Name

trunc - returns largest integer not greater than a scalar or each vector component.

Synopsis

float  trunc(float x);
float1 trunc(float1 x);
float2 trunc(float2 x);
float3 trunc(float3 x);
float4 trunc(float4 x);

half   trunc(half x);
half1  trunc(half1 x);
half2  trunc(half2 x);
half3  trunc(half3 x);
half4  trunc(half4 x);

fixed  trunc(fixed x);
fixed1 trunc(fixed1 x);
fixed2 trunc(fixed2 x);
fixed3 trunc(fixed3 x);
fixed4 trunc(fixed4 x);

Parameters

x

Vector or scalar which to truncate.

Description

Returns the integral value nearest to but no larger in magnitude than x.

Reference Implementation

trunc for a float3 vector could be implemented like this.

float3 trunc(float3 v)
{
  float3 rv;
  int i;

  for (i=0; i<3; i++) {
    float x = v[i];

    rv[i] = x < 0 ? -floor(-x) : floor(x);
  }
  return rv;
}

Profile Support

trunc is supported in all profiles except fp20.

See Also

ceil, floor, round