Let's just figure it out!
potatohead
Posts: 10,253
After reading through the discussions, I think we all agree on the fact that we need to produce what will fund our futures. I don't know what that is. I don't care what that is either. Just that it happens.
Really, the big thing for me is choking on the idea of letting go of projects and thoughts in progress with the FPGA P2, and to some degree the FPGA development project continuing overall.
I like the FPGA work and have learned a TON doing it with everybody trying to get to the next viable, possible, practical design. The next step is programming the FPGA and learning on that level. I'm very much into this aspect of things.
I also really want to make some stuff that isn't quite appropriate for the P1.
Factoring all of that down for me means:
1. Get the chip sorted. We get one, and it needs to fund the future. Whatever that is, I'm excited about it.
2. Continue on in FPGA land, but with more consideration given to power / process realities.
3. I want to know what to code for. Right now, in my spare time, I am really enjoying some coding on the FPGA. Of course, that only makes sense if there is a real end to things, as in a chip.
I think that means the more P1 like COG design in the near term. Because we will be integrating selected P2 features prototyped already, plus the killer I/O pins and fuses, we need to do FPGA work to validate all of that and maximize it. This process favors that P1 like COG design.
So let's get it sorted.
And it will mean continuing on with the P2 design idea, maybe in parallel, and don't call it P2. Since the timeline will very likely change, I think we should blow it out and really think about a System On Chip oriented design on a 3-5 year timeline. Something else will end up being P2, or the P1 COG like design will be P2.
The FPGA work gets pointed toward, Parallax System On A Chip.
Take all that great stuff we learned and go that direction, as there is a clear trend alignment for it, and considerable time window for profit on it all too. Chip has a vision on this many of us share. It is a compelling vision, and we can realize it in FPGA for now, and I for one really don't want to let it die.
The time window for a P1 variant is smaller, but we need funding to go the longer haul. This all makes great sense to me, and I feel good about the possibilities having gotten over the shock of it all! (good grief, that's a painful thing, don't you agree!)
Seems to me, this path does a few things. It gets all the P1 users something they really need; namely, an upgrade. Gotta get that done. Seriously. It also doesn't waste the development, nor the work setting up on how we do FPGA oriented things. Ken has boards and a program in the pipe. That shouldn't fall off the map. Too much good there.
And it lets us take the current design for the P2, expand it some, target a more appropriate process, and nail it cold. Maybe even quickly, keeping the process considerations in mind this time, so that we end up with something viable and we know it. To me, that's the lesson learned. Primary one, among many others.
If doing that takes 6 more years, it's gonna suck some, but it will suck less with a new chip to jam on in the interim.
Take this thread to express your macro level thoughts. Tech details, wants, needs, go in the ones we've got going already. Just how are you getting your head wrapped around what happened? Fun times everyone. Or, ignore it and carry on. I just wanted to express a few things bubbling around in my head.
Really, the big thing for me is choking on the idea of letting go of projects and thoughts in progress with the FPGA P2, and to some degree the FPGA development project continuing overall.
I like the FPGA work and have learned a TON doing it with everybody trying to get to the next viable, possible, practical design. The next step is programming the FPGA and learning on that level. I'm very much into this aspect of things.
I also really want to make some stuff that isn't quite appropriate for the P1.
Factoring all of that down for me means:
1. Get the chip sorted. We get one, and it needs to fund the future. Whatever that is, I'm excited about it.
2. Continue on in FPGA land, but with more consideration given to power / process realities.
3. I want to know what to code for. Right now, in my spare time, I am really enjoying some coding on the FPGA. Of course, that only makes sense if there is a real end to things, as in a chip.
I think that means the more P1 like COG design in the near term. Because we will be integrating selected P2 features prototyped already, plus the killer I/O pins and fuses, we need to do FPGA work to validate all of that and maximize it. This process favors that P1 like COG design.
So let's get it sorted.
