When was the P2 released?
samuell
Posts: 554
in Propeller 2
Hi,
I've been away from the forums and haven't followed the P2 development. I was surprised to see that the P2 was on sale on Mouser. Probably, this has been answered before, but when was the P2 released?
Kind regards, Samuel Lourenço
Comments
There wasn't really a big announce I don't think. Proptool is only just getting finished now, and docs still have a long haul ahead.
EDIT: Silicon was on schedule for release around 12 months back but Covid-19 hit and deliveries went batty worldwide. Not sure when Parallax actually received the first bulk shipment.
Fortunatelly, this information is on V35 docs (page 4):
"2020_06_01 Received 7,000 Rev C chips from ON Semi."
Thanks @Yanomani . However, were those Rev C chips considered final, or engineering samples? I ask this, because the one I've bought from Mouser has no "ES" suffix.
Hi @samuell
Mouser Global product page for the P2X8C4M64P currently displays the following image:
https://mouser.com/images/marketingid/2020/img/106066264.png?v=010421.0540
which matches a non-ES, Rev. C silicon produced at Philippines during the year 2020, based on the marked lot code (LFX2042).
Again (and for historical tracking of versions/lots), based on V35 docs (page 01), there were REV C chips (currently accepted as final) that yet shows the ES appended to the part ID, marked with lot code LHU2019. If they were Rev. B (ES), the lot code would show LPD1941.
As usual, a fresh picture taken from the device you received, would help clarify things a litle bit.
Thanks! No photo needed. My sample has the lot code you referred, exactly. It is LFX2042. So, I would assume it would be the 42th week of 2020?
Edit: Released in October 2020?
Adding to my previous comments, I've decided to piece together some information in order to correlate and determine the release date. So:
If anyone could fill in the blanks about Rev. C, I would appreciate it. I don't have a Rev. C board, nor I can't find any detailed photos. The chip seems to be an engineering sample, though, because the letters seem to be there in the few photos I could find of the chip.
Nevertheless, I think I have enough data to correlate. So, I could assume that the first production chips were dated from October 2020, received in January 2021, and released after. I would need this info for an article on my blog. However, if not possible, the lot date is very telling.
Anyway, @cgracey, could you help me?
A little more info...
Rev A ES chips - there were 10 globtops which were put into the QFP carriers by OnSemi (marked PCAE1838D 22477-900) and some of these were placed on P2D2 boards. There were ?? chips packaged properly. Some of these were used in the P2-EVAL Rev A boards.
Note: Rev B chips have timing changes, silicon changes including instruction differences.
Rev B ES chips - there were some globtops by OnSemi sent to Parallax (around 10? chips IIRC) which preceeded the ones which were packaged properly (IIRC around 160) and marked LPD1941. IIRC there were 100 of these placed into P2-EVAL Rev B boards.
Note: Rev C chips are Rev B chips with a metal layer modification to cut the track to the second ADC pin to reduce ADC noise. Otherwise the silicon is identical to Rev B silicon.
REV C ES chips - I am unsure if there were any globtops. I cannot recall how many there were of these (mine are marked LHV2019 - there may have been two batches). These are the same as the final production Rev C chips.
Rev C production chips - Same as REV C ES chips produced thru the OnSemi production plant.
I received a Prop123 FPGA board in November 2015,
an Eval Board with Rev A ES (LWB1843) in December 2018,
an Eval Board with Rev B glob-top (PCAL1930D) in September 2019,
and an Eval Board with Rev B ES (LPD1941) in November 2019.
PS: Here's a photo of my half knackered glob-top (that is still my main testing unit):
Oh man, what did you do to it? That's brutal!
There was no Rev C glob-top. On Semi required an ES run before production, so Rev C ES (LHV2019) would have arrived, presumably June 2020, late due to Covid-19. And production (LFX2042) in October 2020.
Production is release, effectively. It just wasn't a big flashy launch.
Hehe. There may have been prior electrical damage but I think the main culprit was cooking it in the kitchen oven to desolder one component. Browned the whole board with that! Since then I've discovered various faulty I/O. The messiness is me attempting to recover certain I/O that I thought may have lifted off the board. I don't think any of that worked though, the faults must be deeper in the IC. I just avoid using those pins now.
