(vec3 v) {\n");
 		WRITE(p, "   return length(v) == 0.0 ? vec3(0.0, 0.0, 1.0) : normalize(v);\n");
@@ -766,28 +587,13 @@ bool GenerateVertexShader(const VShaderID &id, char *buffer, const ShaderLanguag
 	} else {
 		// Step 1: World Transform / Skinning
 		if (!enableBones) {
-			if (doBezier || doSpline) {
-				// Hardware tessellation
-				WRITE(p, "  Tess tess;\n");
-				if (compat.shaderLanguage == HLSL_D3D11) {
-					WRITE(p, "  tessellate(In, tess);\n");
-				} else {
-					WRITE(p, "  tessellate(tess);\n");
-				}
-
-				WRITE(p, "  vec3 worldpos = mul(vec4(tess.pos.xyz, 1.0), u_world).xyz;\n");
-				if (hasNormalTess) {
-					WRITE(p, "  mediump vec3 worldnormal = normalizeOr001(mul(vec4(%stess.nrm, 0.0), u_world).xyz);\n", flipNormalTess ? "-" : "");
-				} else {
-					WRITE(p, "  mediump vec3 worldnormal = normalizeOr001(mul(vec4(0.0, 0.0, %s1.0, 0.0), u_world).xyz);\n", flipNormalTess ? "-" : "");
-				}
+		
+			// No skinning, just standard T&L.
+			WRITE(p, "  vec3 worldpos = mul(vec4(position, 1.0), u_world).xyz;\n");
+			if (hasNormal) {
+				WRITE(p, "  mediump vec3 worldnormal = normalizeOr001(mul(vec4(%snormal, 0.0), u_world).xyz);\n", flipNormal ? "-" : "");
 			} else {
-				// No skinning, just standard T&L.
-				WRITE(p, "  vec3 worldpos = mul(vec4(position, 1.0), u_world).xyz;\n");
-				if (hasNormal)
-					WRITE(p, "  mediump vec3 worldnormal = normalizeOr001(mul(vec4(%snormal, 0.0), u_world).xyz);\n", flipNormal ? "-" : "");
-				else
-					WRITE(p, "  mediump vec3 worldnormal = normalizeOr001(mul(vec4(0.0, 0.0, %s1.0, 0.0), u_world).xyz);\n", flipNormal ? "-" : "");
+				WRITE(p, "  mediump vec3 worldnormal = normalizeOr001(mul(vec4(0.0, 0.0, %s1.0, 0.0), u_world).xyz);\n", flipNormal ? "-" : "");
 			}
 		} else {
 			static const char * const rescale[4] = {"", " * 1.9921875", " * 1.999969482421875", ""}; // 2*127.5f/128.f, 2*32767.5f/32768.f, 1.0f};
@@ -878,11 +684,6 @@ bool GenerateVertexShader(const VShaderID &id, char *buffer, const ShaderLanguag
 		// TODO: Declare variables for dots for shade mapping if needed.
 
 		const char *srcCol = "color0";
-		if (doBezier || doSpline) {
-			// TODO: Probably, should use hasColorTess but FF4 has a problem with drawing the background.
-			srcCol = "tess.col";
-		}
-
 		if (lightUberShader && hasColor) {
 			p.F("  vec4 ambientColor = ((u_lightControl & (1u << 0x14u)) != 0x0u) ? %s : u_matambientalpha;\n", srcCol);
 			if (enableLighting) {
@@ -1115,10 +916,7 @@ bool GenerateVertexShader(const VShaderID &id, char *buffer, const ShaderLanguag
 		} else {
 			// Lighting doesn't affect color.
 			if (hasColor) {
-				if (doBezier || doSpline)
-					WRITE(p, "  %sv_color0 = tess.col;\n", compat.vsOutPrefix);
-				else
-					WRITE(p, "  %sv_color0 = color0;\n", compat.vsOutPrefix);
+				WRITE(p, "  %sv_color0 = color0;\n", compat.vsOutPrefix);
 			} else {
 				WRITE(p, "  %sv_color0 = u_matambientalpha;\n", compat.vsOutPrefix);
 				if (bugs.Has(Draw::Bugs::MALI_CONSTANT_LOAD_BUG) && g_Config.bVendorBugChecksEnabled) {
@@ -1139,19 +937,13 @@ bool GenerateVertexShader(const VShaderID &id, char *buffer, const ShaderLanguag
 			case GE_TEXMAP_UNKNOWN: // Not sure what this is, but Riviera uses it.  Treating as coords works.
 				if (scaleUV) {
 					if (hasTexcoord) {
-						if (doBezier || doSpline)
-							WRITE(p, "  %sv_texcoord = vec3(tess.tex.xy * u_uvscaleoffset.xy + u_uvscaleoffset.zw, 0.0);\n", compat.vsOutPrefix);
-						else
-							WRITE(p, "  %sv_texcoord = vec3(texcoord.xy * u_uvscaleoffset.xy, 0.0);\n", compat.vsOutPrefix);
+						WRITE(p, "  %sv_texcoord = vec3(texcoord.xy * u_uvscaleoffset.xy, 0.0);\n", compat.vsOutPrefix);
 					} else {
 						WRITE(p, "  %sv_texcoord = splat3(0.0);\n", compat.vsOutPrefix);
 					}
 				} else {
 					if (hasTexcoord) {
-						if (doBezier || doSpline)
-							WRITE(p, "  %sv_texcoord = vec3(tess.tex.xy * u_uvscaleoffset.xy + u_uvscaleoffset.zw, 0.0);\n", compat.vsOutPrefix);
-						else
-							WRITE(p, "  %sv_texcoord = vec3(texcoord.xy * u_uvscaleoffset.xy + u_uvscaleoffset.zw, 0.0);\n", compat.vsOutPrefix);
+						WRITE(p, "  %sv_texcoord = vec3(texcoord.xy * u_uvscaleoffset.xy + u_uvscaleoffset.zw, 0.0);\n", compat.vsOutPrefix);
 					} else {
 						WRITE(p, "  %sv_texcoord = vec3(u_uvscaleoffset.zw, 0.0);\n", compat.vsOutPrefix);
 					}
@@ -1163,39 +955,28 @@ bool GenerateVertexShader(const VShaderID &id, char *buffer, const ShaderLanguag
 					std::string temp_tc;
 					switch (uvProjMode) {
 					case GE_PROJMAP_POSITION:  // Use model space XYZ as source
-						if (doBezier || doSpline)
-							temp_tc = "vec4(tess.pos, 1.0)";
-						else
-							temp_tc = "vec4(position, 1.0)";
+						temp_tc = "vec4(position, 1.0)";
 						break;
 					case GE_PROJMAP_UV:  // Use unscaled UV as source
-						{
-							// prescale is false here.
-							if (hasTexcoord) {
-								if (doBezier || doSpline)
-									temp_tc = "vec4(tess.tex.xy, 0.0, 1.0)";
-								else
-									temp_tc = "vec4(texcoord.xy, 0.0, 1.0)";
-							} else {
-								temp_tc = "vec4(0.0, 0.0, 0.0, 1.0)";
-							}
+						if (hasTexcoord) {
+							temp_tc = "vec4(texcoord.xy, 0.0, 1.0)";
+						} else {
+							temp_tc = "vec4(0.0, 0.0, 0.0, 1.0)";
 						}
 						break;
 					case GE_PROJMAP_NORMALIZED_NORMAL:  // Use normalized transformed normal as source
-						if ((doBezier || doSpline) && hasNormalTess)
-							temp_tc = StringFromFormat("length(tess.nrm) == 0.0 ? vec4(0.0, 0.0, 0.0, 1.0) : vec4(normalize(%stess.nrm), 1.0)", flipNormalTess ? "-" : "");
-						else if (hasNormal)
+						if (hasNormal) {
 							temp_tc = StringFromFormat("length(normal) == 0.0 ? vec4(0.0, 0.0, 0.0, 1.0) : vec4(normalize(%snormal), 1.0)", flipNormal ? "-" : "");
-						else
+						} else {
 							temp_tc = "vec4(0.0, 0.0, 1.0, 1.0)";
+						}
 						break;
 					case GE_PROJMAP_NORMAL:  // Use non-normalized transformed normal as source
-						if ((doBezier || doSpline) && hasNormalTess)
-							temp_tc = flipNormalTess ? "vec4(-tess.nrm, 1.0)" : "vec4(tess.nrm, 1.0)";
-						else if (hasNormal)
+						if (hasNormal) {
 							temp_tc = flipNormal ? "vec4(-normal, 1.0)" : "vec4(normal, 1.0)";
-						else
+						} else {
 							temp_tc = "vec4(0.0, 0.0, 1.0, 1.0)";
+						}
 						break;
 					}
 					// Transform by texture matrix. XYZ as we are doing projection mapping.
diff --git a/GPU/D3D11/DrawEngineD3D11.cpp b/GPU/D3D11/DrawEngineD3D11.cpp
index c1a67cdc12..c5ce1084e1 100644
--- a/GPU/D3D11/DrawEngineD3D11.cpp
+++ b/GPU/D3D11/DrawEngineD3D11.cpp
@@ -85,9 +85,6 @@ void DrawEngineD3D11::InitDeviceObjects() {
 	pushVerts_ = new PushBufferD3D11(device_, VERTEX_PUSH_SIZE, D3D11_BIND_VERTEX_BUFFER);
 	pushInds_ = new PushBufferD3D11(device_, INDEX_PUSH_SIZE, D3D11_BIND_INDEX_BUFFER);
 
