Having discussed in detail the derivation of one particular block algorithm, we now describe examples of the performance that has been achieved with a variety of block algorithms. The clock speeds for the computers involved in the timings are listed in Table 3.2.
-------------------------------------------
Clock Speed
-------------------------------------------
CONVEX C-4640 135 MHz 7.41 ns
CRAY C90 240 MHz 4.167 ns
DEC 3000-500X Alpha 200 MHz 5.0 ns
IBM POWER2 model 590 66 MHz 15.15 ns
IBM RISC Sys/6000-550 42 MHz 23.81 ns
SGI POWER CHALLENGE 75 MHz 13.33 ns
-------------------------------------------
See Gallivan et al. [42] and Dongarra et al. [31] for an alternative survey of algorithms for dense linear algebra on high-performance computers.