DYNAMICLINEAR Procedure

GPU Statement

GPU <option>;

USEGPU <option>;

The GPU statement specifies that the model perform calculations by using graphics processing units (GPUs). Note: The model reverts to CPU processing for efficiency when the value of the NSIMULATIONS= option in the DYNAMICLINEAR statement is less than 100.

You can specify the following option:

PRECISION=FP32 | TF32 | FP64: specifies the computation precision that the model uses for calculations on GPU hardware. The value must be FP32, TF32, or FP64. For more information, see the section Precision of GPU Computation in PROC DYNAMICLINEAR. By default, PRECISION=FP32.

Last updated: July 09, 2026