-	tessDataTransferD3D11 = new TessellationDataTransferD3D11(context_, device_);
-	tessDataTransfer = tessDataTransferD3D11;
-
 	draw_->SetInvalidationCallback(std::bind(&DrawEngineD3D11::Invalidate, this, std::placeholders::_1));
 }
 
@@ -97,9 +94,6 @@ void DrawEngineD3D11::DestroyDeviceObjects() {
 	}
 
 	ClearInputLayoutMap();
-	delete tessDataTransferD3D11;
-	tessDataTransferD3D11 = nullptr;
-	tessDataTransfer = nullptr;
 	delete pushVerts_;
 	delete pushInds_;
 	pushVerts_ = nullptr;
@@ -347,7 +341,7 @@ void DrawEngineD3D11::Flush() {
 
 		D3D11VertexShader *vshader;
 		D3D11FragmentShader *fshader;
-		shaderManager_->GetShaders(prim, dec_->VertexType(), &vshader, &fshader, pipelineState_, useHWTransform, useHWTessellation_, decOptions_.expandAllWeightsToFloat, applySkinInDecode_, clipInfoFlags_);
+		shaderManager_->GetShaders(prim, dec_->VertexType(), &vshader, &fshader, pipelineState_, useHWTransform, decOptions_.expandAllWeightsToFloat, applySkinInDecode_, clipInfoFlags_);
 		ID3D11InputLayout *inputLayout;
 		SetupDecFmtForDraw(vshader, dec_->GetDecVtxFmt(), dec_->VertexType(), &inputLayout);
 		context_->PSSetShader(fshader->GetShader(), nullptr, 0);
@@ -459,7 +453,7 @@ void DrawEngineD3D11::Flush() {
 		if (action == SW_DRAW_INDEXED) {
 			D3D11VertexShader *vshader;
 			D3D11FragmentShader *fshader;
-			shaderManager_->GetShaders(prim, swDec->VertexType(), &vshader, &fshader, pipelineState_, false, false, decOptions_.expandAllWeightsToFloat, true, clipInfoFlags_);
+			shaderManager_->GetShaders(prim, swDec->VertexType(), &vshader, &fshader, pipelineState_, false, decOptions_.expandAllWeightsToFloat, true, clipInfoFlags_);
 			context_->PSSetShader(fshader->GetShader(), nullptr, 0);
 			context_->VSSetShader(vshader->GetShader(), nullptr, 0);
 			shaderManager_->UpdateUniforms(framebufferManager_->UseBufferedRendering());
@@ -519,95 +513,3 @@ void DrawEngineD3D11::Flush() {
 	framebufferManager_->SetColorUpdated(gstate_c.skipDrawReason);
 	gpuCommon_->NotifyFlush();
 }
-
-TessellationDataTransferD3D11::TessellationDataTransferD3D11(ID3D11DeviceContext *context, ID3D11Device *device)
-	: context_(context), device_(device) {
-	desc.Usage = D3D11_USAGE_DYNAMIC;
-	desc.BindFlags = D3D11_BIND_SHADER_RESOURCE;
-	desc.CPUAccessFlags = D3D11_CPU_ACCESS_WRITE;
-	desc.MiscFlags = D3D11_RESOURCE_MISC_BUFFER_STRUCTURED;
-}
-
-TessellationDataTransferD3D11::~TessellationDataTransferD3D11() {
-}
-
-void TessellationDataTransferD3D11::SendDataToShader(const SimpleVertex *const *points, int size_u, int size_v, u32 vertType, const Spline::Weight2D &weights) {
-	struct TessData {
-		float pos[3]; float pad1;
-		float uv[2]; float pad2[2];
-		float color[4];
-	};
-
-	int size = size_u * size_v;
-
-	if (prevSize < size || !buf[0]) {
-		prevSize = size;
-		buf[0].Reset();
-		view[0].Reset();
-
-		desc.ByteWidth = size * sizeof(TessData);
-		desc.StructureByteStride = sizeof(TessData);
-		device_->CreateBuffer(&desc, nullptr, &buf[0]);
-		if (buf[0])
-			device_->CreateShaderResourceView(buf[0].Get(), nullptr, &view[0]);
-		if (!buf[0] || !view[0])
-			return;
-		context_->VSSetShaderResources(0, 1, view[0].GetAddressOf());
-	}
-	D3D11_MAPPED_SUBRESOURCE map{};
-	HRESULT hr = context_->Map(buf[0].Get(), 0, D3D11_MAP_WRITE_DISCARD, 0, &map);
-	if (FAILED(hr))
-		return;
-	uint8_t *data = (uint8_t *)map.pData;
-
-	float *pos = (float *)(data);
-	float *tex = (float *)(data + offsetof(TessData, uv));
-	float *col = (float *)(data + offsetof(TessData, color));
-	int stride = sizeof(TessData) / sizeof(float);
-
-	CopyControlPoints(pos, tex, col, stride, stride, stride, points, size, vertType);
-
-	context_->Unmap(buf[0].Get(), 0);
-
-	using Spline::Weight;
-
-	// Weights U
-	if (prevSizeWU < weights.size_u || !buf[1]) {
-		prevSizeWU = weights.size_u;
-		buf[1].Reset();
-		view[1].Reset();
-
-		desc.ByteWidth = weights.size_u * sizeof(Weight);
-		desc.StructureByteStride = sizeof(Weight);
-		device_->CreateBuffer(&desc, nullptr, &buf[1]);
-		if (buf[1])
-			device_->CreateShaderResourceView(buf[1].Get(), nullptr, &view[1]);
-		if (!buf[1] || !view[1])
-			return;
-		context_->VSSetShaderResources(1, 1, view[1].GetAddressOf());
-	}
-	hr = context_->Map(buf[1].Get(), 0, D3D11_MAP_WRITE_DISCARD, 0, &map);
-	if (SUCCEEDED(hr))
-		memcpy(map.pData, weights.u, weights.size_u * sizeof(Weight));
-	context_->Unmap(buf[1].Get(), 0);
-
-	// Weights V
-	if (prevSizeWV < weights.size_v) {
-		prevSizeWV = weights.size_v;
-		buf[2].Reset();
-		view[2].Reset();
-
-		desc.ByteWidth = weights.size_v * sizeof(Weight);
-		desc.StructureByteStride = sizeof(Weight);
-		device_->CreateBuffer(&desc, nullptr, &buf[2]);
-		if (buf[2])
-			device_->CreateShaderResourceView(buf[2].Get(), nullptr, &view[2]);
-		if (!buf[2] || !view[2])
-			return;
-		context_->VSSetShaderResources(2, 1, view[2].GetAddressOf());
-	}
-	hr = context_->Map(buf[2].Get(), 0, D3D11_MAP_WRITE_DISCARD, 0, &map);
-	if (SUCCEEDED(hr))
-		memcpy(map.pData, weights.v, weights.size_v * sizeof(Weight));
-	context_->Unmap(buf[2].Get(), 0);
-}
diff --git a/GPU/D3D11/DrawEngineD3D11.h b/GPU/D3D11/DrawEngineD3D11.h
index e7c567e370..3944f72a36 100644
--- a/GPU/D3D11/DrawEngineD3D11.h
+++ b/GPU/D3D11/DrawEngineD3D11.h
@@ -37,22 +37,6 @@ class ShaderManagerD3D11;
 class TextureCacheD3D11;
 class FramebufferManagerD3D11;
 
