diff --git a/bin/resources/shaders/opengl/convert.glsl b/bin/resources/shaders/opengl/convert.glsl
index 2fae5d922f0f3..c3e5a616ff286 100644
--- a/bin/resources/shaders/opengl/convert.glsl
+++ b/bin/resources/shaders/opengl/convert.glsl
@@ -378,4 +378,21 @@ void ps_yuv()
 }
 #endif
 
+#if defined(ps_stencil_image_init_0) || defined(ps_stencil_image_init_1)
+
+void main()
+{
+    SV_Target0 = vec4(0x7FFFFFFF);
+
+    #ifdef ps_stencil_image_init_0
+        if((127.5f / 255.0f) < sample_c().a) // < 0x80 pass (== 0x80 should not pass)
+            SV_Target0 = vec4(-1);
+    #endif
+    #ifdef ps_stencil_image_init_1
+        if(sample_c().a < (127.5f / 255.0f)) // >= 0x80 pass
+            SV_Target0 = vec4(-1);
+    #endif
+}
+#endif
+
 #endif
diff --git a/bin/resources/shaders/opengl/tfx_fs.glsl b/bin/resources/shaders/opengl/tfx_fs.glsl
index df9f30083c362..57719dcd04309 100644
--- a/bin/resources/shaders/opengl/tfx_fs.glsl
+++ b/bin/resources/shaders/opengl/tfx_fs.glsl
@@ -74,36 +74,12 @@ layout(binding = 1) uniform sampler2D PaletteSampler;
 layout(binding = 2) uniform sampler2D RtSampler; // note 2 already use by the image below
 #endif
 
-#ifndef DISABLE_GL42_image
-#if PS_DATE > 0
-// Performance note: images mustn't be declared if they are unused. Otherwise it will
-// require extra shader validation.
-
-// FIXME how to declare memory access
-layout(r32i, binding = 3) uniform iimage2D img_prim_min;
-// WARNING:
-// You can't enable it if you discard the fragment. The depth is still
-// updated (shadow in Shin Megami Tensei Nocturne)
-//
-// early_fragment_tests must still be enabled in the first pass of the 2 passes algo
-// First pass search the first primitive that will write the bad alpha value. Value
-// won't be written if the fragment fails the depth test.
-//
-// In theory the best solution will be do
-// 1/ copy the depth buffer
-// 2/ do the full depth (current depth writes are disabled)
-// 3/ restore the depth buffer for 2nd pass
-// Of course, it is likely too costly.
-#if PS_DATE == 1 || PS_DATE == 2
-layout(early_fragment_tests) in;
-#endif
+#if PS_DATE == 3
+layout(binding = 3) uniform sampler2D img_prim_min;
 
 // I don't remember why I set this parameter but it is surely useless
 //layout(pixel_center_integer) in vec4 gl_FragCoord;
 #endif
-#else
-// use basic stencil
-#endif
 
 vec4 fetch_rt()
 {
@@ -814,8 +790,7 @@ void ps_main()
  	 	discard;
 #endif
 
-#if PS_DATE != 0
-#if ((PS_DATE & 3) == 1 || (PS_DATE & 3) == 2)
+#if PS_DATE >= 5
 
 #if PS_WRITE_RG == 1
     // Pseudo 16 bits access.
@@ -833,25 +808,19 @@ void ps_main()
 #endif
 
     if (bad) {
-#if PS_DATE >= 5 || defined(DISABLE_GL42_image)
         discard;
-#else
-        imageStore(img_prim_min, ivec2(gl_FragCoord.xy), ivec4(-1));
-        return;
-#endif
     }
 
 #endif
 
-#if PS_DATE == 3 && !defined(DISABLE_GL42_image)
-    int stencil_ceil = imageLoad(img_prim_min, ivec2(gl_FragCoord.xy)).r;
+#if PS_DATE == 3
+    int stencil_ceil = int(texelFetch(img_prim_min, ivec2(gl_FragCoord.xy), 0).r);
     // Note gl_PrimitiveID == stencil_ceil will be the primitive that will update
     // the bad alpha value so we must keep it.
 
