Linux Certif

Toute la documentation sur la certification Linux LPI

clamp.3Cg

NAME

clamp - returns smallest integer not less than a scalar or each vector component.

SYNOPSIS

   float  clamp(float x, float a, float b);
   float1 clamp(float1 x, float1 a, float1 b);
   float2 clamp(float2 x, float2 a, float2 b);
   float3 clamp(float3 x, float3 a, float3 b);
   float4 clamp(float4 x, float4 a, float4 b);
 
 
 

   half   clamp(half x, half a, half b);
   half1  clamp(half1 x, half1 a, half1 b);
   half2  clamp(half2 x, half2 a, half2 b);
   half3  clamp(half3 x, half3 a, half3 b);
   half4  clamp(half4 x, half4 a, half4 b);
 
 
 

   fixed  clamp(fixed x, fixed a, fixed b);
   fixed1 clamp(fixed1 x, fixed1 a, fixed1 b);
   fixed2 clamp(fixed2 x, fixed2 a, fixed2 b);
   fixed3 clamp(fixed3 x, fixed3 a, fixed3 b);
   fixed4 clamp(fixed4 x, fixed4 a, fixed4 b);
 
 
 

PARAMETERS

x

Vector or scalar to clamp.

a

Vector or scalar for bottom of clamp range.

b

Vector or scalar for top of clamp range.

DESCRIPTION

Returns x clamped to the range [a,b] as follows:

1) Returns a if x is less than a; else

2) Returns b if x is greater than b; else

3) Returns x otherwise.

For vectors, the returned vector contains the clamped result of each element of the vector x clamped using the respective element of vectors a and b.

REFERENCE IMPLEMENTATION

clamp for float scalars could be implemented like this.

   float clamp(float x, float a, float b)
   {
     return max(a, min(b, x));
   }
 
 
 

PROFILE SUPPORT

clamp is supported in all profiles except fp20.

SEE ALSO

max, min, saturate