-class TessellationDataTransferD3D11 : public TessellationDataTransfer {
-private:
-	ID3D11DeviceContext *context_;
-	ID3D11Device *device_;
-	Microsoft::WRL::ComPtr<ID3D11Buffer> buf[3]{};
-	Microsoft::WRL::ComPtr<ID3D11ShaderResourceView> view[3]{};
-	D3D11_BUFFER_DESC desc{};
-	int prevSize = 0;
-	int prevSizeWU = 0, prevSizeWV = 0;
-public:
-	TessellationDataTransferD3D11(ID3D11DeviceContext *context, ID3D11Device *device);
-	~TessellationDataTransferD3D11();
-	// Send spline/bezier's control points and weights to vertex shader through structured shader buffer.
-	void SendDataToShader(const SimpleVertex *const *points, int size_u, int size_v, u32 vertType, const Spline::Weight2D &weights) override;
-};
-
 // Handles transform, lighting and drawing.
 class DrawEngineD3D11 : public DrawEngineCommon {
 public:
@@ -140,8 +124,5 @@ private:
 	D3D11StateKeys keys_{};
 	D3D11DynamicState dynState_{};
 
-	// Hardware tessellation
-	TessellationDataTransferD3D11 *tessDataTransferD3D11 = nullptr;
-
 	int lastRenderStepId_ = -1;
 };
diff --git a/GPU/D3D11/GPU_D3D11.cpp b/GPU/D3D11/GPU_D3D11.cpp
index 77e580cf86..260ffde59b 100644
--- a/GPU/D3D11/GPU_D3D11.cpp
+++ b/GPU/D3D11/GPU_D3D11.cpp
@@ -83,7 +83,6 @@ u32 GPU_D3D11::CheckGPUFeatures() const {
 	u32 features = GPUCommonHW::CheckGPUFeatures();
 
 	features |= GPU_USE_TEXTURE_FLOAT;
-	features |= GPU_USE_INSTANCE_RENDERING;
 	features |= GPU_USE_TEXTURE_LOD_CONTROL;
 
 	uint32_t fmt4444 = draw_->GetDataFormatSupport(Draw::DataFormat::A4R4G4B4_UNORM_PACK16);
diff --git a/GPU/D3D11/ShaderManagerD3D11.cpp b/GPU/D3D11/ShaderManagerD3D11.cpp
index 52b006add9..337c8a49d7 100644
--- a/GPU/D3D11/ShaderManagerD3D11.cpp
+++ b/GPU/D3D11/ShaderManagerD3D11.cpp
@@ -186,13 +186,13 @@ void ShaderManagerD3D11::BindUniforms() {
 	context_->PSSetConstantBuffers(0, 1, ps_cbs);
 }
 
-void ShaderManagerD3D11::GetShaders(int prim, u32 vertexType, D3D11VertexShader **vshader, D3D11FragmentShader **fshader, const ComputedPipelineState &pipelineState, bool useHWTransform, bool useHWTessellation, bool weightsAsFloat, bool useSkinInDecode, ClipInfoFlags clipInfoFlags) {
+void ShaderManagerD3D11::GetShaders(int prim, u32 vertexType, D3D11VertexShader **vshader, D3D11FragmentShader **fshader, const ComputedPipelineState &pipelineState, bool useHWTransform, bool weightsAsFloat, bool useSkinInDecode, ClipInfoFlags clipInfoFlags) {
 	VShaderID VSID;
 	FShaderID FSID;
 
 	if (gstate_c.IsDirty(DIRTY_VERTEXSHADER_STATE)) {
 		gstate_c.Clean(DIRTY_VERTEXSHADER_STATE);
-		ComputeVertexShaderID(&VSID, vertexType, useHWTransform, useHWTessellation, weightsAsFloat, useSkinInDecode, clipInfoFlags);
+		ComputeVertexShaderID(&VSID, vertexType, useHWTransform, weightsAsFloat, useSkinInDecode, clipInfoFlags);
 	} else {
 		VSID = lastVSID_;
 	}
diff --git a/GPU/D3D11/ShaderManagerD3D11.h b/GPU/D3D11/ShaderManagerD3D11.h
index 0935e66f44..da4f3d8899 100644
--- a/GPU/D3D11/ShaderManagerD3D11.h
+++ b/GPU/D3D11/ShaderManagerD3D11.h
@@ -88,7 +88,7 @@ public:
 	ShaderManagerD3D11(Draw::DrawContext *draw, ID3D11Device *device, ID3D11DeviceContext *context, D3D_FEATURE_LEVEL featureLevel);
 	~ShaderManagerD3D11();
 
-	void GetShaders(int prim, u32 vertexType, D3D11VertexShader **vshader, D3D11FragmentShader **fshader, const ComputedPipelineState &pipelineState, bool useHWTransform, bool useHWTessellation, bool weightsAsFloat, bool useSkinInDecode, ClipInfoFlags clipInfoFlags);
+	void GetShaders(int prim, u32 vertexType, D3D11VertexShader **vshader, D3D11FragmentShader **fshader, const ComputedPipelineState &pipelineState, bool useHWTransform, bool weightsAsFloat, bool useSkinInDecode, ClipInfoFlags clipInfoFlags);
 	void ClearShaders() override;
 	void DirtyLastShader() override;
 
diff --git a/GPU/GLES/DrawEngineGLES.cpp b/GPU/GLES/DrawEngineGLES.cpp
index 5bd549cf07..97bf678982 100644
--- a/GPU/GLES/DrawEngineGLES.cpp
+++ b/GPU/GLES/DrawEngineGLES.cpp
@@ -57,15 +57,10 @@ DrawEngineGLES::DrawEngineGLES(Draw::DrawContext *draw) : inputLayoutMap_(16), d
 	decOptions_.expand8BitNormalsToFloat = false;
 
 	InitDeviceObjects();
-
-	tessDataTransferGLES = new TessellationDataTransferGLES(render_);
-	tessDataTransfer = tessDataTransferGLES;
 }
 
 DrawEngineGLES::~DrawEngineGLES() {
 	DestroyDeviceObjects();
-
-	delete tessDataTransferGLES;
 }
 
 void DrawEngineGLES::DeviceLost() {
@@ -147,7 +142,6 @@ void DrawEngineGLES::EndFrame() {
 	FrameData &frameData = frameData_[render_->GetCurFrame()];
 	frameData.pushIndex->End();
 	frameData.pushVertex->End();
-	tessDataTransferGLES->EndFrame();
 }
 
 struct GlTypeInfo {
@@ -275,7 +269,7 @@ void DrawEngineGLES::Flush() {
 		lastUseHwTransform_ = useHWTransform;
 	}
 
-	Shader *vshader = shaderManager_->ApplyVertexShader(useHWTransform, useHWTessellation_, dec_->VertexType(), decOptions_.expandAllWeightsToFloat, applySkinInDecode_ || !useHWTransform, clipInfoFlags_, &vsid);
+	Shader *vshader = shaderManager_->ApplyVertexShader(useHWTransform, dec_->VertexType(), decOptions_.expandAllWeightsToFloat, applySkinInDecode_ || !useHWTransform, clipInfoFlags_, &vsid);
 
 	useHWTransform = vshader->UseHWTransform();  // In case shader compilation failed and it fell back. However, this can no longer really happen... Need to fix this.
 