     if (gl_PrimitiveID > stencil_ceil) {
         discard;
     }
-#endif
 #endif
 
     vec4 C = ps_color();
@@ -934,19 +903,15 @@ void ps_main()
 #endif
 
     // Get first primitive that will write a failling alpha value
-#if PS_DATE == 1 && !defined(DISABLE_GL42_image)
+#if PS_DATE == 1
     // DATM == 0
     // Pixel with alpha equal to 1 will failed (128-255)
-    if (C.a > 127.5f) {
-        imageAtomicMin(img_prim_min, ivec2(gl_FragCoord.xy), gl_PrimitiveID);
-    }
+    SV_Target0 = (C.a > 127.5f) ? vec4(gl_PrimitiveID) : vec4(0x7FFFFFFF);
     return;
-#elif PS_DATE == 2 && !defined(DISABLE_GL42_image)
+#elif PS_DATE == 2
     // DATM == 1
     // Pixel with alpha equal to 0 will failed (0-127)
-    if (C.a < 127.5f) {
-        imageAtomicMin(img_prim_min, ivec2(gl_FragCoord.xy), gl_PrimitiveID);
-    }
+    SV_Target0 = (C.a < 127.5f) ? vec4(gl_PrimitiveID) : vec4(0x7FFFFFFF);
     return;
 #endif
 
diff --git a/pcsx2/GS/GS.cpp b/pcsx2/GS/GS.cpp
index e3d4110383584..5314cc8acc6f1 100644
--- a/pcsx2/GS/GS.cpp
+++ b/pcsx2/GS/GS.cpp
@@ -1477,12 +1477,10 @@ void GSApp::Init()
 	m_default_configuration["OsdShowIndicators"]                          = "1";
 	m_default_configuration["OsdScale"]                                   = "100";
 	m_default_configuration["override_GL_ARB_copy_image"]                 = "-1";
-	m_default_configuration["override_GL_ARB_clear_texture"]              = "-1";
 	m_default_configuration["override_GL_ARB_clip_control"]               = "-1";
 	m_default_configuration["override_GL_ARB_direct_state_access"]        = "-1";
 	m_default_configuration["override_GL_ARB_draw_buffers_blend"]         = "-1";
 	m_default_configuration["override_GL_ARB_gpu_shader5"]                = "-1";
-	m_default_configuration["override_GL_ARB_shader_image_load_store"]    = "-1";
 	m_default_configuration["override_GL_ARB_texture_barrier"]            = "-1";
 	m_default_configuration["OverrideTextureBarriers"]                    = "-1";
 	m_default_configuration["OverrideGeometryShaders"]                    = "-1";
diff --git a/pcsx2/GS/Renderers/OpenGL/GLLoader.cpp b/pcsx2/GS/Renderers/OpenGL/GLLoader.cpp
index 9124722b154c0..86bfc1acf2a92 100644
--- a/pcsx2/GS/Renderers/OpenGL/GLLoader.cpp
+++ b/pcsx2/GS/Renderers/OpenGL/GLLoader.cpp
@@ -148,10 +148,8 @@ namespace GLLoader
 	bool has_dual_source_blend = false;
 	bool found_framebuffer_fetch = false;
 	bool found_geometry_shader = true; // we require GL3.3 so geometry must be supported by default
-	bool found_GL_ARB_clear_texture = false;
 	// DX11 GPU
 	bool found_GL_ARB_gpu_shader5 = false;             // Require IvyBridge
-	bool found_GL_ARB_shader_image_load_store = false; // Intel IB. Nvidia/AMD miss Mesa implementation.
 	bool found_GL_ARB_texture_barrier = false;
 
