fp_contract pragma
Determines whether floating-point contraction takes place. A floating-point contraction is an instruction such as Fused-Multiply-Add (FMA) that combines two separate floating point operations into a single instruction. Use of these instructions can affect floating-point precision, because instead of rounding after each operation, the processor may round only once after both operations.
Syntax
#pragma fp_contract ({on|off})
Remarks
When you use the default compiler options, fp_contract is off, which tells the compiler to preserve individual floating-point instructions. Set fp_contract to on to use floating-point contraction instructions where possible. This behavior is new in Visual Studio 2022 version 17.0. In previous compiler versions, fp_contract defaulted to on.
When you use the default compiler options, fp_contract is on. This setting tells the compiler to use floating-point contraction instructions where possible. Set fp_contract to off to preserve individual floating-point instructions. In Visual Studio 2022 version 17.0 and later, fp_contract defaults to off.
For more information on floating-point behavior, see /fp (Specify floating-point behavior).
Other floating-point pragma directives include:
Example
The /fp:fast compiler option enables contractions by default, but the #pragma fp_contract (off) directive in this example turns them off. The code generated from this sample won't use a fused-multiply-add instruction even when it's available on the target processor. If you comment out #pragma fp_contract (off), the generated code may use a fused-multiply-add instruction if it's available.
// pragma_directive_fp_contract.cpp
// On x86 and x64 compile with: /O2 /fp:fast /arch:AVX2
#include <stdio.h>
// remove the following line to enable FP contractions
#pragma fp_contract (off)
int main() {
double z, b, t;
for (int i = 0; i < 10; i++) {
b = i * 5.5;
t = i * 56.025;
z = t * i + b;
printf("out = %.15e\n", z);
}
}
out = 0.000000000000000e+00
out = 6.152500000000000e+01
out = 2.351000000000000e+02
out = 5.207249999999999e+02
out = 9.184000000000000e+02
out = 1.428125000000000e+03
out = 2.049900000000000e+03
out = 2.783725000000000e+03
out = 3.629600000000000e+03
out = 4.587525000000000e+03