Optimize this EMBM code

[Leadwerks]

I wanted to see if there is a way to eliminate the per-fragment matrix multiplier in my EMBM shader. I admit I don’t really understand what is going on here, and I just tried different things until I found something that worked.

The part I am interested in is in the second block of code, but I included the first just so you could see what is happening.

Here is the part that gets the normal from the normal map:

	#ifdef LW_BUMPMAP
		#ifdef LW_LIGHTMAP
			lightdir = gl_NormalMatrix * ( ( texture2D(radiositymap,gl_TexCoord[1].st).xyz - 0.5 ));
		#else
			#ifdef LW_LIGHTS
				lightdir = BumpVector;			
			#else	
				lightdir = gl_NormalMatrix * ((BumpVector-0.5)*2.0);
			#endif
		#endif
		bumpcolor = texture2D(bumpmap,basetexcoord);	
		vec3 halfvec = normalize(normalize(lightdir) + normalize(-ModelVertex));	
		normal = normalize(bumpcolor.xyz - 0.5);
		normal = T * normal.x + B * normal.y + N * normal.z;
		lightcolor = lightcolor * max(0.0,dot(normal,lightdir)) * 2.0;
		specular = pow(max(0.0, dot(halfvec,normal)),8.0) * bumpcolor.w * 0.5;
		#ifdef LW_EMBM
			specular = 0.0;
		#endif
	#else
		normal = N;
	#endif

Here is the part that perturbs the cubemap coords by the normal:

		#ifdef LW_EMBM		
			vec3 cubecoord = gl_TexCoord[3].xyz;			
			normal *= gl_NormalMatrix;
			cubecoord += vec3(normal.x,-normal.y,-normal.z) * 100.0;
			//vec3 cubecoord = gl_TexCoord[3].xyz * vec3(normal.x,normal.y,-normal.z) * 100.0;
			gl_FragColor = gl_FragColor * alpha + textureCube( cubemap,cubecoord * vec3(1,1,-1) ) * (1.0 - alpha);
		#else
			gl_FragColor = gl_FragColor * alpha + textureCube( cubemap,gl_TexCoord[3].xyz * vec3(1,1,-1) ) * (1.0 - alpha);
		#endif

[Leadwerks]

I wanted to see if there is another way to handle this part:

vec3 cubecoord = gl_TexCoord[3].xyz;
normal *= gl_NormalMatrix;
cubecoord += vec3(normal.x,-normal.y,-normal.z) * 100.0;

It seems kind of weird that the normal should be multiplied by the normal matrix.