@@ -458,80 +452,3 @@ bail:
 	framebufferManager_->SetColorUpdated(gstate_c.skipDrawReason);
 	gpuCommon_->NotifyFlush();
 }
-
-// TODO: Refactor this to a single USE flag.
-bool DrawEngineGLES::SupportsHWTessellation() {
-	bool hasTexelFetch = gl_extensions.GLES3 || (!gl_extensions.IsGLES && gl_extensions.VersionGEThan(3, 3, 0)) || gl_extensions.EXT_gpu_shader4;
-	return hasTexelFetch && gstate_c.UseAll(GPU_USE_VERTEX_TEXTURE_FETCH | GPU_USE_TEXTURE_FLOAT | GPU_USE_INSTANCE_RENDERING);
-}
-
-bool DrawEngineGLES::UpdateUseHWTessellation(bool enable) const {
-	return enable && SupportsHWTessellation();
-}
-
-void TessellationDataTransferGLES::SendDataToShader(const SimpleVertex *const *points, int size_u, int size_v, u32 vertType, const Spline::Weight2D &weights) {
-	bool hasColor = (vertType & GE_VTYPE_COL_MASK) != 0;
-	bool hasTexCoord = (vertType & GE_VTYPE_TC_MASK) != 0;
-
-	int size = size_u * size_v;
-	float *pos = new float[size * 4];
-	float *tex = hasTexCoord ? new float[size * 4] : nullptr;
-	float *col = hasColor ? new float[size * 4] : nullptr;
-	int stride = 4;
-
-	CopyControlPoints(pos, tex, col, stride, stride, stride, points, size, vertType);
-	// Removed the 1D texture support, it's unlikely to be relevant for performance.
-	// Control Points
-	if (prevSizeU < size_u || prevSizeV < size_v) {
-		prevSizeU = size_u;
-		prevSizeV = size_v;
-		if (data_tex[0])
-			renderManager_->DeleteTexture(data_tex[0]);
-		data_tex[0] = renderManager_->CreateTexture(GL_TEXTURE_2D, size_u * 3, size_v, 1, 1);
-		renderManager_->TextureImage(data_tex[0], 0, size_u * 3, size_v, 1, Draw::DataFormat::R32G32B32A32_FLOAT, nullptr, GLRAllocType::NONE, false);
-		renderManager_->FinalizeTexture(data_tex[0], 0, false);
-	}
-	renderManager_->BindTexture(TEX_SLOT_SPLINE_POINTS, data_tex[0]);
-	// Position
-	renderManager_->TextureSubImage(TEX_SLOT_SPLINE_POINTS, data_tex[0], 0, 0, 0, size_u, size_v, Draw::DataFormat::R32G32B32A32_FLOAT, (u8 *)pos, GLRAllocType::NEW);
-	// Texcoord
-	if (hasTexCoord)
-		renderManager_->TextureSubImage(TEX_SLOT_SPLINE_POINTS, data_tex[0], 0, size_u, 0, size_u, size_v, Draw::DataFormat::R32G32B32A32_FLOAT, (u8 *)tex, GLRAllocType::NEW);
-	// Color
-	if (hasColor)
-		renderManager_->TextureSubImage(TEX_SLOT_SPLINE_POINTS, data_tex[0], 0, size_u * 2, 0, size_u, size_v, Draw::DataFormat::R32G32B32A32_FLOAT, (u8 *)col, GLRAllocType::NEW);
-
-	// Weight U
-	if (prevSizeWU < weights.size_u) {
-		prevSizeWU = weights.size_u;
-		if (data_tex[1])
-			renderManager_->DeleteTexture(data_tex[1]);
-		data_tex[1] = renderManager_->CreateTexture(GL_TEXTURE_2D, weights.size_u * 2, 1, 1, 1);
-		renderManager_->TextureImage(data_tex[1], 0, weights.size_u * 2, 1, 1, Draw::DataFormat::R32G32B32A32_FLOAT, nullptr, GLRAllocType::NONE, false);
-		renderManager_->FinalizeTexture(data_tex[1], 0, false);
-	}
-	renderManager_->BindTexture(TEX_SLOT_SPLINE_WEIGHTS_U, data_tex[1]);
-	renderManager_->TextureSubImage(TEX_SLOT_SPLINE_WEIGHTS_U, data_tex[1], 0, 0, 0, weights.size_u * 2, 1, Draw::DataFormat::R32G32B32A32_FLOAT, (u8 *)weights.u, GLRAllocType::NONE);
-
-	// Weight V
-	if (prevSizeWV < weights.size_v) {
-		prevSizeWV = weights.size_v;
-		if (data_tex[2])
-			renderManager_->DeleteTexture(data_tex[2]);
-		data_tex[2] = renderManager_->CreateTexture(GL_TEXTURE_2D, weights.size_v * 2, 1, 1, 1);
-		renderManager_->TextureImage(data_tex[2], 0, weights.size_v * 2, 1, 1, Draw::DataFormat::R32G32B32A32_FLOAT, nullptr, GLRAllocType::NONE, false);
-		renderManager_->FinalizeTexture(data_tex[2], 0, false);
-	}
-	renderManager_->BindTexture(TEX_SLOT_SPLINE_WEIGHTS_V, data_tex[2]);
-	renderManager_->TextureSubImage(TEX_SLOT_SPLINE_WEIGHTS_V, data_tex[2], 0, 0, 0, weights.size_v * 2, 1, Draw::DataFormat::R32G32B32A32_FLOAT, (u8 *)weights.v, GLRAllocType::NONE);
-}
-
-void TessellationDataTransferGLES::EndFrame() {
-	for (int i = 0; i < 3; i++) {
-		if (data_tex[i]) {
-			renderManager_->DeleteTexture(data_tex[i]);
-			data_tex[i] = nullptr;
-		}
-	}
-	prevSizeU = prevSizeV = prevSizeWU = prevSizeWV = 0;
-}
diff --git a/GPU/GLES/DrawEngineGLES.h b/GPU/GLES/DrawEngineGLES.h
index 052b2bc92f..9220f5da25 100644
--- a/GPU/GLES/DrawEngineGLES.h
+++ b/GPU/GLES/DrawEngineGLES.h
@@ -40,23 +40,6 @@ struct TransformedVertex;
 
 struct DecVtxFormat;
 
