diff --git a/src/gl/scene/gl_flats.cpp b/src/gl/scene/gl_flats.cpp
index 91c7f1c7e..19eb3b9e1 100644
--- a/src/gl/scene/gl_flats.cpp
+++ b/src/gl/scene/gl_flats.cpp
@@ -298,9 +298,9 @@ void GLFlat::DrawSubsectors(int pass, bool istrans)
 				if (gl_drawinfo->ss_renderflags[sub-subsectors]&renderflags || istrans)
 				{
 					if (pass == GLPASS_ALL) lightsapplied = SetupSubsectorLights(lightsapplied, sub);
-					//drawcalls.Clock();
+					drawcalls.Clock();
 					glDrawArrays(GL_TRIANGLE_FAN, index, sub->numlines);
-					//drawcalls.Unclock();
+					drawcalls.Unclock();
 					flatvertices += sub->numlines;
 					flatprimitives++;
 				}
diff --git a/src/gl/shaders/gl_shader.cpp b/src/gl/shaders/gl_shader.cpp
index b056def8e..db3e68298 100644
--- a/src/gl/shaders/gl_shader.cpp
+++ b/src/gl/shaders/gl_shader.cpp
@@ -293,8 +293,18 @@ FShader *FShaderManager::Compile (const char *ShaderName, const char *ShaderPath
 
 void FShader::ApplyMatrices(VSMatrix *proj, VSMatrix *view)
 {
-	glProgramUniformMatrix4fv(hShader, projectionmatrix_index, 1, false, proj->get());
-	glProgramUniformMatrix4fv(hShader, viewmatrix_index, 1, false, view->get());
+
+	if (gl.flags & RFL_SEPARATE_SHADER_OBJECTS)
+	{
+		glProgramUniformMatrix4fv(hShader, projectionmatrix_index, 1, false, proj->get());
+		glProgramUniformMatrix4fv(hShader, viewmatrix_index, 1, false, view->get());
+	}
+	else
+	{
+		Bind();
+		glUniformMatrix4fv(projectionmatrix_index, 1, false, proj->get());
+		glUniformMatrix4fv(viewmatrix_index, 1, false, view->get());
+	}
 }
 
 
@@ -486,7 +496,16 @@ void FShaderManager::SetWarpSpeed(unsigned int eff, float speed)
 		FShader *sh = mTextureEffects[eff];
 
 		float warpphase = gl_frameMS * speed / 1000.f;
-		glProgramUniform1f(sh->GetHandle(), sh->timer_index, warpphase);
+		if (gl.flags & RFL_SEPARATE_SHADER_OBJECTS)
+		{
+			glProgramUniform1f(sh->GetHandle(), sh->timer_index, warpphase);
+		}
+		else
+		{
+			// not so pretty...
+			sh->Bind();
+			glUniform1f(sh->timer_index, warpphase);
+		}
 	}
 }
 
diff --git a/src/gl/system/gl_interface.cpp b/src/gl/system/gl_interface.cpp
index e1c70b14a..f00202fd8 100644
--- a/src/gl/system/gl_interface.cpp
+++ b/src/gl/system/gl_interface.cpp
@@ -133,6 +133,8 @@ void gl_LoadExtensions()
 	if (CheckExtension("GL_ARB_texture_compression")) gl.flags|=RFL_TEXTURE_COMPRESSION;
 	if (CheckExtension("GL_EXT_texture_compression_s3tc")) gl.flags|=RFL_TEXTURE_COMPRESSION_S3TC;
 	if (CheckExtension("GL_ARB_buffer_storage")) gl.flags |= RFL_BUFFER_STORAGE;
+	if (CheckExtension("GL_ARB_separate_shader_objects")) gl.flags |= RFL_SEPARATE_SHADER_OBJECTS;
+	
 
 	glGetIntegerv(GL_MAX_TEXTURE_SIZE,&gl.max_texturesize);
 	glPixelStorei(GL_UNPACK_ALIGNMENT, 1);
diff --git a/src/gl/system/gl_interface.h b/src/gl/system/gl_interface.h
index 74beb72a4..0f569d2c4 100644
--- a/src/gl/system/gl_interface.h
+++ b/src/gl/system/gl_interface.h
@@ -9,9 +9,10 @@ enum RenderFlags
 	RFL_TEXTURE_COMPRESSION=1,
 	RFL_TEXTURE_COMPRESSION_S3TC=2,
 
-	RFL_BUFFER_STORAGE = 8,
-	RFL_SHADER_STORAGE_BUFFER = 16,
-	RFL_BASEINDEX = 32,
+	RFL_SEPARATE_SHADER_OBJECTS = 4,	// we need this extension for glProgramUniform. On hardware not supporting it we need some rather clumsy workarounds
+	RFL_BUFFER_STORAGE = 8,				// allows persistently mapped buffers, which are the only efficient way to actually use a dynamic vertex buffer. If this isn't present, a workaround with uniform arrays is used.
+	RFL_SHADER_STORAGE_BUFFER = 16,		// to be used later for a parameter buffer
+	RFL_BASEINDEX = 32,					// currently unused
 };
 
 enum TexMode