 	static bool mandatory(const std::string& ext)
@@ -266,10 +264,6 @@ namespace GLLoader
 		{
 			// GL4.0
 			found_GL_ARB_gpu_shader5 = optional("GL_ARB_gpu_shader5");
-			// GL4.2
-			found_GL_ARB_shader_image_load_store = optional("GL_ARB_shader_image_load_store");
-			// GL4.4
-			found_GL_ARB_clear_texture = optional("GL_ARB_clear_texture");
 			// GL4.5
 			optional("GL_ARB_direct_state_access");
 			// Mandatory for the advance HW renderer effect. Unfortunately Mesa LLVMPIPE/SWR renderers doesn't support this extension.
diff --git a/pcsx2/GS/Renderers/OpenGL/GLLoader.h b/pcsx2/GS/Renderers/OpenGL/GLLoader.h
index 30ff6f752c2d4..ecdb6c00dfd23 100644
--- a/pcsx2/GS/Renderers/OpenGL/GLLoader.h
+++ b/pcsx2/GS/Renderers/OpenGL/GLLoader.h
@@ -43,7 +43,5 @@ namespace GLLoader
 	extern bool found_framebuffer_fetch;
 	extern bool found_geometry_shader;
 	extern bool found_GL_ARB_gpu_shader5;
-	extern bool found_GL_ARB_shader_image_load_store;
-	extern bool found_GL_ARB_clear_texture;
 	extern bool found_GL_ARB_texture_barrier;
 } // namespace GLLoader
diff --git a/pcsx2/GS/Renderers/OpenGL/GSDeviceOGL.cpp b/pcsx2/GS/Renderers/OpenGL/GSDeviceOGL.cpp
index db8b4e3dcba2b..668e25c851552 100644
--- a/pcsx2/GS/Renderers/OpenGL/GSDeviceOGL.cpp
+++ b/pcsx2/GS/Renderers/OpenGL/GSDeviceOGL.cpp
@@ -217,7 +217,7 @@ bool GSDeviceOGL::Create(HostDisplay* display)
 	// optional features based on context
 	m_features.broken_point_sampler = GLLoader::vendor_id_amd;
 	m_features.geometry_shader = GLLoader::found_geometry_shader;
-	m_features.image_load_store = GLLoader::found_GL_ARB_shader_image_load_store && GLLoader::found_GL_ARB_clear_texture;
+	m_features.image_load_store = true;
 	if (GSConfig.OverrideTextureBarriers == 0)
 		m_features.texture_barrier = GLLoader::found_framebuffer_fetch; // Force Disabled
 	else if (GSConfig.OverrideTextureBarriers == 1)
@@ -353,21 +353,23 @@ bool GSDeviceOGL::Create(HostDisplay* display)
 		}
 	}
 
+	// these all share the same vertex shader
+	const auto convert_glsl = Host::ReadResourceFileToString("shaders/opengl/convert.glsl");
+	if (!convert_glsl.has_value())
+	{
+		Host::ReportErrorAsync("GS", "Failed to read shaders/opengl/convert.glsl.");
+		return false;
+	}
+
 	// ****************************************************************
 	// convert
 	// ****************************************************************
 	{
 		GL_PUSH("GSDeviceOGL::Convert");
 
-		// these all share the same vertex shader
-		const auto shader = Host::ReadResourceFileToString("shaders/opengl/convert.glsl");
-		if (!shader.has_value())
-		{
-			Host::ReportErrorAsync("GS", "Failed to read shaders/opengl/convert.glsl.");
-			return false;
-		}
 
-		m_convert.vs = GetShaderSource("vs_main", GL_VERTEX_SHADER, m_shader_common_header, *shader, {});
+
+		m_convert.vs = GetShaderSource("vs_main", GL_VERTEX_SHADER, m_shader_common_header, *convert_glsl, {});
 
 		for (size_t i = 0; i < std::size(m_convert.ps); i++)
 		{
@@ -375,7 +377,7 @@ bool GSDeviceOGL::Create(HostDisplay* display)
 			const std::string macro_sel = (static_cast<ShaderConvert>(i) == ShaderConvert::RGBA_TO_8I) ?
                                               StringUtil::StdStringFromFormat("#define PS_SCALE_FACTOR %d\n", GSConfig.UpscaleMultiplier) :
                                               std::string();
-			const std::string ps(GetShaderSource(name, GL_FRAGMENT_SHADER, m_shader_common_header, *shader, macro_sel));
+			const std::string ps(GetShaderSource(name, GL_FRAGMENT_SHADER, m_shader_common_header, *convert_glsl, macro_sel));
 			if (!m_shader_cache.GetProgram(&m_convert.ps[i], m_convert.vs, {}, ps))
 				return false;
 			m_convert.ps[i].SetFormattedName("Convert pipe %s", name);
@@ -527,6 +529,15 @@ bool GSDeviceOGL::Create(HostDisplay* display)
 		m_date.dss = new GSDepthStencilOGL();
 		m_date.dss->EnableStencil();
 		m_date.dss->SetStencil(GL_ALWAYS, GL_REPLACE);
+
+		for (size_t i = 0; i < std::size(m_date.primid_ps); i++)
+		{
+			const std::string ps(GetShaderSource(
+				StringUtil::StdStringFromFormat("ps_stencil_image_init_%d", i),
+				GL_FRAGMENT_SHADER, m_shader_common_header, *convert_glsl, {}));
+			m_shader_cache.GetProgram(&m_date.primid_ps[i], m_convert.vs, {}, ps);
+			m_date.primid_ps[i].SetFormattedName("PrimID Destination Alpha Init %d", i);
+		}
 	}
 
