IPP Errata
----------

* Chapter 6. The check on the size of the input value n was incorrect in the Get_args functions in cuda_vec_add.cu and cuda_vec_add_cp.cu. Both files have been up- dated. Thanks to Xianyu, Zhu for spotting these errors. (March 7, 2023)
* In Section 2.3.5, pp. 46-47, Prof. Mohammed Zahran wrote that snooping cache coherence is rarely used anymore.  The reason is that busses aren't scalable, and that broadcasts on other types of interconnects are expensive. Thanks to Prof. Zahran for this comment.  (October 30, 2023)
* In Section 6.11.4, page 325, of "An Introduction to Parallel Programming, 2nd Edition", the code at the bottom says "__shfl_down_sync_sync ...", when it should be "__shfl_down_sync". Thanks to Prof. Eleazar Leal for spotting this error. (April 22, 2023)