-class TessellationDataTransferGLES : public TessellationDataTransfer {
-private:
-	GLRTexture *data_tex[3]{};
-	int prevSizeU = 0, prevSizeV = 0;
-	int prevSizeWU = 0, prevSizeWV = 0;
-	GLRenderManager *renderManager_;
-public:
-	TessellationDataTransferGLES(GLRenderManager *renderManager)
-			: renderManager_(renderManager) { }
-	~TessellationDataTransferGLES() {
-		EndFrame();
-	}
-	// Send spline/bezier's control points and weights to vertex shader through floating point texture.
-	void SendDataToShader(const SimpleVertex *const *points, int size_u, int size_v, u32 vertType, const Spline::Weight2D &weights) override;
-	void EndFrame();  // Queues textures for deletion.
-};
-
 // Handles transform, lighting and drawing.
 class DrawEngineGLES : public DrawEngineCommon {
 public:
@@ -90,11 +73,6 @@ public:
 
 	void ClearInputLayoutMap();
 
-	static bool SupportsHWTessellation() ;
-
-protected:
-	bool UpdateUseHWTessellation(bool enable) const override;
-
 private:
 	void Invalidate(InvalidationCallbackFlags flags);
 
@@ -131,7 +109,4 @@ private:
 
 	int bufferDecimationCounter_ = 0;
 	int lastRenderStepId_ = -1;
-
-	// Hardware tessellation
-	TessellationDataTransferGLES *tessDataTransferGLES;
 };
diff --git a/GPU/GLES/GPU_GLES.cpp b/GPU/GLES/GPU_GLES.cpp
index ea1f569fb7..96234e3a45 100644
--- a/GPU/GLES/GPU_GLES.cpp
+++ b/GPU/GLES/GPU_GLES.cpp
@@ -107,13 +107,6 @@ GPU_GLES::GPU_GLES(GraphicsContext *gfxCtx, Draw::DrawContext *draw)
 			INFO_LOG(Log::G3D, "Shader cache disabled. Not loading.");
 		}
 	}
-
-	if (g_Config.bHardwareTessellation) {
-		// Log information that we disable hardware tessellation if device is unsupported.
-		if (!drawEngine_.SupportsHWTessellation()) {
-			ERROR_LOG(Log::G3D, "Hardware Tessellation is unsupported, falling back to software tessellation");
-		}
-	}
 }
 
 GPU_GLES::~GPU_GLES() {
@@ -141,16 +134,6 @@ u32 GPU_GLES::CheckGPUFeatures() const {
 	if (gl_extensions.GLES3 || !gl_extensions.IsGLES)
 		features |= GPU_USE_TEXTURE_LOD_CONTROL;
 
-	bool canUseInstanceID = gl_extensions.EXT_draw_instanced || gl_extensions.ARB_draw_instanced;
-	bool canDefInstanceID = gl_extensions.IsGLES || gl_extensions.EXT_gpu_shader4 || gl_extensions.VersionGEThan(3, 1);
-	bool instanceRendering = gl_extensions.GLES3 || (canUseInstanceID && canDefInstanceID);
-	if (instanceRendering)
-		features |= GPU_USE_INSTANCE_RENDERING;
-
-	int maxVertexTextureImageUnits = gl_extensions.maxVertexTextureUnits;
-	if (maxVertexTextureImageUnits >= 3) // At least 3 for hardware tessellation
-		features |= GPU_USE_VERTEX_TEXTURE_FETCH;
-
 	if (gl_extensions.ARB_texture_float || gl_extensions.OES_texture_float)
 		features |= GPU_USE_TEXTURE_FLOAT;
 
diff --git a/GPU/GLES/ShaderManagerGLES.cpp b/GPU/GLES/ShaderManagerGLES.cpp
index b7752b447e..34cf666006 100644
--- a/GPU/GLES/ShaderManagerGLES.cpp
+++ b/GPU/GLES/ShaderManagerGLES.cpp
@@ -607,12 +607,6 @@ void LinkedShader::UpdateUniforms(const ShaderID &vsid, const ShaderLanguageDesc
 			if (u_lightspecular[i] != -1) SetColorUniform3(render_, &u_lightspecular[i], gstate.lcolor[i * 3 + 2]);
 		}
 	}
-
-	if (dirty & DIRTY_BEZIERSPLINE) {
-		if (u_spline_counts != -1) {
-			render_->SetUniformI1(&u_spline_counts, gstate_c.spline_num_points_u);
-		}
-	}
 }
 
 static constexpr size_t CODE_BUFFER_SIZE = 32768;
@@ -707,10 +701,10 @@ Shader *ShaderManagerGLES::CompileVertexShader(VShaderID VSID) {
 	return new Shader(render_, codeBuffer_, desc, params);
 }
 
-Shader *ShaderManagerGLES::ApplyVertexShader(bool useHWTransform, bool useHWTessellation, u32 vertexType, bool weightsAsFloat, bool useSkinInDecode, ClipInfoFlags clipInfoFlags, VShaderID *VSID) {
+Shader *ShaderManagerGLES::ApplyVertexShader(bool useHWTransform, u32 vertexType, bool weightsAsFloat, bool useSkinInDecode, ClipInfoFlags clipInfoFlags, VShaderID *VSID) {
 	if (gstate_c.IsDirty(DIRTY_VERTEXSHADER_STATE)) {
 		gstate_c.Clean(DIRTY_VERTEXSHADER_STATE);
-		ComputeVertexShaderID(VSID, vertexType, useHWTransform, useHWTessellation, weightsAsFloat, useSkinInDecode, clipInfoFlags);
+		ComputeVertexShaderID(VSID, vertexType, useHWTransform, weightsAsFloat, useSkinInDecode, clipInfoFlags);
 	} else {
 		*VSID = lastVSID_;
 	}
@@ -743,7 +737,7 @@ Shader *ShaderManagerGLES::ApplyVertexShader(bool useHWTransform, bool useHWTess
 
 		// Can still work with software transform.
 		VShaderID vsidTemp;
-		ComputeVertexShaderID(&vsidTemp, vertexType, false, false, weightsAsFloat, true, clipInfoFlags);
+		ComputeVertexShaderID(&vsidTemp, vertexType, false, weightsAsFloat, true, clipInfoFlags);
 		vs = CompileVertexShader(vsidTemp);
 	}
 
diff --git a/GPU/GLES/ShaderManagerGLES.h b/GPU/GLES/ShaderManagerGLES.h
index f3bd32ea29..db35142c8d 100644
--- a/GPU/GLES/ShaderManagerGLES.h
+++ b/GPU/GLES/ShaderManagerGLES.h
@@ -172,7 +172,7 @@ public:
 
 	// This is the old ApplyShader split into two parts, because of annoying information dependencies.
 	// If you call ApplyVertexShader, you MUST call ApplyFragmentShader soon afterwards.
-	Shader *ApplyVertexShader(bool useHWTransform, bool useHWTessellation, u32 vertexType, bool weightsAsFloat, bool useSkinInDecode, ClipInfoFlags clipInfoFlags, VShaderID *VSID);
+	Shader *ApplyVertexShader(bool useHWTransform, u32 vertexType, bool weightsAsFloat, bool useSkinInDecode, ClipInfoFlags clipInfoFlags, VShaderID *VSID);
 	LinkedShader *ApplyFragmentShader(VShaderID VSID, Shader *vs, const ComputedPipelineState &pipelineState, ClipInfoFlags clipInfoFlags);
 
 	void DeviceLost() override;
diff --git a/GPU/GPUCommonHW.cpp b/GPU/GPUCommonHW.cpp
index 84323c66a1..d8e91c112c 100644
--- a/GPU/GPUCommonHW.cpp
+++ b/GPU/GPUCommonHW.cpp
@@ -1328,15 +1328,7 @@ void GPUCommonHW::Execute_Bezier(u32 op, u32 diff) {
 