 	// ****************************************************************
@@ -803,41 +814,26 @@ void GSDeviceOGL::ClearDepth(GSTexture* t)
 
 	GL_PUSH("Clear Depth %d", T->GetID());
 
-	if (0 && GLLoader::found_GL_ARB_clear_texture)
-	{
-		// I don't know what the driver does but it creates
-		// some slowdowns on Harry Potter PS
-		// Maybe it triggers some texture relocations, or maybe
-		// it clears also the stencil value (2 times slower)
-		//
-		// Let's disable this code for the moment.
+	OMSetFBO(m_fbo);
+	// RT must be detached, if RT is too small, depth won't be fully cleared
+	// AT tolenico 2 map clip bug
+	OMAttachRt(NULL);
+	OMAttachDs(T);
 
-		// Don't bother with Depth_Stencil insanity
-		T->Clear(NULL);
+	// TODO: check size of scissor before toggling it
+	glDisable(GL_SCISSOR_TEST);
+	const float c = 0.0f;
+	if (GLState::depth_mask)
+	{
+		glClearBufferfv(GL_DEPTH, 0, &c);
 	}
 	else
 	{
-		OMSetFBO(m_fbo);
-		// RT must be detached, if RT is too small, depth won't be fully cleared
-		// AT tolenico 2 map clip bug
-		OMAttachRt(NULL);
-		OMAttachDs(T);
-
-		// TODO: check size of scissor before toggling it
-		glDisable(GL_SCISSOR_TEST);
-		const float c = 0.0f;
-		if (GLState::depth_mask)
-		{
-			glClearBufferfv(GL_DEPTH, 0, &c);
-		}
-		else
-		{
-			glDepthMask(true);
-			glClearBufferfv(GL_DEPTH, 0, &c);
-			glDepthMask(false);
-		}
-		glEnable(GL_SCISSOR_TEST);
+		glDepthMask(true);
+		glClearBufferfv(GL_DEPTH, 0, &c);
+		glDepthMask(false);
 	}
+	glEnable(GL_SCISSOR_TEST);
 }
 
 void GSDeviceOGL::ClearStencil(GSTexture* t, u8 c)
@@ -938,39 +934,15 @@ GSDepthStencilOGL* GSDeviceOGL::CreateDepthStencil(OMDepthStencilSelector dssel)
 	return dss;
 }
 
-void GSDeviceOGL::InitPrimDateTexture(GSTexture* rt, const GSVector4i& area)
+GSTexture* GSDeviceOGL::InitPrimDateTexture(GSTexture* rt, const GSVector4i& area, bool datm)
 {
 	const GSVector2i& rtsize = rt->GetSize();
 
-	// Create a texture to avoid the useless clean@0
-	if (m_date.t == NULL)
-		m_date.t = CreateTexture(rtsize.x, rtsize.y, false, GSTexture::Format::PrimID);
-
-	// Clean with the max signed value
-	const int max_int = 0x7FFFFFFF;
-	static_cast<GSTextureOGL*>(m_date.t)->Clear(&max_int, area);
-
-	glBindImageTexture(3, static_cast<GSTextureOGL*>(m_date.t)->GetID(), 0, false, 0, GL_READ_WRITE, GL_R32I);
-#ifdef ENABLE_OGL_DEBUG
-	// Help to see the texture in apitrace
-	PSSetShaderResource(3, m_date.t);
-#endif
-}
-
-void GSDeviceOGL::RecycleDateTexture()
-{
-	if (m_date.t)
-	{
-		//static_cast<GSTextureOGL*>(m_date.t)->Save(format("/tmp/date_adv_%04ld.csv", GSState::s_n));
-
-		Recycle(m_date.t);
-		m_date.t = NULL;
-	}
-}
+	GSTexture* tex = CreateRenderTarget(rtsize.x, rtsize.y, GSTexture::Format::PrimID, false);
 