And it will mean continuing on with the P2 design idea, maybe in parallel, and don't call it P2. Since the timeline will very likely change, I think we should blow it out and really think about a System On Chip oriented design on a 3-5 year timeline. Something else will end up being P2, or the P1 COG like design will be P2.
The FPGA work gets pointed toward, Parallax System On A Chip.
Take all that great stuff we learned and go that direction, as there is a clear trend alignment for it, and considerable time window for profit on it all too. Chip has a vision on this many of us share. It is a compelling vision, and we can realize it in FPGA for now, and I for one really don't want to let it die.
The time window for a P1 variant is smaller, but we need funding to go the longer haul. This all makes great sense to me, and I feel good about the possibilities having gotten over the shock of it all! (good grief, that's a painful thing, don't you agree!)
Seems to me, this path does a few things. It gets all the P1 users something they really need; namely, an upgrade. Gotta get that done. Seriously. It also doesn't waste the development, nor the work setting up on how we do FPGA oriented things. Ken has boards and a program in the pipe. That shouldn't fall off the map. Too much good there.
And it lets us take the current design for the P2, expand it some, target a more appropriate process, and nail it cold. Maybe even quickly, keeping the process considerations in mind this time, so that we end up with something viable and we know it. To me, that's the lesson learned. Primary one, among many others.
If doing that takes 6 more years, it's gonna suck some, but it will suck less with a new chip to jam on in the interim.
Take this thread to express your macro level thoughts. Tech details, wants, needs, go in the ones we've got going already. Just how are you getting your head wrapped around what happened? Fun times everyone. Or, ignore it and carry on. I just wanted to express a few things bubbling around in my head.
Comments
If you give me a bigger, better P1B or a P2/P3, I will always find a way to max it out. If it does not have some technological feature, I will always find a work around, or add another IC. Who knows what new technology will be around in 5 years, 10 years.
My thing as a software engineer is always, how to solve a customer problem. How can I make an app that will solve a medical problem, make a new discovery, make life better.
So, lets say you have a quadcopter, or robot. Can I use that for ocean rescue, check water /sewer lines for breaks, help as a sentry for a military operation. A thousand other ideas. Look around and if you don't see a problem, you are not looking hard enough. Take that knowledge and technology and talent, and don't let it go to waste. Imagine. Invent. Grow.
-Phil
Many of us use the terms "we" and "us" because Parallax have allowed us to partake in discussions to the point we "feel" included. These are just semantics and I am sure Chip, Ken and Parallax take no offence to us using those terms.
All we can do is give them our opinions and thoughts, nothing more.
I know. It's just that so much energy, sweat, tears, sturm und drang -- whatever you want to call it -- has been expended here that could have more profitably been put to use doing neat stuff with the P1. It seems that the forum has become more about what we want than what we already have. This used to be a happy place and, to me, happiness isn't about getting what you want, it's about appreciating what you have and not wanting what you don't -- or can't -- have.
-Phil
That I have placed interest in P2 and am interested in it's development takes nothing away from P1. Really, a lot of that P1 work / fun wasn't shared here because it didn't need to be. Secondly, I have done some P1 work I chose not to share due to some of us here thinking it's more important that my work comply with some vision or other they had, which complicated just getting something done, was less enjoyable, and or messed with simply learning or doing something for the hell of it.
If you want to open that door, I'm not entirely sure you won't regret it. Truth is you could Phil. There are lots of reasons why people do what they do, and frankly it's selfish to expect everybody to maximize the P1 when it's a big world out there. Realistically, we should be so lucky as to get the time we do.
Edit: Regarding iterations, blood sweat and such. When Chip was working on P1, he had a difficult time being alone on it, frequently getting negative reactions. This was painful, and he almost didn't finish. Ask him.
He didn't want that on P2, and is sharing his personal explorations with the rest of us, finding it stimulating and motivating. That's cool! And it's not a waste at all. Just so you know.
