> Also, GPUs are distinctively not systolic arrays. Hennesy's "Computer Architecture: a quantitative approach" has a good primer on Single instruction Multiple Threads (SIMT) architectures typically seen in GPUs.

Good catch, as pointed out by wikipedia:

"They are sometimes classified as multiple-instruction single-data (MISD) architectures under Flynn's taxonomy, but this classification is questionable because a strong argument can be made to distinguish systolic arrays from any of Flynn's four categories: SISD, SIMD, MISD, MIMD"

https://en.wikipedia.org/wiki/Systolic_array