...BLAS¹

from Basic Linear Algebra Subroutines

.
.
.
.

... project²

https://www.cs.utk.edu/~ghenry/distrib/

.
.
.
.

... ATLAS³

https://www.netlib.org/atlas/index.html

.
.
.
.

... semaphores⁴

it can also be done with mutexes.

.
.
.
.

... counter⁵

this is a 64 bits counter which value is incremented each cycle and start counting at power on of processor

.
.
.
.

... implementation⁶

with recent version f77 level 1 blas compete with those of ASCI Red project

.
.
.
.

... processors⁷

an introduction on how to use these counters is available at https://www.cs.utk.edu/~ghenry/distrib/mon_counters

.
.
.
.

... loads⁸

loads necessary to perform computation

.
.
.
.

... loads⁹

loads effectively issued

.
.
.
.

... MESI¹⁰

Modified, Exclusive, Shared and Invalid

.
.
.
.

... misses¹¹

PII and up use a write-back/write-allocate strategy while the PPRO use write-trough.

.
.
.
.

... dimension¹²

distance between first elements of columns.

.
.
.
.

... Phipac¹³

https://www.icsi.berkeley.edu/~bilmes/phipac/

.
.
.
.

... MTL¹⁴

https://www.lsc.nd.edu/research/mtl/

.
.
.
.