Martin: Do you have a Propeller GCC board configuration file for your DNA boards?
Yes, I posted them a little higher up in this thread. There's a ticket in the repository about adding the two files to the tree. They're also on the wiki
I've been working with these new boards and I have some questions-
The Windbond chip comes mounted in the U12 socket but looking at the Windbond data sheet they talk about pulling the /Hold and /WE pins high. I look at the schematic for the DNA board and I see that position U9 looks like it is setup for this chip and shares the same Clock/DO/DI pins leaving the CS pin by itself.
So it looks to me that the Windbond chip would be better in the U9 position but can it share the lines with the uSD card and both work ok?
I'd like to experiment with the Windbond chip so am wondering if there is any spin code that I can play with for this chip? Can it be used similar to the uSD card for storing files, data, etc.?
U9 might be better for Spin/PASM, originally that is where it is supposed to be, but also works at quad speed in U12 (but only with PropGCC thus far)
Have you made any attempt to measure the relative performance of the single bit SPI flash chips vs. the quad SPI flash chips? Are the quad chips noticably faster?
I don't have any benchmark numbers, but running the count.bas file on the same chip in both sockets shows about a 2x speed improvement with running quad in U12.
I'd like to experiment with the Windbond chip so am wondering if there is any spin/pasm code that I can play with for this chip? Can it be used similar to the uSD card for storing files, data, etc.?
@Don, I don't see why not, it's just another storage device. Have you checked the OBEX for code?
As for the RTC chip, I had mistakenly thought that code was already in the obex, but it turns out the code (written by Roy Eltham) is in this thread. (also added to the wiki)
I need feedback on the wiki especially the PropGCC section. Is it understandable at all? Is it missing anything? What would you like to see more of? What needs to be clarified?
@Don, It's possible FSRW or KyeFAT could use the Winbond chip by changing the pin assignments. I have no idea, I haven't had time to look further into it.
I need feedback on the wiki especially the PropGCC section. Is it understandable at all? Is it missing anything? What would you like to see more of? What needs to be clarified?
I don't have the Winbond chip, so haven't played with this part.
Curious about the setup - /opt/parallax/propeller-load/.
Might be clearer what's involved.
My install went to
"C:\Program Files\SimpleIDE\bin\SimpleIDE.exe"
(and didn't seem to work anywhere else.)
There are \BIN and \Translation folders under that.
Very impressive to go from concept, through prototype and into production in (relatively speaking) such a short time. There is some hope that the GG USB platform will survive and thrive after all. It remains to be seen if I can get back to my projects that could use this board. Far too much OT and homestead duties during the summer months taking priority.
Just got my Winbond flash chips today and realized they are 8MBit chips not 1MBit chips so I now have 1MByte of code space to work with in PropGCC! That's awesome, should be plenty of room to do most anything I need to get done with the robot. Tried out the chip in U12 today and everything works perfectly and easily. Thanks again for making these up. It's nice to keep using my platform modules with updated board.
Martin: I wanted to let you know that I pushed your dna.cfg and dna-sqi.cfg configuration files to the propgcc repository on Google Code. I also pushed one I made called dna-sqi-nway.cfg that uses an n-way set-associative cache driver. It's still under construction though so it may not perform quite as well as the current direct-mapped cache driver.
What advantages does this new cache driver have, when finished, over the old?
A multi-way cache has the potential of increasing the cache hit rate which should improve performance. However, that has to be balanced against against either smaller cache lines or fewer cache index entries per "way" in order to avoid using up N times as much hub memory for an N-way cache. I'm still playing with those parameters so it remains to be seen whether this new cache driver will improve performance or not.
- The USB connector has been changed to a different, more robust part.
- The microSD slot is now a push-in-push-out type.
- The crystal lead sockets are more compact.
- Pads for a CR12xx battery clip (for the RTC) have been added to the back.
- Reset pin added to the lower headers.
Most of these changes are to consolidate parts in my inventory and make the DNA easier to assemble. The RTC will run for 10 years on a single CR1216 (or CR1225) battery. There are shipping restrictions with lithium batteries so you may have to source the battery locally.
- The USB connector has been changed to a different, more robust part.
- The microSD slot is now a push-in-push-out type.
- The crystal lead sockets are more compact.
- Pads for a CR12xx battery clip (for the RTC) have been added to the back.
- Reset pin added to the lower headers.
Most of these changes are to consolidate parts in my inventory and make the DNA easier to assemble. The RTC will run for 10 years on a single CR1216 (or CR1225) battery. There are shipping restrictions with lithium batteries so you may have to source the battery locally.
Neat board i never knew about! I'd love to get ahold of a relatively cheap board with spi flash! Can't wait to buy one. (What's the price BTW? Couldn't find it being sold anymore)
Neat board i never knew about! I'd love to get ahold of a relatively cheap board with spi flash! Can't wait to buy one. (What's the price BTW? Couldn't find it being sold anymore)
Comments
Yes, I posted them a little higher up in this thread. There's a ticket in the repository about adding the two files to the tree. They're also on the wiki
The Windbond chip comes mounted in the U12 socket but looking at the Windbond data sheet they talk about pulling the /Hold and /WE pins high. I look at the schematic for the DNA board and I see that position U9 looks like it is setup for this chip and shares the same Clock/DO/DI pins leaving the CS pin by itself.
So it looks to me that the Windbond chip would be better in the U9 position but can it share the lines with the uSD card and both work ok?
I'd like to experiment with the Windbond chip so am wondering if there is any spin code that I can play with for this chip? Can it be used similar to the uSD card for storing files, data, etc.?
Thanks.
Don
I don't have any benchmark numbers, but running the count.bas file on the same chip in both sockets shows about a 2x speed improvement with running quad in U12.
As for the RTC chip, I had mistakenly thought that code was already in the obex, but it turns out the code (written by Roy Eltham) is in this thread. (also added to the wiki)
but sounds like it uses Femto Basic. I was hoping to find something similar to FSRW or Kye's FAT object.
http://1mgh.com/dna
@Don, It's possible FSRW or KyeFAT could use the Winbond chip by changing the pin assignments. I have no idea, I haven't had time to look further into it.
I don't have the Winbond chip, so haven't played with this part.
Curious about the setup - /opt/parallax/propeller-load/.
Might be clearer what's involved.
My install went to
"C:\Program Files\SimpleIDE\bin\SimpleIDE.exe"
(and didn't seem to work anywhere else.)
There are \BIN and \Translation folders under that.
The .cfg files are located in c:\propgcc\propeller-load
Very impressive to go from concept, through prototype and into production in (relatively speaking) such a short time. There is some hope that the GG USB platform will survive and thrive after all. It remains to be seen if I can get back to my projects that could use this board. Far too much OT and homestead duties during the summer months taking priority.
Congrat's and great work!
$0
"n-way set-associative cache driver" Will that allow me to use a Frammis array with my Gamblepudding plots?
rev.f
- The USB connector has been changed to a different, more robust part.
- The microSD slot is now a push-in-push-out type.
- The crystal lead sockets are more compact.
- Pads for a CR12xx battery clip (for the RTC) have been added to the back.
- Reset pin added to the lower headers.
Most of these changes are to consolidate parts in my inventory and make the DNA easier to assemble. The RTC will run for 10 years on a single CR1216 (or CR1225) battery. There are shipping restrictions with lithium batteries so you may have to source the battery locally.
Clip: http://www.digikey.com/product-detail/en/0/BAT-HLD-012-SMT-ND
Unfortunately, it's sold out. :-(
Sold here? Can I be first in line?
Just wanted one.