Thanks Steve I have never used a cpp or the GNU chain so I am glad you answered the question.
I agree with the comments that the code snippets should be kept separate from the OBEX or it will get too cluttered. So, unless Parallax want to jump in now, I guess the wiki will do. I am not a fan of the wiki because it is more open to abuse than a privately held site like the OBEX.
Yep, however, ours is archived, and does have revision control, with a few of us looking at it. No real harm is gonna get done. I take a full archive, from time to time as well. Actually, it's getting time to do that again.
We have not used it much, but the other thing the Wiki can do is have some basic discussion. Say somebody sees a improvement, or bug, or a use case limit of some kind. Those could be discussed, with that discussion contained with the code.
@Heater, I like code repositories like that. Where? Sourceforge, or something? There is a conflict in my mind over just snippets, and projects and a code repository. Thinking of a greater library that can be pulled when people want it maybe? Best of? Curious about how that could / might work.
Self-contained PASM code fragments are nice for when you want to just add some functionality to your existing PASM code. It's a different story if you want to access to a variety of PASM functions from Spin.
If you want to re-use a cog to call a bunch of PASM functions from Spin, you can simply cognew the code each time of course. But then you get the 512longs * 16clocks of load time, followed by run time, followed by the time spent checking to see if the function is done.
You can also do an overlay loader cog (which can minimize load times, but suffers the same "checking to see if we're done" problem), but of course you have to have specify an interface (which is usually a bit cumbersome) on how to call & use the PASM overlay fragments. I'm attaching a fairly simple implementation of this.
The reason I bring this up is that if there is interest in a repository of PASM overlay functions, it might be nice to adopt a common interface if possible (certainly doesn't have to be mine!) to ease code reuse.
Here's my example code:
{{
Test PFunc - the PASM Functionalizer
Jonathan "lonesock" Dummer
}}
' set for the standard 5MHz crystal...this must be correct for the serial code to work
CON
_CLKMODE = XTAL1 + PLL16X
_XINFREQ = 5_000_000
OBJ
term : "FullDuplexSerial"
pf : "PFunc"
PUB main | i, s, tSpin, tPF
' init my objects
term.start( 31, 30, 0, 115_200 )
waitcnt( (clkfreq >> 2) + cnt )
pf.start
' OK, try testing something
pf.run( @pf_SimpleTest, @i )
term.str( string( 13, 13, "The simple test: 105 =?= " ) )
term.dec( i )
term.tx( 13 )
' Try the 2nd test
i := 254870
's := i
tSpin := -cnt
s := ^^i
tSpin += cnt
tPF := -cnt
pf.run( @pf_TestSQRT, @i )
tPF += cnt
term.str( string( "The sqrt test: " ) )
term.dec( s )
term.str( string( " =?= " ) )
term.dec( i )
term.str( string( 13, "Timing (PF vs Spin): " ) )
term.dec( tPF )
term.str( string( " vs " ) )
term.dec( tSpin )
term.tx( 13 )
'' simple example: write a value into (fake) PAR
DAT
ORG pf#PAR_address
pf_SimpleTest long startvalue-pf_SimpleTest ' length of this routine (excluding this register)
mov value, startvalue
add value, #5
wrlong value, pf#PAR_address
jmp #0 ' done with the function
startvalue long 100
value res 1
'' harder example, take the square-root of an integer
DAT
ORG pf#PAR_address
pf_TestSQRT long mask-pf_TestSQRT ' length of this routine (excluding this register)
rdlong input, pf#PAR_address
mov root, #0 'reset root
:loop or root,mask 'set trial bit
cmpsub input,root wc 'subtract root from input if fits
sumnc root,mask 'cancel trial bit, set root bit if fit
shr root,#1 'shift root down
ror mask,#2 wc 'shift mask down
if_nc jmp #:loop 'loop until mask empty
wrlong root, pf#PAR_address
jmp #0 ' done with the function
mask long $40000000
input res 1
root res 1
And here is the output:
The simple test: 105 =?= 105
The sqrt test: 504 =?= 504
Timing (PF vs Spin): 3696 vs 1680
lonesock: I have posted a fast overlay loader (pasm) in OBEX. It is the basis of ZiCog's non-resident emulation code segments.
Chris_D: SO where are we going to put them? OBEX or Wiki ? I prefer obex so long as we have a separate section called "code snippets" or whatever. We need to ask Parallax to create a new section. Ken/Bump..?
Cluso99, I agree with you, a code snippets section on the object exchange would make a lot of sense. It was the first place I started looking for things like that.
I like code repositories like that. Where? Sourceforge, or something? There is a conflict in my mind over just snippets, and projects and a code repository. Thinking of a greater library that can be pulled when people want it maybe? Best of? Curious about how that could / might work.
Google code is great as a code repository. Of course, somebody has to own it. If a parallax desiginee controled user registration, google code might be an option for non-obex code explorations. Of course, google code is the only one I have experience with, so it might not be the most suitable.
Chris_D, any chance you could host something? Or at least, start collecting some fragments? I've got a few I can send you if you want to start acting as the 'keeper of the code fragments'.
I think it would be best to keep this on the Parallax website where people would naturally go to find such things. BTW, not much creeps me out, but your avatar creeps me out - yikes.
@ Dr_Acula
I have no website or webspace anywhere but I certainly wouldn't mind being the keeper of the code fragments till I can convince Paralllax to provide a proper home. Perhaps we shold post these code-frags here in this thread?
