Prop123_A9_Prop2_8cogs_v32h.rbf loaded and working ok.
Cold start
----------------------------------------------------------------
Parallax P2 .:.:--TAQOZ--:.:. V1.0 180526-0250 140F
----------------------------------------------------------------
TAQOZ# dir
.SDSS16G 5D59_6020 PROP 32k 15,189M
PROP $0000_3F80 0
B I n f. o $1D00_3F80 110
S y s t. e $1BC0_3F80 6,619,245
SYSTEM~1 $0000_4040 0
BUGZ $0000_4100 330,112
FOZ $0000_43C0 245,760
INVADE $0000_45C0 83,520
MAZE $0000_4680 127,744
VGA $0000_4780 172,992
ok
TAQOZ# lsio
P:00000000001111111111222222222233333333334444444444555555555566
P:01234567890123456789012345678901234567890123456789012345678901
=:~~~~~~~~~d~~~~~~~~~~~~~~~dd~~~~~~d~~~~~~~~~~~~~~~~~~~~~hhhh~h~ ok
TAQOZ#
Thanks Chip but I can't seem to use the new PNut either, it is still reporting a 1024 limit when I try it on ROM_140F.spin2
The about dialog reports PNut_v32h.exe as v32f
Thanks Chip but I can't seem to use the new PNut either, it is still reporting a 1024 limit when I try it on ROM_140F.spin2
The about dialog reports PNut_v32h.exe as v32f
I see the problem now. It's just an equate change in the compiler source. I won't be able to get to it until I get back tonight. Short label names will buy you more symbols, for now.
Peter,
140i has your 140f taqoz and compiles with pnut. I removed some labels and made others local (beginning with ".") to reduce the labels I use.
Thanks, I've made room for another 128+ bytes while adding extra functionality and also printing the date and time in DIR. More words are device sensitive so that to print a null terminated string we normally supply the address (usually via the symbol) to PRINT$. Now if we have an open file we can supply the offset into that file or print the entire file with "0 SD PRINT$" etc.
I'm doing the same with BACKUP and RESTORE so that you can open a file and specify "0 SD BACKUP" to backup or if you want to access the first hidden 4M of the SD (less the MBR) you specify it by sector, in this case sector 1 to skip the MBR and type "0 SD BACKUP" .
SD - select the SD card
SF - select the Serial Flash
Device sensitive words such as DUMP, PRINT$, CTYPE (but not BACKUP.RESTORE) revert back to normal hub RAM after execution.
I would like to specify file name but code space is very tight to process and cleanup file names and then search whereas the starting sector number is the file handle if you like. You can always create a constant if you like when working off a fixed SD with" $1B780 == BOOTFILE" and then just use it "BOOTFILE FOPEN".
Are there any other features, both to do with the SD card and with TAQOZ that we could include (given the limits). Format might be nice but might not be as useful as some other functions.
I would like to specify file name but code space is very tight to process and cleanup file names and then search whereas the starting sector number is the file handle if you like.
Is this made simpler, if you impose some rules on file names ?
eg I could see a use for (eg) easy loading of future boot-patch files, where giving a name is easier for users to manage.
.... Format might be nice but might not be as useful as some other functions.
Can Format go into the expanded TAQOZ ? - the one too large for ROM that users download ?
To me Format is a blunt hammer, and a code crash might call Format ?
Is the write protect switch on SD cards checked here ?
Thanks Chip. Even though I'm using p2asm as part of my build tool I still need a sanity check on the final output as I have been caught
Can you include the latest ROM into these images? I've got a whole heap of extra stuff in there since a few days ago including all the FAT32 and DIR and FOPEN functions.
I'm still making some changes to the latest version but I will finish this later on as I'm taking the cheese and kisses out to dinner.
Great, that works! It's a pity we didn't have a CLI version of it though as I could easily tie that in with a preprocessor like I do now with p2asm and m4.
Great, that works! It's a pity we didn't have a CLI version of it though as I could easily tie that in with a preprocessor like I do now with p2asm and m4.
