|
|
![[parent]](http://images.planetmath.org:8080/images/uparrow.png) Viewing Message
|
|
|
| ``Re: how is a Hertz energy?''
by Algeboy on 2007-07-02 11:00:12 |
|
| That would make sense. That way one could avoid the usual problem of listing "14 hours on a Sparc, 27 days on a Pentium III..." which doesn't really give you a sense of how much overall computing was done.
I see such lists in computer algebra results some times and it makes me laugh because I know these data will be relevent only for about a year. Using a standardized measure such as sum of the cycles seems to be something you could use as a long term bench mark. You could run the same problem on faster machines which might get the answer quicker in real time because they can fit more cycles in a second, but to determine if you have an actual computational improvement you ought to be able to do it in fewer over all computing cycles, thus a lower Hz measure rather than a lower times measure. This way it is not about fast/slow computers but rather efficient programs.
So using Hertz in this way is somewhat like the concept of polynomial time measurements. These are measures of the number of steps made in the program and are not directly a measure of energy usage as that varries from computer to computer.
Thanks perucho. |
| | [ reply | up | top ] | |
|
|
|
|
|
|
|
