in my geometry shader I have something like this:

float r[10]={0,0,0,0,0,0,0,0,0,0};
float4 g;

uint max=Tex2Da.Load(uint3(0,0,0));

for(uint i=0;i<max;i++)  //max is never greater than 10
{
     g  = Tex2Db.Load(uint3(0,0,0));

     uint mod=i%4;
     r   = g[mod];
}

g is a vector but you access it as an array of vectors.

yeah that can be done in HLSL, there is nothing wrong about that, even if i change

r[i] = g[mod];

for this:

r[i] = g.x;

it will return the same error

The rest of the shader is ugly but ok here is it complete

struct VS_PPLL
{
    float4 Pos : PITION;
	uint Id : anceID;
};

struct PS_Ff
{
    float4 Pos : SV_POSITION;
	float4 Frc : FFOO;
};


[maxvertexcount(17)] 
void GSInstJoFo( point VS_PPLL input[1] , inout PointStream<PS_Ff> duo )
{
	PS_Ff output;

	uint cC=uint(C.y);

	uint3 u=uint3(input[0].Id%cC,input[0].Id/cC,0);

	output.Pos.x=(float(u.x)/C.y)*2.0f-1.0f+1.0f/C.y;
	output.Pos.y=1.0f-(float(u.y)/C.y)*2.0f-1.0f/C.y;
	output.Pos.zw=float2(0.0f,1.0f);

	output.Frc=Tex2Db.Load(u);

	uint nJo=uint(Tex2Dg.Load(u));
	     if(!(any(output.Frc.xyz))){
	duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();
	duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();}
	else if(nJo==0 && any(output.Frc.xyz)){ duo.Append( output );
	duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();
	duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();}
	else
	{
		uint i=0;
		float3 p1=Tex2Da.Load(u).xyz;
		uint3 e=u;
		float4 jntsClss;
		float3 dSum=float3(0.0f,0.0f,0.0f); float mCntr=0.0f;
		float doit,l;
		float3 arrD[16];
		uint4 arrJP[4];
		uint fl,fl2;
		for(i=0;i<nJo;i++)
		{
			fl=i%4;
			if(fl==0)
			{
				fl2=i/4;
				e.x = (u.x*4) + fl2;
				arrJP[fl2]     = uint4(Tex2Dc.Load(e));
				jntsClss = Tex2Df.Load(e);
			}

			e.x=arrJP[fl2][fl]%cC;
			e.y=arrJP[fl2][fl]/cC;

			arrD=Tex2Da.Load(e).xyz;

			arrD=arrD-p1;
			if(any(arrD))
			{
				l=length(arrD);
				if(l<ballCLe(jntsClss[fl]))
				{
					arrD=normalize(arrD);
					doit=dot(arrD,output.Frc.xyz);
					if(doit!=0.0f)
					{
						dSum+=doit*arrD;
						mCntr++;
					}
					else
					{
						arrD=float3(0.0f,0.0f,0.0f);
					}
				}
			}
		}
		if(mCntr==0.0f){duo.Append( output );
		duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();
		duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();duo.RestartStrip();}
		else
		{
			dSum=normalize(dSum);
			dSum=dSum*dot(dSum,output.Frc.xyz);
			float3 dPer=normalize(output.Frc.xyz-dSum);
			dPer=dPer*dot(dPer,output.Frc.xyz);

			dSum=dSum/mCntr;

			output.Frc.xyz=dSum+dPer;
			duo.Append( output );duo.RestartStrip();

			for(i=0;i<nJo;i++)
			{
				if(any(arrD))
				{
					output.Frc.xyz=arrD*dot(arrD,dSum);

					fl=i%4;fl2=i/4;

					u.xy=uint2(arrJP[fl2][fl]%cC,arrJP[fl2][fl]/cC);
					output.Pos.x=(float(u.x)/C.y)*2.0f-1.0f+1.0f/C.y;
					output.Pos.y=1.0f-(float(u.y)/C.y)*2.0f-1.0f/C.y;

					duo.Append( output );duo.RestartStrip();
				}
				else{duo.RestartStrip();}
			}
			for(i=0;i<16-nJo;i++)
			{
				duo.RestartStrip();
			}
		}	
	}
}

The error points to this, the one that is more down in the code

output.Pos.x=(float(u.x)/C.y)*2.0f-1.0f+1.0f/C.y;
output.Pos.y=1.0f-(float(u.y)/C.y)*2.0f-1.0f/C.y;

which depends on "u"

"u" depends on the array "arrJP"

"arrJP" depends on "Tex2Dc.Load(e)"

"Tex2Dc.Load(e)" depends on "e"

"e" depends on the loop number "i"

"i" depends on the loop wich has max value "nJo"

"nJo" depends on "Tex2Dg.Load(u)"

"Tex2Dg.Load(u)" depends on "u"

finally "u" depends on VS_PPLL--> Id

C.y is a constant

----------------------------------------------------------------------------------------------------------------

ERROR IS FIXED BY INITIALIZING

float3 arrD[16];
uint4 arrJP[4];