Boston Linux & Unix (BLU) Home | Calendar | Mail Lists | List Archives | Desktop SIG | Hardware Hacking SIG
Wiki | Flickr | PicasaWeb | Video | Maps & Directions | Installfests | Keysignings
Linux Cafe | Meeting Notes | Blog | Linux Links | Bling | About BLU

BLU Discuss list archive


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

To swap or not to swap that is the question



I haven't done the re-scaling of this analogy, but to many years ago I
took a performance class.  The instructor used an illustration that if
you assume that moving stuff from one register to another took one
instruction cycle, then to scale it up to 1 second per instruction cycle
to help give you a sense of scale:

Register to register:	1 second
Register to cache:	10 seconds
Register to memory:	1 minute
Register to drum:		8 hours
   (sometimes implemented as a head per track disk, old technology now,
	often used for virtual memory back then)
Register to disk:		2.8 days

This is from memory, but it gives the scale of things.  I wonder what
the comparison would be today for a reasonably fast machine, like a 2GHz
CPU? (I'm a little to out of it to look up the info now days)

Someone interested?

Might want to add some relative information for:
	UPB Flash key fob
	CDRom
	Disk 7200RPM 8mSec ATA100
	Disk 10000RPM 6mSec ATA100
		Average rotational delay and seek time matter in
addition to the		speed of data transmission
	Disk 10000RPM Ultra160 SCSI
	Gigabit attached NAS
	SAN?

Inquiring minds want to know :)
 





BLU is a member of BostonUserGroups
BLU is a member of BostonUserGroups
We also thank MIT for the use of their facilities.

Valid HTML 4.01! Valid CSS!



Boston Linux & Unix / webmaster@blu.org