 	// We need to dirty UVSCALEOFFSET here because we look at the submit type when setting that uniform.
 	gstate_c.Dirty(DIRTY_RASTER_STATE | DIRTY_VERTEXSHADER_STATE | DIRTY_UVSCALEOFFSET);
-	if (drawEngineCommon_->CanUseHardwareTessellation(surface.primType)) {
-		gstate_c.submitType = SubmitType::HW_BEZIER;
-		if (gstate_c.spline_num_points_u != surface.num_points_u) {
-			gstate_c.Dirty(DIRTY_BEZIERSPLINE);
-			gstate_c.spline_num_points_u = surface.num_points_u;
-		}
-	} else {
-		gstate_c.submitType = SubmitType::BEZIER;
-	}
+	gstate_c.submitType = SubmitType::BEZIER;
 
 	int bytesRead = 0;
 	gstate_c.UpdateUVScaleOffset();
@@ -1408,15 +1400,7 @@ void GPUCommonHW::Execute_Spline(u32 op, u32 diff) {
 
 	// We need to dirty UVSCALEOFFSET here because we look at the submit type when setting that uniform.
 	gstate_c.Dirty(DIRTY_RASTER_STATE | DIRTY_VERTEXSHADER_STATE | DIRTY_UVSCALEOFFSET);
-	if (drawEngineCommon_->CanUseHardwareTessellation(surface.primType)) {
-		gstate_c.submitType = SubmitType::HW_SPLINE;
-		if (gstate_c.spline_num_points_u != surface.num_points_u) {
-			gstate_c.Dirty(DIRTY_BEZIERSPLINE);
-			gstate_c.spline_num_points_u = surface.num_points_u;
-		}
-	} else {
-		gstate_c.submitType = SubmitType::SPLINE;
-	}
+	gstate_c.submitType = SubmitType::SPLINE;
 
 	int bytesRead = 0;
 	gstate_c.UpdateUVScaleOffset();
diff --git a/GPU/GPUState.cpp b/GPU/GPUState.cpp
index 157784df9a..59bd57d16c 100644
--- a/GPU/GPUState.cpp
+++ b/GPU/GPUState.cpp
@@ -375,29 +375,29 @@ static constexpr const char * g_gpuUseFlagNames[32] = {
 	"GPU_USE_VS_RANGE_CULLING",
 	"GPU_USE_BLEND_MINMAX",
 	"GPU_USE_LOGIC_OP",
-	"GPU_USE_FRAGMENT_UBERSHADER",
-	"GPU_USE_TEXTURE_NPOT",
+	"N/A",
+	"N/A",
 	"GPU_USE_ANISOTROPY",
 	"GPU_USE_CLEAR_RAM_HACK",
-	"GPU_USE_INSTANCE_RENDERING",
-	"GPU_USE_VERTEX_TEXTURE_FETCH",
+	"N/A",
+	"N/A",
 	"GPU_USE_TEXTURE_FLOAT",
 	"GPU_USE_16BIT_FORMATS",
 	"GPU_USE_DEPTH_CLAMP",
 	"GPU_USE_TEXTURE_LOD_CONTROL",
 	"GPU_USE_DEPTH_TEXTURE",
-	"GPU_USE_ACCURATE_DEPTH",
-	"GPU_USE_GS_CULLING",
 	"N/A",
+	"N/A",
+	"GPU_USE_FRAMEBUFFER_ARRAYS",
 	"GPU_USE_FRAMEBUFFER_FETCH",
-	"GPU_SCALE_DEPTH_FROM_24BIT_TO_16BIT",
+	"N/A",
 	"GPU_ROUND_FRAGMENT_DEPTH_TO_16BIT",
 	"GPU_ROUND_DEPTH_TO_16BIT",
 	"GPU_USE_CLIP_DISTANCE",
 	"GPU_USE_CULL_DISTANCE",
-	"N/A", // bit 26
-	"N/A", // bit 27
-	"N/A", // bit 28
+	"GPU_USE_SHADER_BLENDING", // bit 26
+	"GPU_USE_NONBUFFERED_FLIP", // bit 27
+	"GPU_USE_PRE_ROTATION", // bit 28
 	"GPU_USE_VIRTUAL_REALITY",
 	"GPU_USE_SINGLE_PASS_STEREO",
 	"GPU_USE_SIMPLE_STEREO_PERSPECTIVE",
diff --git a/GPU/GPUState.h b/GPU/GPUState.h
index e8f5e7d824..70023b638b 100644
--- a/GPU/GPUState.h
+++ b/GPU/GPUState.h
@@ -475,8 +475,8 @@ enum : u32 {
 	// Free bits: 6-7
 	GPU_USE_ANISOTROPY = FLAG_BIT(8),
 	GPU_USE_CLEAR_RAM_HACK = FLAG_BIT(9),
-	GPU_USE_INSTANCE_RENDERING = FLAG_BIT(10),
-	GPU_USE_VERTEX_TEXTURE_FETCH = FLAG_BIT(11),
+	// Free bit: 10
+	// Free bit: 11
 	GPU_USE_TEXTURE_FLOAT = FLAG_BIT(12),
 	GPU_USE_16BIT_FORMATS = FLAG_BIT(13),
 	GPU_USE_DEPTH_CLAMP = FLAG_BIT(14),
diff --git a/GPU/Software/TransformUnit.h b/GPU/Software/TransformUnit.h
index c6ce6722b7..17fc2053f4 100644
--- a/GPU/Software/TransformUnit.h
+++ b/GPU/Software/TransformUnit.h
@@ -187,7 +187,4 @@ public:
 		FreeAlignedMemory(p);
 	}
 #endif
-
-protected:
-	bool UpdateUseHWTessellation(bool enable) const override { return false; }
 };
diff --git a/GPU/Vulkan/DrawEngineVulkan.cpp b/GPU/Vulkan/DrawEngineVulkan.cpp
index 9e74f06b9e..1f53b4b91c 100644
--- a/GPU/Vulkan/DrawEngineVulkan.cpp
+++ b/GPU/Vulkan/DrawEngineVulkan.cpp
@@ -64,9 +64,6 @@ void DrawEngineVulkan::InitDeviceObjects() {
 		BindingType::UNIFORM_BUFFER_DYNAMIC_ALL,  // uniforms
 		BindingType::UNIFORM_BUFFER_DYNAMIC_VERTEX,  // lights
 		BindingType::UNIFORM_BUFFER_DYNAMIC_VERTEX,  // bones
-		BindingType::STORAGE_BUFFER_VERTEX,  // tess
-		BindingType::STORAGE_BUFFER_VERTEX,
-		BindingType::STORAGE_BUFFER_VERTEX,
 	};
 
 	VulkanContext *vulkan = (VulkanContext *)draw_->GetNativeObject(Draw::NativeObject::CONTEXT);
@@ -94,9 +91,6 @@ void DrawEngineVulkan::InitDeviceObjects() {
 	res = vkCreateSampler(device, &samp, nullptr, &nullSampler_);
 	_dbg_assert_(VK_SUCCESS == res);
 
-	tessDataTransferVulkan = new TessellationDataTransferVulkan(vulkan);
-	tessDataTransfer = tessDataTransferVulkan;
-
 	draw_->SetInvalidationCallback(std::bind(&DrawEngineVulkan::Invalidate, this, std::placeholders::_1));
 }
 
@@ -115,10 +109,6 @@ void DrawEngineVulkan::DestroyDeviceObjects() {
 
 	draw_->SetInvalidationCallback(InvalidationCallback());
 
-	delete tessDataTransferVulkan;
-	tessDataTransfer = nullptr;
-	tessDataTransferVulkan = nullptr;
-
 	pushUBO_ = nullptr;
 
 	if (pushVertex_) {
@@ -166,8 +156,6 @@ void DrawEngineVulkan::BeginFrame() {
 	pushVertex_->BeginFrame();
 	pushIndex_->BeginFrame();
 