I have noticed a difference between the two A9 images when starting TAQOZ.
The BeMicro -A9 starts with all cogs but the P123-A9 starts with 1 cog.
I have verified the cog leds are working correctly on the P123-A9 board with my own code.
I have noticed a difference between the two A9 images when starting TAQOZ.
The BeMicro -A9 starts with all cogs but the P123-A9 starts with 1 cog.
I have verified the cog leds are working correctly on the P123-A9 board with my own code.
My later version of TAQOZ just starts one cog to help with debugging but the user can start other cogs with <cog> NEWCOG.
I changed the way I automount since it would throw an error if no card was present. Now it checks first and ignores automount and the subsequent error if there is no card.
How many clocks does it take for an OUTL #pin to be seen at the output pin from the start (or the end - please specify which) of the OUTx instruction?
How many clocks does it take for a TESTP #pin to sample the input pin prior to the start (or the end - please specify which) of the TESTP instruction? 6 from the start/end???
How many clocks does it take for an OUTL #pin to be seen at the output pin from the start (or the end - please specify which) of the OUTx instruction?
How many clocks does it take for a TESTP #pin to sample the input pin prior to the start (or the end - please specify which) of the TESTP instruction? 6 from the start/end???
Didn't I see some test code for this go past recently ? There are pipeline FF's both ways.
I believe 3 clocks for DIR/OUT to pin, and 2 clocks for pin to IN.
I think cluso wants to pair OUT and TESTP, and the question seems to come to 'how many NOPs are needed to test the new value, not the old one' ?
Is that one NOP or 2 NOPs, or ? , needed ? (and is the Silicon the same as the FPGA at all MHz ? )
I believe 3 clocks for DIR/OUT to pin, and 2 clocks for pin to IN.
I think cluso wants to pair OUT and TESTP, and the question seems to come to 'how many NOPs are needed to test the new value, not the old one' ?
Is that one NOP or 2 NOPs, or ? , needed ? (and is the Silicon the same as the FPGA at all MHz ? )
You can sample after the clock high and low, and the data will still be from before the high.
Chip,
Is it 1 clock after the OUTx instruction completes that the output pin will be set?
And is it 2 clocks prior to the TESTP instruction commences (ie 4 clocks from the end of TESTP) ?
This is what I see, but it's relative, not specific to the start of the instruction.
It requires a waitx #3 to see the output back to input. Waitx #2 doesn't see the output.
Comments
All thee things are going to make the P2 very easy to experiment with right out of the box!
Thanks Chip but I can't seem to use the new PNut either, it is still reporting a 1024 limit when I try it on ROM_140F.spin2
The about dialog reports PNut_v32h.exe as v32f
I see the problem now. It's just an equate change in the compiler source. I won't be able to get to it until I get back tonight. Short label names will buy you more symbols, for now.
140i has your 140f taqoz and compiles with pnut. I removed some labels and made others local (beginning with ".") to reduce the labels I use.
Thanks, I've made room for another 128+ bytes while adding extra functionality and also printing the date and time in DIR. More words are device sensitive so that to print a null terminated string we normally supply the address (usually via the symbol) to PRINT$. Now if we have an open file we can supply the offset into that file or print the entire file with "0 SD PRINT$" etc.
I'm doing the same with BACKUP and RESTORE so that you can open a file and specify "0 SD BACKUP" to backup or if you want to access the first hidden 4M of the SD (less the MBR) you specify it by sector, in this case sector 1 to skip the MBR and type "0 SD BACKUP" .
SD - select the SD card
SF - select the Serial Flash
Device sensitive words such as DUMP, PRINT$, CTYPE (but not BACKUP.RESTORE) revert back to normal hub RAM after execution.
I would like to specify file name but code space is very tight to process and cleanup file names and then search whereas the starting sector number is the file handle if you like. You can always create a constant if you like when working off a fixed SD with" $1B780 == BOOTFILE" and then just use it "BOOTFILE FOPEN".