-void GSDeviceOGL::Barrier(GLbitfield b)
-{
-	glMemoryBarrier(b);
+	GL_PUSH("PrimID Destination Alpha Clear");
+	StretchRect(rt, GSVector4(area) / GSVector4(rtsize).xyxy(), tex, GSVector4(area), m_date.primid_ps[datm], false);
+	return tex;
 }
 
 std::string GSDeviceOGL::GetShaderSource(const std::string_view& entry, GLenum type, const std::string_view& common_header, const std::string_view& glsl_h_code, const std::string_view& macro_sel)
@@ -1000,16 +972,6 @@ std::string GSDeviceOGL::GenGlslHeader(const std::string_view& entry, GLenum typ
 	if (GLLoader::found_GL_ARB_gpu_shader5)
 		header += "#extension GL_ARB_gpu_shader5 : enable\n";
 
-	if (GLLoader::found_GL_ARB_shader_image_load_store)
-	{
-		// Need GL version 420
-		header += "#extension GL_ARB_shader_image_load_store: require\n";
-	}
-	else
-	{
-		header += "#define DISABLE_GL42_image\n";
-	}
-
 	if (m_features.framebuffer_fetch)
 		header += "#define HAS_FRAMEBUFFER_FETCH 1\n";
 	else
@@ -1856,6 +1818,8 @@ void GSDeviceOGL::RenderHW(GSHWDrawConfig& config)
 
 	GSVector2i rtsize = (config.rt ? config.rt : config.ds)->GetSize();
 
+	GSTexture* primid_texture = nullptr;
+
 	// Destination Alpha Setup
 	switch (config.destination_alpha)
 	{
@@ -1863,7 +1827,7 @@ void GSDeviceOGL::RenderHW(GSHWDrawConfig& config)
 		case GSHWDrawConfig::DestinationAlphaMode::Full:
 			break; // No setup
 		case GSHWDrawConfig::DestinationAlphaMode::PrimIDTracking:
-			InitPrimDateTexture(config.rt, config.drawarea);
+			primid_texture = InitPrimDateTexture(config.rt, config.drawarea, config.datm);
 			break;
 		case GSHWDrawConfig::DestinationAlphaMode::StencilOne:
 			if (m_features.texture_barrier)
@@ -1933,11 +1897,6 @@ void GSDeviceOGL::RenderHW(GSHWDrawConfig& config)
 		PSSetShaderResource(2, config.rt);
 
 	SetupSampler(config.sampler);
-	OMSetBlendState(config.blend.enable, s_gl_blend_factors[config.blend.src_factor],
-		s_gl_blend_factors[config.blend.dst_factor], s_gl_blend_ops[config.blend.op],
-		config.blend.constant_enable, config.blend.constant);
-	OMSetColorMaskState(config.colormask);
-	SetupOM(config.depth);
 
 	if (m_vs_cb_cache.Update(config.cb_vs))
 	{
@@ -1983,35 +1942,32 @@ void GSDeviceOGL::RenderHW(GSHWDrawConfig& config)
 
 	if (config.destination_alpha == GSHWDrawConfig::DestinationAlphaMode::PrimIDTracking)
 	{
-		GL_PUSH("Date GL42");
-		// It could be good idea to use stencil in the same time.
-		// Early stencil test will reduce the number of atomic-load operation
-
-		// Create an r32i image that will contain primitive ID
-		// Note: do it at the beginning because the clean will dirty the FBO state
-		//dev->InitPrimDateTexture(rtsize.x, rtsize.y);
+		GL_PUSH("Destination Alpha PrimID Init");
 
-		// I don't know how much is it legal to mount rt as Texture/RT. No write is done.
-		// In doubt let's detach RT.
-		OMSetRenderTargets(NULL, config.ds, &config.scissor);
+		OMSetRenderTargets(primid_texture, config.ds, &config.scissor);
+		OMColorMaskSelector mask;
+		mask.wrgba = 0;
+		mask.wr = true;
+		OMSetColorMaskState(mask);
+		OMSetBlendState(true, GL_ONE, GL_ONE, GL_MIN);
+		OMDepthStencilSelector dss = config.depth;
+		dss.zwe = 0; // Don't write depth
+		SetupOM(dss);
 
-		// Don't write anything on the color buffer
-		// Neither in the depth buffer
-		glDepthMask(false);
 		// Compute primitiveID max that pass the date test (Draw without barrier)
 		DrawIndexedPrimitive();
 
