The latest article in IBM's Programming high-performance applications on the Cell BE processor is online. Part 4 is titled "Program the SPU for performance". The introduction:

Write optimal code for the Cell Broadband Engine? (Cell BE) processor's synergistic processing unit (SPU) and have your programs running lightning fast. This installment of Programming high-performance applications on the Cell BE processor covers SIMD vector programming, branch elimination, loop unrolling, instruction scheduling, and branch hinting techniques. Previous installments have covered the basics of the Sony? PLAYSTATION? 3, the Cell BE architecture, and SPU programming.

Available at www-128.ibm.com/developerworks/power/library/pa-linuxps3-4