Skip to main content

Unit 3.2.5 Blocking for the L1, L2, and L3 caches

The blocking for the various memory layers is captured in the following figure:

Figure 3.2.9. Illustration of the five loops around the micro-kernel. PowerPoint source for figure.

PowerPoint source used in video.

Using our prior naming convention, which of the implementations

  • Gemm_IJP_JI_MRxNRKernel.c

  • Gemm_JPI_JI_MRxNRKernel.c

  • Gemm_PJI_JI_MRxNRKernel.c

best captures the loop structure illustrated in Figure 3.2.9?