-	tessDataTransferVulkan->SetPushPool(pushUBO_);
-
 	DirtyAllUBOs();
 
 	AssertEmpty();
@@ -318,7 +306,7 @@ void DrawEngineVulkan::Flush() {
 			VulkanVertexShader *vshader = nullptr;
 			VulkanFragmentShader *fshader = nullptr;
 
-			shaderManager_->GetShaders(prim, dec_->VertexType(), &vshader, &fshader, pipelineState_, true, useHWTessellation_, decOptions_.expandAllWeightsToFloat, applySkinInDecode_, clipInfoFlags_);
+			shaderManager_->GetShaders(prim, dec_->VertexType(), &vshader, &fshader, pipelineState_, true, decOptions_.expandAllWeightsToFloat, applySkinInDecode_, clipInfoFlags_);
 			_dbg_assert_msg_(vshader->UseHWTransform(), "Bad vshader");
 			VulkanPipeline *pipeline = pipelineManager_->GetOrCreatePipeline(renderManager, pipelineLayout_, pipelineKey_, &dec_->decFmt, vshader, fshader, true, 0, framebufferManager_->GetMSAALevel(), false);
 			if (!pipeline || !pipeline->pipeline) {
@@ -351,8 +339,6 @@ void DrawEngineVulkan::Flush() {
 		UpdateUBOs();
 
 		int descCount = 6;
-		if (tess)
-			descCount = 9;
 		int descSetIndex;
 		PackedDescriptor *descriptors = renderManager->PushDescriptorSet(descCount, &descSetIndex);
 		descriptors[0].image.view = imageView;
@@ -375,14 +361,7 @@ void DrawEngineVulkan::Flush() {
 		descriptors[5].buffer.buffer = boneBuf;
 		descriptors[5].buffer.range = sizeof(UB_VS_Bones);
 		descriptors[5].buffer.offset = 0;
-		if (tess) {
-			const VkDescriptorBufferInfo *bufInfo = tessDataTransferVulkan->GetBufferInfo();
-			for (int j = 0; j < 3; j++) {
-				descriptors[j + 6].buffer.buffer = bufInfo[j].buffer;
-				descriptors[j + 6].buffer.range = bufInfo[j].range;
-				descriptors[j + 6].buffer.offset = bufInfo[j].offset;
-			}
-		}
+
 		// TODO: Can we avoid binding all three when not needed? Same below for hardware transform.
 		// Think this will require different descriptor set layouts.
 		const uint32_t dynamicUBOOffsets[3] = {
@@ -485,7 +464,7 @@ void DrawEngineVulkan::Flush() {
 				VulkanVertexShader *vshader = nullptr;
 				VulkanFragmentShader *fshader = nullptr;
 
-				shaderManager_->GetShaders(prim, swDec->VertexType(), &vshader, &fshader, pipelineState_, false, false, decOptions_.expandAllWeightsToFloat, true, clipInfoFlags_);
+				shaderManager_->GetShaders(prim, swDec->VertexType(), &vshader, &fshader, pipelineState_, false, decOptions_.expandAllWeightsToFloat, true, clipInfoFlags_);
 				_dbg_assert_msg_(!vshader->UseHWTransform(), "Bad vshader");
 				VulkanPipeline *pipeline = pipelineManager_->GetOrCreatePipeline(renderManager, pipelineLayout_, pipelineKey_, &swDec->decFmt, vshader, fshader, false, 0, framebufferManager_->GetMSAALevel(), false);
 				if (!pipeline || !pipeline->pipeline) {
@@ -605,38 +584,3 @@ void DrawEngineVulkan::UpdateUBOs() {
 		dirtyUniforms_ &= ~DIRTY_BONE_UNIFORMS;
 	}
 }
-
-void TessellationDataTransferVulkan::SendDataToShader(const SimpleVertex *const *points, int size_u, int size_v, u32 vertType, const Spline::Weight2D &weights) {
-	// SSBOs that are not simply float1 or float2 need to be padded up to a float4 size. vec3 members
-	// also need to be 16-byte aligned, hence the padding.
-	struct TessData {
-		float pos[3]; float pad1;
-		float uv[2]; float pad2[2];
-		float color[4];
-	};
-
-	int size = size_u * size_v;
-
-	int ssboAlignment = vulkan_->GetPhysicalDeviceProperties().properties.limits.minStorageBufferOffsetAlignment;
-	uint8_t *data = (uint8_t *)push_->Allocate(size * sizeof(TessData), ssboAlignment, &bufInfo_[0].buffer, (uint32_t *)&bufInfo_[0].offset);
-	bufInfo_[0].range = size * sizeof(TessData);
-
-	float *pos = (float *)(data);
-	float *tex = (float *)(data + offsetof(TessData, uv));
-	float *col = (float *)(data + offsetof(TessData, color));
-	int stride = sizeof(TessData) / sizeof(float);
-
-	CopyControlPoints(pos, tex, col, stride, stride, stride, points, size, vertType);
-
-	using Spline::Weight;
-
-	// Weights U
-	data = (uint8_t *)push_->Allocate(weights.size_u * sizeof(Weight), ssboAlignment, &bufInfo_[1].buffer, (uint32_t *)&bufInfo_[1].offset);
-	memcpy(data, weights.u, weights.size_u * sizeof(Weight));
-	bufInfo_[1].range = weights.size_u * sizeof(Weight);
-
-	// Weights V
-	data = (uint8_t *)push_->Allocate(weights.size_v * sizeof(Weight), ssboAlignment, &bufInfo_[2].buffer, (uint32_t *)&bufInfo_[2].offset);
-	memcpy(data, weights.v, weights.size_v * sizeof(Weight));
-	bufInfo_[2].range = weights.size_v * sizeof(Weight);
-}
diff --git a/GPU/Vulkan/DrawEngineVulkan.h b/GPU/Vulkan/DrawEngineVulkan.h
index 196159afb2..f941772d53 100644
--- a/GPU/Vulkan/DrawEngineVulkan.h
+++ b/GPU/Vulkan/DrawEngineVulkan.h
@@ -64,20 +64,6 @@ struct DrawEngineVulkanStats {
 
 class VulkanRenderManager;
 
-class TessellationDataTransferVulkan : public TessellationDataTransfer  {
-public:
-	TessellationDataTransferVulkan(VulkanContext *vulkan) : vulkan_(vulkan) {}
-
-	void SetPushPool(VulkanPushPool *push) { push_ = push; }
-	// Send spline/bezier's control points and weights to vertex shader through structured shader buffer.
-	void SendDataToShader(const SimpleVertex *const *points, int size_u, int size_v, u32 vertType, const Spline::Weight2D &weights) override;
-	const VkDescriptorBufferInfo *GetBufferInfo() { return bufInfo_; }
-private:
-	VulkanContext *vulkan_;
-	VulkanPushPool *push_;  // Updated each frame.
-	VkDescriptorBufferInfo bufInfo_[3]{};
-};
-
 enum {
 	DRAW_BINDING_TEXTURE = 0,
 	DRAW_BINDING_2ND_TEXTURE = 1,
@@ -227,7 +213,4 @@ private:
 
 	int tessOffset_ = 0;
 	FBOTexState fboTexBindState_ = FBO_TEX_NONE;
-
-	// Hardware tessellation
-	TessellationDataTransferVulkan *tessDataTransferVulkan = nullptr;
 };
diff --git a/GPU/Vulkan/GPU_Vulkan.cpp b/GPU/Vulkan/GPU_Vulkan.cpp
index aaa5871baa..36375f4152 100644
--- a/GPU/Vulkan/GPU_Vulkan.cpp
+++ b/GPU/Vulkan/GPU_Vulkan.cpp
@@ -214,8 +214,6 @@ u32 GPU_Vulkan::CheckGPUFeatures() const {
 
