Getting Started with Conditional Numerical Reproducibility

Developer Reference for Intel® oneAPI Math Kernel Library for C

Download PDF

ID 766684

Date 4/28/2026

Version

Public

Getting Started with Conditional Numerical Reproducibility

Intel® oneAPI Math Kernel Library (oneMKL) offers functions and environment variables to help you get reproducible results. You can configure Intel® oneMKL using functions or environment variables, but the functions provide more flexibility.

The following specific examples introduce you to the conditional numerical reproducibility.

While these examples recommend aligning input and output data, you can supply unaligned data to Intel® oneMKL functions running in the CNR mode, but refer to Reproducibility Conditions for details related to data alignment.

Intel CPUs Supporting Intel AVX2

To ensure Intel® oneMKL calls return the same results on every Intel CPU supporting Intel AVX2 instructions:

Make sure that your application uses a fixed number of threads
(Recommended) Properly align input and output arrays in Intel® oneMKL function calls
Do either of the following:
- Call mkl_cbwr_set(MKL_CBWR_AVX2)
- Set the environment variable:
  
  Linux: export MKL_CBWR = AVX2
  
  Windows: set MKL_CBWR = AVX2

NOTE:

On non-Intel CPUs and on Intel CPUs that do not support Intel AVX2, this environment setting may cause results to differ because the AUTO branch is used instead, while the above function call returns an error and does not enable the CNR mode.

Intel CPUs Supporting Intel SSE2

To ensure Intel® oneMKL calls return the same results on every Intel CPU supporting Intel SSE2 instructions:

Make sure that your application uses a fixed number of threads
(Recommended) Properly align input and output arrays in Intel® oneMKL function calls
Do either of the following:
- Call mkl_cbwr_set(MKL_CBWR_SSE2)
- Set the environment variable:
  
  Linux: export MKL_CBWR = SSE2
  
  Windows: set MKL_CBWR = SSE2

NOTE:

On non-Intel CPUs, this environment setting may cause results to differ because the AUTO branch is used instead, while the above function call returns an error and does not enable the CNR mode.

Intel or Intel-Compatible CPUs Supporting Intel SSE2

On non-Intel CPUs, only the MKL_CBWR_AUTO and MKL_CBWR_COMPATIBLE options are supported for function calls and only AUTO and COMPATIBLE options for environment settings.

To ensure Intel® oneMKL calls return the same results on all Intel or Intel compatible CPUs supporting Intel SSE2 instructions:

Make sure that your application uses a fixed number of threads
(Recommended) Properly align input and output arrays in Intel® oneMKL function calls
Do either of the following:
- Call mkl_cbwr_set(MKL_CBWR_COMPATIBLE)
- Set the environment variable:
  
  Linux: export MKL_CBWR = COMPATIBLE
  
  Windows: set MKL_CBWR = COMPATIBLE

NOTE:

The special MKL_CBWR_COMPATIBLE/COMPATIBLE option is provided because Intel and Intel compatible CPUs have a few instructions, such as approximation instructions rcpps/rsqrtps, that may return different results. This option ensures that Intel® oneMKL does not use these instructions and forces a single Intel SSE2 only code path to be executed.

Intel GPUs

To ensure that selected oneMKL calls return the same results on every Intel GPU, do either of the following:

Call mkl_cbwr_set(MKL_CBWR_AUTO) .
Set the following environment variable:
- Linux: export MKL_CBWR = AUTO
- Windows: set MKL_CBWR=AUTO .

Next Steps

See Setting the Environment Variable for Conditional Numerical Reproducibility for details of specifying the branch using environment variables.
See the following sections in the Intel® oneMKL Developer Reference:
Support Functions for Conditional Numerical Reproducibility for how to configure the CNR mode of Intel® oneMKL using functions.
Intel® oneMKL PARDISO - Parallel Direct Sparse Solver Interface for how to configure the CNR mode for PARDISO.

Product and Performance Information
Performance varies by use, configuration and other factors. Learn more at www.Intel.com/PerformanceIndex . Notice revision #20201201

Product and Performance Information

Performance varies by use, configuration and other factors. Learn more at www.Intel.com/PerformanceIndex .

Notice revision #20201201

NOTE:

Code Examples

Select Your Language

Using Intel.com Search

Quick Links

Recent Searches

Advanced Search

Only search in

Developer Reference for Intel® oneAPI Math Kernel Library for C

Getting Started with Conditional Numerical Reproducibility

Intel CPUs Supporting Intel AVX2

Intel CPUs Supporting Intel SSE2

Intel or Intel-Compatible CPUs Supporting Intel SSE2

Intel GPUs

Next Steps