-		// Ask PS to discard shader above the primitiveID max
-		glDepthMask(GLState::depth_mask);
-
 		psel.ps.date = 3;
 		config.alpha_second_pass.ps.date = 3;
 		SetupPipeline(psel);
-
-		// Be sure that first pass is finished !
-		Barrier(GL_SHADER_IMAGE_ACCESS_BARRIER_BIT);
+		PSSetShaderResource(3, primid_texture);
 	}
 
+	OMSetBlendState(config.blend.enable, s_gl_blend_factors[config.blend.src_factor],
+		s_gl_blend_factors[config.blend.dst_factor], s_gl_blend_ops[config.blend.op],
+		config.blend.constant_enable, config.blend.constant);
+	OMSetColorMaskState(config.colormask);
+	SetupOM(config.depth);
 	OMSetRenderTargets(hdr_rt ? hdr_rt : config.rt, config.ds, &config.scissor);
 
 	SendHWDraw(config, psel.ps.IsFeedbackLoop());
@@ -2063,8 +2019,8 @@ void GSDeviceOGL::RenderHW(GSHWDrawConfig& config)
 		}
 	}
 
-	if (config.destination_alpha == GSHWDrawConfig::DestinationAlphaMode::PrimIDTracking)
-		RecycleDateTexture();
+	if (primid_texture)
+		Recycle(primid_texture);
 	if (draw_rt_clone)
 		Recycle(draw_rt_clone);
 
diff --git a/pcsx2/GS/Renderers/OpenGL/GSDeviceOGL.h b/pcsx2/GS/Renderers/OpenGL/GSDeviceOGL.h
index addf11732f0fa..78136edb62885 100644
--- a/pcsx2/GS/Renderers/OpenGL/GSDeviceOGL.h
+++ b/pcsx2/GS/Renderers/OpenGL/GSDeviceOGL.h
@@ -272,7 +272,7 @@ class GSDeviceOGL final : public GSDevice
 	struct
 	{
 		GSDepthStencilOGL* dss = nullptr;
-		GSTexture* t = nullptr;
+		GL::Program primid_ps[2];
 	} m_date;
 
 	struct
@@ -338,8 +338,7 @@ class GSDeviceOGL final : public GSDevice
 	void ClearDepth(GSTexture* t) final;
 	void ClearStencil(GSTexture* t, u8 c) final;
 
-	void InitPrimDateTexture(GSTexture* rt, const GSVector4i& area);
-	void RecycleDateTexture();
+	GSTexture* InitPrimDateTexture(GSTexture* rt, const GSVector4i& area, bool datm);
 
 	bool DownloadTexture(GSTexture* src, const GSVector4i& rect, GSTexture::GSMap& out_map) final;
 
