Posted by Jerry Eckert.
Engineer from Looking for a team in Raleigh, NC sponsored by .
Posted on 11/11/99 10:26 PM MST
Does anyone know how the access times for the regular RAM and the scratchpad RAM on the BS-2sx compare?
Thanks.
Jerry
Posted by Joe Johnson. [PICTURE: SAME | NEW | HELP]
Engineer on team #47, Chief Delphi, from Pontiac Central High School and Delphi Automotive Systems.
Posted on 11/12/99 4:11 AM MST
In Reply to: BS-2sx RAM access times posted by Jerry Eckert on 11/11/99 10:26 PM MST:
Jerry,
Very good question. I regret that I don’t know. BUT I can guess. I suppose it is almost the same as the normal RAM on the STAMP2. In any case, It must be many times faster than writing to EEPROM which was our only option if we ran out of variable space last year (and then it was only an option for stuff that changed once or twice per reboot not once or twice per loop – the EEPROMs are only advertized to be writeable a set number of times).
I have another good question, perhaps others have guesses they can provide.
Can RAM in the scratchpad be named like regular RAM (i.e. Variables). EEPROM space can be named (using the DATA command), but access to it is not as easy as RAM. For example, in order to use EEPROM space I would have to READ from a named address and effectively store it in a variable prior to using it in a calculation. Then to replace the value, I would have to WRITE a value to the named address.
It would be a lot simplier if we could treat the scratch RAM more like variables than like named mailboxes.
Any guessers?
Joe J.
Posted by Jerry Eckert.
Engineer from Looking for a team in Raleigh, NC sponsored by .
Posted on 11/12/99 1:23 PM MST
In Reply to: Don’t know But I can guess… posted by Joe Johnson on 11/12/99 4:11 AM MST:
: Jerry,
: Can RAM in the scratchpad be named like regular RAM (i.e. Variables). EEPROM space can be named (using the DATA command), but access to it is not as easy as RAM. For example, in order to use EEPROM space I would have to READ from a named address and effectively store it in a variable prior to using it in a calculation. Then to replace the value, I would have to WRITE a value to the named address.
: It would be a lot simplier if we could treat the scratch RAM more like variables than like named mailboxes.
Joe,
Thanks for the reply, and allow me to return the guess…
I assume you can assign symbolic names to the scratchpad locations by defining symbolic constants
(e.g., SCRATCHPAD_1 CON 1
SCRATCHPAD_2 CON 2
etc.)
It looks to me like their goal was to implement the scratchpad with as few changes to the compiler as possible.
: Any guessers?
: Joe J.