 	// Mandatory features on Vulkan, which may be checked in "centralized" code
 	features |= GPU_USE_TEXTURE_LOD_CONTROL;
-	features |= GPU_USE_INSTANCE_RENDERING;
-	features |= GPU_USE_VERTEX_TEXTURE_FETCH;
 	features |= GPU_USE_TEXTURE_FLOAT;
 
 	if (!draw_->GetBugs().Has(Draw::Bugs::PVR_BAD_16BIT_TEXFORMATS)) {
diff --git a/GPU/Vulkan/ShaderManagerVulkan.cpp b/GPU/Vulkan/ShaderManagerVulkan.cpp
index 276895ad8c..9366a1d7f1 100644
--- a/GPU/Vulkan/ShaderManagerVulkan.cpp
+++ b/GPU/Vulkan/ShaderManagerVulkan.cpp
@@ -244,7 +244,7 @@ uint64_t ShaderManagerVulkan::UpdateUniforms(bool useBufferedRendering) {
 	return dirty;
 }
 
-void ShaderManagerVulkan::GetShaders(int prim, u32 vertexType, VulkanVertexShader **vshader, VulkanFragmentShader **fshader, const ComputedPipelineState &pipelineState, bool useHWTransform, bool useHWTessellation, bool weightsAsFloat, bool useSkinInDecode, ClipInfoFlags clipInfoFlags) {
+void ShaderManagerVulkan::GetShaders(int prim, u32 vertexType, VulkanVertexShader **vshader, VulkanFragmentShader **fshader, const ComputedPipelineState &pipelineState, bool useHWTransform, bool weightsAsFloat, bool useSkinInDecode, ClipInfoFlags clipInfoFlags) {
 	VulkanContext *vulkan = (VulkanContext *)draw_->GetNativeObject(Draw::NativeObject::CONTEXT);
 
 	VShaderID VSID;
@@ -254,7 +254,7 @@ void ShaderManagerVulkan::GetShaders(int prim, u32 vertexType, VulkanVertexShade
 	if (gstate_c.IsDirty(DIRTY_VERTEXSHADER_STATE)) {
 		gstate_c.Clean(DIRTY_VERTEXSHADER_STATE);
 		recomputedVS = true;
-		ComputeVertexShaderID(&VSID, vertexType, useHWTransform, useHWTessellation, weightsAsFloat, useSkinInDecode, clipInfoFlags);
+		ComputeVertexShaderID(&VSID, vertexType, useHWTransform, weightsAsFloat, useSkinInDecode, clipInfoFlags);
 		if (VSID == lastVSID_) {
 			_dbg_assert_(lastVShader_ != nullptr);
 			vs = lastVShader_;
diff --git a/GPU/Vulkan/ShaderManagerVulkan.h b/GPU/Vulkan/ShaderManagerVulkan.h
index 5a72bda512..6104880feb 100644
--- a/GPU/Vulkan/ShaderManagerVulkan.h
+++ b/GPU/Vulkan/ShaderManagerVulkan.h
@@ -102,7 +102,7 @@ public:
 	void DeviceLost() override;
 	void DeviceRestore(Draw::DrawContext *draw) override;
 
-	void GetShaders(int prim, u32 vertexType, VulkanVertexShader **vshader, VulkanFragmentShader **fshader, const ComputedPipelineState &pipelineState, bool useHWTransform, bool useHWTessellation, bool weightsAsFloat, bool useSkinInDecode, ClipInfoFlags clipInfoFlags);
+	void GetShaders(int prim, u32 vertexType, VulkanVertexShader **vshader, VulkanFragmentShader **fshader, const ComputedPipelineState &pipelineState, bool useHWTransform, bool weightsAsFloat, bool useSkinInDecode, ClipInfoFlags clipInfoFlags);
 	void ClearShaders() override;
 	void DirtyLastShader() override;
 
diff --git a/UI/GameSettingsScreen.cpp b/UI/GameSettingsScreen.cpp
index 4b1b6fa2e7..af4cd23013 100644
--- a/UI/GameSettingsScreen.cpp
+++ b/UI/GameSettingsScreen.cpp
@@ -159,35 +159,6 @@ void GameSettingsScreen::PreCreateViews() {
 	ReloadAllThemeInfo();
 }
 
-// This needs before run CheckGPUFeatures()
-// TODO: Remove this if fix the issue
-static bool CheckSupportShaderTessellationGLES() {
-#if PPSSPP_PLATFORM(UWP)
-	return true;
-#else
-	// TODO: Make work with non-GL backends
-	int maxVertexTextureImageUnits = gl_extensions.maxVertexTextureUnits;
-	bool vertexTexture = maxVertexTextureImageUnits >= 3; // At least 3 for hardware tessellation
-
-	bool textureFloat = gl_extensions.ARB_texture_float || gl_extensions.OES_texture_float;
-	bool hasTexelFetch = gl_extensions.GLES3 || (!gl_extensions.IsGLES && gl_extensions.VersionGEThan(3, 3, 0)) || gl_extensions.EXT_gpu_shader4;
-
-	return vertexTexture && textureFloat && hasTexelFetch;
-#endif
-}
-
-static bool DoesBackendSupportHWTess() {
-	switch (GetGPUBackend()) {
-	case GPUBackend::OPENGL:
-		return CheckSupportShaderTessellationGLES();
-	case GPUBackend::VULKAN:
-	case GPUBackend::DIRECT3D11:
-		return true;
-	default:
-		return false;
-	}
-}
-
 static bool UsingHardwareTextureScaling() {
 	// For now, Vulkan only.
 	return g_Config.bTexHardwareScaling && GetGPUBackend() == GPUBackend::VULKAN && !g_Config.bSoftwareRendering;
@@ -563,14 +534,6 @@ void GameSettingsScreen::CreateGraphicsSettings(UI::ViewGroup *graphicsSettings)
 	swSkin->SetDisabledPtr(&g_Config.bSoftwareRendering);
 	graphicsSettings->Add(new SettingHint(gr->T("SoftwareSkinning Tip", "Combine skinned model draws on the CPU, faster in most games"), swSkin));
 
-	if (DoesBackendSupportHWTess()) {
-		CheckBox *tessellationHW = graphicsSettings->Add(new CheckBox(&g_Config.bHardwareTessellation, gr->T("Hardware Tessellation")));
-		tessellationHW->SetEnabledFunc([]() {
-			return !g_Config.bSoftwareRendering && g_Config.bHardwareTransform;
-		});
-		graphicsSettings->Add(new SettingHint(gr->T("HardwareTessellation Tip", "Uses hardware to make curves"), tessellationHW));
-	}
-
 	graphicsSettings->Add(new ItemHeader(gr->T("Texture upscaling")));
 
 	if (GetGPUBackend() == GPUBackend::VULKAN) {
diff --git a/libretro/libretro.cpp b/libretro/libretro.cpp
index b66e5bf763..8bc85d61b9 100644
--- a/libretro/libretro.cpp
+++ b/libretro/libretro.cpp
@@ -818,15 +818,6 @@ static void check_variables(CoreParameter &coreParam)
          g_Config.bSoftwareSkinning = true;
    }
 
-   var.key = "ppsspp_hardware_tesselation";
-   if (environ_cb(RETRO_ENVIRONMENT_GET_VARIABLE, &var) && var.value)
-   {
-      if (!strcmp(var.value, "disabled"))
-         g_Config.bHardwareTessellation = false;
-      else
-         g_Config.bHardwareTessellation = true;
-   }
-
    var.key = "ppsspp_lower_resolution_for_effects";
    if (environ_cb(RETRO_ENVIRONMENT_GET_VARIABLE, &var) && var.value)
    {