@@ -391,6 +390,4 @@ class GSDeviceOGL final : public GSDevice
 	void SetupOM(OMDepthStencilSelector dssel);
 	GLuint GetSamplerID(PSSamplerSelector ssel);
 	GLuint GetPaletteSamplerID();
-
-	void Barrier(GLbitfield b);
 };
diff --git a/pcsx2/GS/Renderers/OpenGL/GSTextureOGL.cpp b/pcsx2/GS/Renderers/OpenGL/GSTextureOGL.cpp
index 160c6c9a83f26..37647e7e91e25 100644
--- a/pcsx2/GS/Renderers/OpenGL/GSTextureOGL.cpp
+++ b/pcsx2/GS/Renderers/OpenGL/GSTextureOGL.cpp
@@ -189,8 +189,8 @@ GSTextureOGL::GSTextureOGL(Type type, int width, int height, int levels, Format
 	{
 		// 1 Channel integer
 		case Format::PrimID:
-			gl_fmt          = GL_R32I;
-			m_int_format    = GL_RED_INTEGER;
+			gl_fmt          = GL_R32F;
+			m_int_format    = GL_RED;
 			m_int_type      = GL_INT;
 			m_int_shift     = 2;
 			break;
diff --git a/pcsx2/GS/Window/GSSetting.cpp b/pcsx2/GS/Window/GSSetting.cpp
index fd1ffe4541fe0..9516856b60120 100644
--- a/pcsx2/GS/Window/GSSetting.cpp
+++ b/pcsx2/GS/Window/GSSetting.cpp
@@ -168,10 +168,6 @@ const char* dialog_message(int ID, bool* updateText)
 		case IDC_GEOMETRY_SHADER_OVERRIDE:
 			return cvtString("Allows the GPU instead of just the CPU to transform lines into sprites. This reduces CPU load and bandwidth requirement, but it is heavier on the GPU.\n"
 				"Automatic detection is recommended.");
-		case IDC_IMAGE_LOAD_STORE:
-			return cvtString("Allows advanced atomic operations to speed up Accurate DATE.\n"
-				"Only disable this if using Accurate DATE causes (GPU driver) issues.\n\n"
-				"Note: This option is only supported by GPUs which support at least Direct3D 11.");
 		case IDC_LINEAR_PRESENT:
 			return cvtString("Use bilinear filtering when Upscaling/Downscaling the image to the screen. Disable it if you want a sharper/pixelated output.");
 		// Exclusive for Hardware Renderer
diff --git a/pcsx2/GS/Window/GSSetting.h b/pcsx2/GS/Window/GSSetting.h
index 423a19ce37a3b..012f7388795cd 100644
--- a/pcsx2/GS/Window/GSSetting.h
+++ b/pcsx2/GS/Window/GSSetting.h
@@ -88,7 +88,6 @@ enum
 	IDC_SWTHREADS_EDIT,
 	// OpenGL Advanced Settings
 	IDC_GEOMETRY_SHADER_OVERRIDE,
-	IDC_IMAGE_LOAD_STORE,
 	// On-screen Display
 	IDC_OSD_LOG,
 	IDC_OSD_MONITOR,
diff --git a/pcsx2/GS/Window/GSwxDialog.cpp b/pcsx2/GS/Window/GSwxDialog.cpp
index 3640d96d68caf..bb7de46371fa5 100644
--- a/pcsx2/GS/Window/GSwxDialog.cpp
+++ b/pcsx2/GS/Window/GSwxDialog.cpp
@@ -548,7 +548,6 @@ DebugTab::DebugTab(wxWindow* parent)
 	const int space = wxSizerFlags().Border().GetBorderInPixels();
 	PaddedBoxSizer<wxBoxSizer> tab_box(wxVERTICAL);
 
-	auto ogl_hw_prereq = [this]{ return m_is_ogl_hw; };
 	auto vk_ogl_hw_prereq = [this] { return m_is_ogl_hw || m_is_vk_hw; };
 
 	if (g_Conf->DevMode || IsDevBuild)
@@ -584,10 +583,9 @@ DebugTab::DebugTab(wxWindow* parent)
 
 	PaddedBoxSizer<wxStaticBoxSizer> ogl_box(wxVERTICAL, this, "Overrides");
 	auto* ogl_grid = new wxFlexGridSizer(2, space, space);
-	m_ui.addComboBoxAndLabel(ogl_grid, "Texture Barriers:", "OverrideTextureBarriers",                 &theApp.m_gs_generic_list, -1,                           vk_ogl_hw_prereq);
-	m_ui.addComboBoxAndLabel(ogl_grid, "Geometry Shader:",  "OverrideGeometryShaders",                 &theApp.m_gs_generic_list, IDC_GEOMETRY_SHADER_OVERRIDE, vk_ogl_hw_prereq);
-	m_ui.addComboBoxAndLabel(ogl_grid, "Image Load Store:", "override_GL_ARB_shader_image_load_store", &theApp.m_gs_generic_list, IDC_IMAGE_LOAD_STORE,         ogl_hw_prereq);
-	m_ui.addComboBoxAndLabel(ogl_grid, "Dump Compression:", "GSDumpCompression",                       &theApp.m_gs_dump_compression, -1);
+	m_ui.addComboBoxAndLabel(ogl_grid, "Texture Barriers:", "OverrideTextureBarriers", &theApp.m_gs_generic_list, -1,                           vk_ogl_hw_prereq);
+	m_ui.addComboBoxAndLabel(ogl_grid, "Geometry Shader:",  "OverrideGeometryShaders", &theApp.m_gs_generic_list, IDC_GEOMETRY_SHADER_OVERRIDE, vk_ogl_hw_prereq);
+	m_ui.addComboBoxAndLabel(ogl_grid, "Dump Compression:", "GSDumpCompression",       &theApp.m_gs_dump_compression, -1);
 	ogl_box->Add(ogl_grid);
 
 	tab_box->Add(ogl_box.outer, wxSizerFlags().Expand());