I guess we'll just have to agree to disagree. I really don't think we're "helping" in the least. The history of the last several months only proves to me that we're in the way and clouding whatever singular vision Chip might have held for the P2 -- and now P1+. The list of wants from Parallax's volume customers that Ken posted should be engraved in stone. It's really the only thing that matters. The details of how to get there should be left to Parallax.
-Phil
BTW, potatohead, I'm not singling you out for anything. Seriously. I think we're all to blame for the waste heat and entropy surrounding what should be a rather more keenly focused effort.
-Phil
Here's the deal. If you really don't think it's appropriate, go and talk to Chip and Ken and have them close the door for the greater good. If they do that, I'm OK with it. I like Parallax, and I like my peers here. I want the greater good. Always.
But I'm not gonna own doing that and neither is anybody else with any measure of self-respect. That's what I pushed back on. I hear your point of view, and it's good advocacy. Go get after it. Maybe it's the right thing to do.
Honestly, I'm not so sure a more focused effort was possible, and I mean that, due to how Chip works and the overall dynamics inherent in that, Parallax, all of "us" here, etc... Go and find out if it was. Again, you will find I actually advocated that multiple times, and after having conversations with people who I don't choose to name here, found out it was not gonna go down that way. Ever.
So I did the work, then chose to partcipate as I thought made sense and added value. That is all any of us did.
EDIT: And I get it! If you are not singling me out, then you are expressing your macro level thoughts. No worries here. Do I have it right? I may go back and edit some. May have read you wrong.
Yep, I made the edits. Better discussion this way. No worries here Phil.
I think you are projecting yourself and your feelings onto Chip and Ken and whatever else. Chip ASKED for the input. Chip LOVES it when we suggest things that he didn't think of and likes it when it makes the P2 more interested or better. Chip decided he didn't want to work in isolation. So stop saying he should go off on his own and ignore the people he ASKED to give input! Seriously, are you just ignoring what Chip posts and says? I've been to Chip's place and talked to him in person about this stuff, and he wants to do it this way.
You imply that the forum posters are to blame for the P2 being in the predicament it's in, but you are wrong. It's where it is because Chip chose to go down the path that led to here. Chip is still learning how to make chips the new way and learning what can and can't be done with the process. Also, you know what, Chip will be the one that figures out how to get the P2 to work in a way that he is happy with... and that includes collaborating with US.
So, once again, stuff it with the naysaying BS about US and them. Join the collaboration and follow Chip's lead! You'll be less grumpy if you just give in...
Now, we were invited to contribute ideas, say what we wanted in a device, and so on. Remember where it all started with Chip asking "Would you like 16 COGs or more 256K RAM" so many years ago I have forgotten when?
However this in now driving me nuts. I don't know how Chip keeps a cool head. Since the announcement of the power consumption issues we have seen seen hundreds, maybe thousands of suggestions posts. A million ideas put forward, all kinds of suggestions and tweaks. It's impossible to keep up. It seems impossible to even go back to the simpler P1 style COGs without a hundred suggestions as to how fix and improve it. There are whole new threads being stated on these suggestions.
And its not just the technical issues being advised on, it's the whole Parallax business model and marketing approach.
It's all looking like madness from the outside.
I have faith that Chip will keep a cool head and home in on a solution. Perhaps that month in Mexico was the best suggestion to come out of this so far.
P.S. It's kind of frightening that the "More COGs or more RAM" question from way back is essentially still with us here.
Its scary to read back over some of the early threads, eg this one from Sleazy G back in late 2007
Perhaps the ultimate answer is to produce two chips,
"Propeller - CogMAX" with >32 cogs, or as many as can be packed in, plus
"Propeller - RAMjam" with as much memory as will fit around 8 cogs.
Or perhaps the ultimate answer, where the cells of the sram have distributed logic nestled between them, and cogs and ram are one and the same
16 cores, 32 GFLOPS, 2 watts. Scales up to 64 cores so far.
Not that we should be comparing a Propeller to such a thing.