From memory:
Do I remember that correctly?
It sure deserved such a launch. The P2 is a piece of engineering and architecture that many companies would be proud of. The silicon shows quality.
Anyway, I concur that production is release, in this particular case. The silicon was pretty much tested at that point so, presumably, it would be a matter of receiving them and then sending them away.
Regarding your first point, definitely correct. Power gating was implemented and Rev. B samples were less power hungry. About your second point, I have a vague idea of that being discussed.
Both points are correct.
There'll be more fanfare later I'd say. Tools are nearly done I believe. But more docs to do, the silicon google doc is getting messy with disjointed contributions. Doing a big push without good docs would be a mistake.
P2X8C4M64P Approved for Production in November 2019 - https://forums.parallax.com/discussion/170809/p2x8c4m64p-approved-for-production/p1
That was straight after Rev B ES was shipping. Rev C ES was still 6 months out, and production chips nearly a year.
About tools. A C compiler that is able to run under Linux as well as Windows is a must for me. I think Parallax should aim for those users that are experienced in C and don't want to learn Spin for reasons. For instance, my reason for not wanting to learn Spin is that it is an interpreted language, and therefore not as fast as C could be. I prefer to deal with low level stuff any time. So Spin is not worth learning for me, regardless of being a nice language.
There is a C compiler called FlexC that works on Windows and Linux and MAC. There is also FlexSpin which compiles Spin to PASM code and is therefore not interpreted. These and more is available in Flexprop:
https://forums.parallax.com/discussion/164187/flexspin-compiler-for-p2-assembly-spin-basic-and-c-in-one-compiler#latest
Andy
Thanks! I've read about, and used, FlexGUI and Spin 2 GUI. But never tried Flex C.
Pssst! Flex does a killer version of BASIC too.
C will be far more effective on the Prop2 than the Prop1 both because of hubexec and the larger hubRAM. That said, Chip's Spin2 is also several times faster on the Prop2. So, when memory is tight, it'll still have the upper hand over compiled solutions.
Is there any GUI for FlexC? I guess I could build a GUI frontend based on Qt. It is nothing that I'm not experienced with. Basically, input a file path, some parameters, and the GUI would run the commands. I would need to know what steps to do, though.
Also, is there any chance of placing the commands inside the "Build" folder into "/usr/local/bin"? I tried that, but it breaks the shell scripts.
I've not tried to work that out. I'm happily using flexspin and loadp2 in a bash shell. But then, I'm only doing pasm coding as well.
Thanks Evan, for your response. I could have that working. I only need a crash course on how to compile using different commands. Then I could analyse the type of file to be compiled and have the program running the sequence of commands accordingly.
You can see the application below as an example. It runs commands according to the user input, displaying the output as well as any error messages, since it reads from both streams.
Here's an example one-liner I use for assembling and downloading. All I do is change the filename for different programs.
filename=adc-to-millivolts5; echo; flexspin -g -l -2b ${filename}.spin2; echo; loadp2 -p /dev/serial/by-id/usb-Parallax_Inc_Propeller_P2-ES_EVAL_P23YOO42-if00-port0 ${filename}.binary -v -t -b230400 -SINGLEFlexProp is the "native" GUI for FlexC. Someone (I think @pullmoll ) has done a Qt version of FlexProp. There's also Visual Studio Code, which @"Stephen Moraco" has configured for Spin support... not sure if that works for BASIC and C as well, but there's no reason it couldn't.
Evan, I'm curious. Why do you need the "echo"? Are you leaving an empty line between each command?
By the way, I can have the application replacing the filename on the fly. So the application will only need to run two commands.
yep, adds separation for readability.
FlexGui is a simple IDE using QT and can compile C, Basic and Spin on linux, windows, mac.
You can mix languages and use objects fro any of them in any other of them.
youtube video here
Enjoy!
Mike
I need a solution that works on Linux. So, I guess that I'll make one. The commands already exist, and a GUI that invokes them is a simple task.