Are there any other features, both to do with the SD card and with TAQOZ that we could include (given the limits). Format might be nice but might not be as useful as some other functions.
Is this made simpler, if you impose some rules on file names ?
eg I could see a use for (eg) easy loading of future boot-patch files, where giving a name is easier for users to manage.
Can Format go into the expanded TAQOZ ? - the one too large for ROM that users download ?
To me Format is a blunt hammer, and a code crash might call Format ?
Is the write protect switch on SD cards checked here ?
Here it is right here:
https://drive.google.com/file/d/1asHZjmC3Q7e2aQcI42S8R1BdlxUjuY22/view?usp=sharing
Can you include the latest ROM into these images? I've got a whole heap of extra stuff in there since a few days ago including all the FAT32 and DIR and FOPEN functions.
I'm still making some changes to the latest version but I will finish this later on as I'm taking the cheese and kisses out to dinner.
Yes, Chip needs to test it on the latest rom files.
Okay. Sorry. I found the part that matters and fixed it. We can now do 10k DAT symbols:
https://drive.google.com/file/d/1JukT7CKEOJLygqs1GefDUoLerWiOgWpj/view?usp=sharing
CLI. Will work on that....
Thanks for trying it, Brian and Peter.
The BeMicro -A9 starts with all cogs but the P123-A9 starts with 1 cog.
I have verified the cog leds are working correctly on the P123-A9 board with my own code.
My later version of TAQOZ just starts one cog to help with debugging but the user can start other cogs with <cog> NEWCOG.
This is the current boot splash.
and performing a DIR from there:
Sorry to ask again...
How many clocks does it take for an OUTL #pin to be seen at the output pin from the start (or the end - please specify which) of the OUTx instruction?
How many clocks does it take for a TESTP #pin to sample the input pin prior to the start (or the end - please specify which) of the TESTP instruction? 6 from the start/end???
eg
clock /\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\ | : | : | : | : | : | : | : | : | : | : | : | : | : | : | : | : | : | : | out tst ~~~~~|~~~~~ ~~~~~~|~~~~~~~~~~~~~~~~Ugh. The forum doesn't accept the proptool extended charactersDidn't I see some test code for this go past recently ? There are pipeline FF's both ways.
I find this post April 26, added 1 clk delay
I think these tests relate to SmartPins, and show some variance ?
I think cluso wants to pair OUT and TESTP, and the question seems to come to 'how many NOPs are needed to test the new value, not the old one' ?
Is that one NOP or 2 NOPs, or ? , needed ? (and is the Silicon the same as the FPGA at all MHz ? )
You can sample after the clock high and low, and the data will still be from before the high.
Is it 1 clock after the OUTx instruction completes that the output pin will be set?
And is it 2 clocks prior to the TESTP instruction commences (ie 4 clocks from the end of TESTP) ?
This is what I see, but it's relative, not specific to the start of the instruction.
It requires a waitx #3 to see the output back to input. Waitx #2 doesn't see the output.
mov cmdout,#10 .loop drvl #0 waitx #32 outh #0 waitx cmdout testp #0 wc mov lmm_x,cmdout call #@_hubhex8 mov lmm_x,#" " call #@_hubTx mov lmm_x,#"0" if_c mov lmm_x,#"1" call #@_hubTx call #@_hubTxCR djnz cmdout,#.loopclock /\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\ | : | : | : : | : | : | : | : | : | : | : | : | : | : | : | : | : | : | out waitx #3 tst ~~~~~~~~~~|~~~~~~~~~~~~clock /\/\/\/\/\/\/\/\/\/\/\/ | : | : : : : | : | out waitx #3 testp ~~~~~~~~~~~|~~~~~~~~~~~ clock /\/\/\/\/\/\/\/\/\/\/\/ | : | : : : : : | : | out waitx #4 testb ina/inb ~~~~~~~~~~~|~~~~~~~~~~~Note that TESTP needs one less clock than reading the INA/INB register.