Off topic: cost of goods sold (COGS) : Is it this ? or what does COGs stand for?
I could easily give away some webspace, and also I could try to make a directory of such modules. I do not follow this thread on a regular basis, but what about placing such snippets of code in the second half of a 64k eeprom, like in the protoboard?
Comments
I still think them need be in OBEX but Maybe as it is now OBEX need some extension to separate Propeller TOOL and Other type of Compiler's.
As my opinion are that it is good to have them in one place That have separate entries for type of TOOL's used
Yep, however, ours is archived, and does have revision control, with a few of us looking at it. No real harm is gonna get done. I take a full archive, from time to time as well. Actually, it's getting time to do that again.
We have not used it much, but the other thing the Wiki can do is have some basic discussion. Say somebody sees a improvement, or bug, or a use case limit of some kind. Those could be discussed, with that discussion contained with the code.
@Heater, I like code repositories like that. Where? Sourceforge, or something? There is a conflict in my mind over just snippets, and projects and a code repository. Thinking of a greater library that can be pulled when people want it maybe? Best of? Curious about how that could / might work.
Self-contained PASM code fragments are nice for when you want to just add some functionality to your existing PASM code. It's a different story if you want to access to a variety of PASM functions from Spin.
If you want to re-use a cog to call a bunch of PASM functions from Spin, you can simply cognew the code each time of course. But then you get the 512longs * 16clocks of load time, followed by run time, followed by the time spent checking to see if the function is done.
You can also do an overlay loader cog (which can minimize load times, but suffers the same "checking to see if we're done" problem), but of course you have to have specify an interface (which is usually a bit cumbersome) on how to call & use the PASM overlay fragments. I'm attaching a fairly simple implementation of this.
The reason I bring this up is that if there is interest in a repository of PASM overlay functions, it might be nice to adopt a common interface if possible (certainly doesn't have to be mine!) to ease code reuse.
Here's my example code:
{{ Test PFunc - the PASM Functionalizer Jonathan "lonesock" Dummer }} ' set for the standard 5MHz crystal...this must be correct for the serial code to work CON _CLKMODE = XTAL1 + PLL16X _XINFREQ = 5_000_000 OBJ term : "FullDuplexSerial" pf : "PFunc" PUB main | i, s, tSpin, tPF ' init my objects term.start( 31, 30, 0, 115_200 ) waitcnt( (clkfreq >> 2) + cnt ) pf.start ' OK, try testing something pf.run( @pf_SimpleTest, @i ) term.str( string( 13, 13, "The simple test: 105 =?= " ) ) term.dec( i ) term.tx( 13 ) ' Try the 2nd test i := 254870 's := i tSpin := -cnt s := ^^i tSpin += cnt tPF := -cnt pf.run( @pf_TestSQRT, @i ) tPF += cnt term.str( string( "The sqrt test: " ) ) term.dec( s ) term.str( string( " =?= " ) ) term.dec( i ) term.str( string( 13, "Timing (PF vs Spin): " ) ) term.dec( tPF ) term.str( string( " vs " ) ) term.dec( tSpin ) term.tx( 13 ) '' simple example: write a value into (fake) PAR DAT ORG pf#PAR_address pf_SimpleTest long startvalue-pf_SimpleTest ' length of this routine (excluding this register) mov value, startvalue add value, #5 wrlong value, pf#PAR_address jmp #0 ' done with the function startvalue long 100 value res 1 '' harder example, take the square-root of an integer DAT ORG pf#PAR_address pf_TestSQRT long mask-pf_TestSQRT ' length of this routine (excluding this register) rdlong input, pf#PAR_address mov root, #0 'reset root :loop or root,mask 'set trial bit cmpsub input,root wc 'subtract root from input if fits sumnc root,mask 'cancel trial bit, set root bit if fit shr root,#1 'shift root down ror mask,#2 wc 'shift mask down if_nc jmp #:loop 'loop until mask empty wrlong root, pf#PAR_address jmp #0 ' done with the function mask long $40000000 input res 1 root res 1And here is the output:Jonathan
Chris
Chris_D: SO where are we going to put them? OBEX or Wiki ? I prefer obex so long as we have a separate section called "code snippets" or whatever. We need to ask Parallax to create a new section. Ken/Bump..?
PARALLAX HELP!!!!!!!!!!!!!!
Google code is great as a code repository. Of course, somebody has to own it. If a parallax desiginee controled user registration, google code might be an option for non-obex code explorations. Of course, google code is the only one I have experience with, so it might not be the most suitable.
Chris_D, any chance you could host something? Or at least, start collecting some fragments? I've got a few I can send you if you want to start acting as the 'keeper of the code fragments'.
I think it would be best to keep this on the Parallax website where people would naturally go to find such things. BTW, not much creeps me out, but your avatar creeps me out - yikes.
@ Dr_Acula
I have no website or webspace anywhere but I certainly wouldn't mind being the keeper of the code fragments till I can convince Paralllax to provide a proper home. Perhaps we shold post these code-frags here in this thread?
Chris
I could easily give away some webspace, and also I could try to make a directory of such modules. I do not follow this thread on a regular basis, but what about placing such snippets of code in the second half of a 64k eeprom, like in the protoboard?
These subroutines are tools.
best regards
Stefan