My Dilemma
Peter Jakacki
Posts: 10,193
in Propeller 2
Over the last 10 years there have been so many suggestions for features on the planned P2 successor to the P1. With great enthusiasm the forum has pitched in helping Chip with testing and software and design analysis resulting in a really great design for the P2. But that's the problem, it's still a design, it is not a P2.
My dilemma though is this; Do I cave in and throw my ideas and time and testing into the P2 stew with the possible reward that the P2 design actually incorporates some nice features that I'd like to see and know I could use? Sure, I've stocked up on different FPGA boards and ported Tachyon just so I'd have some idea about P2. Or do I refrain from joining in in the vain hope that I will have the possible reward of an actual P2?
Seeing that every new features becomes a monster of a thread and pushes the P2 silicon date further and further into the future, I must weigh that option very cautiously as no matter how good it will be it is of no use to me now, but not joining in seems to make not one iota of difference anyway.
Trouble is I'm too busy designing stuff with the only Prop we've got and I have to make do with it. I don't have the luxury of debating and thinking about features year after year. The thing though is once P2 hits the streets I am so very ready as what I am doing now I will be doing with P2 as well, you know, actually building stuff.
Does anybody actually do stuff with the P1 anymore other than breadboarding and learning? I haven't heard about any commercial projects or designs lately. I know there has been work done on tools to build P1 applications but besides demos where are those applications?
My other options are find a suitable FPGA and soup up a P1V for my needs or go elsewhere.
What do I do?
My dilemma though is this; Do I cave in and throw my ideas and time and testing into the P2 stew with the possible reward that the P2 design actually incorporates some nice features that I'd like to see and know I could use? Sure, I've stocked up on different FPGA boards and ported Tachyon just so I'd have some idea about P2. Or do I refrain from joining in in the vain hope that I will have the possible reward of an actual P2?
Seeing that every new features becomes a monster of a thread and pushes the P2 silicon date further and further into the future, I must weigh that option very cautiously as no matter how good it will be it is of no use to me now, but not joining in seems to make not one iota of difference anyway.
Trouble is I'm too busy designing stuff with the only Prop we've got and I have to make do with it. I don't have the luxury of debating and thinking about features year after year. The thing though is once P2 hits the streets I am so very ready as what I am doing now I will be doing with P2 as well, you know, actually building stuff.
Does anybody actually do stuff with the P1 anymore other than breadboarding and learning? I haven't heard about any commercial projects or designs lately. I know there has been work done on tools to build P1 applications but besides demos where are those applications?
My other options are find a suitable FPGA and soup up a P1V for my needs or go elsewhere.
What do I do?
Comments
Can you do what you need with something else which is on the shelf now?
Can you afford to wait at least a year for P2 in silicon?
Will waiting cause lost opportunities?
If you wait, can you risk the P2 not working in silicon?
Would that time waiting be better spent learning to use, for instance, FPGAs?
How sensitive to component costs are your products?
How critical to your business plans is the P2?
One should never underestimate lost opportunities. Waiting for the 'right' product to be available might mean that whilst your product, when it arrives, will be the best it's not much good if customers are already buying an inferior one. Getting people to change is difficult.
I hope so, I've stayed on board this long rather than slink away quietly like many.
Now if I hear both you and Ken say "we are almost there" then that would inspire me even more.
Now if I hear "we are there"......
Maybe we can start targeting that elusive "Spring date" that Dave Hein was couting down to last year....
Yeah, Ken's not the only one with a "beef"
Recently, I've redesigned my P1-based prototype and discovered that I can get by with a lot less computing power (speed/pins/cogs/memory) than I thought. So, while the P2 moves ahead, I'm moving ahead with a P1 version of my project. I figure it might be ready just about the time that the P2 is ready to sample, and I can decide then whether to release the P1 version or wait for the P2. It's pretty amazing what the P1 can do, even 10 years after its release, especially when its features match the application. Still, I enjoy following the progress of the P2 and will continue to do so for as long as I believe that it is moving forward. Perhaps that's because I have a specific application in mind that can take advantage of part of what it has to offer.
Anyway, thanks for expressing your dilemma. I'd like to say that I know how you feel, but your computing needs and ability are way above mine. But at least I'm moving forward (albeit slowly) with a potential product that is nearly a perfect match for the P1. I know the forum has quieted down a lot. That's a sign that the P2 is overdue, I guess. But I think we'll get at least half the activity back when the P2 is released (though it will take time to ramp back up).
PS: Amusingly, when I checked the spelling of "albeit" (as I wondered if it had a space), it gave this example: "He was making progress, albeit rather slowly." That's what I just said about me! Oh, well, one step at a time. Meanwhile, Chip is chipping away (sorry). Actually, it looks like he's down to the polishing phase. It's been a long road, but it's worth it...because, like the P1, the P2 will allow lots of creative and useful things.
I don't think that your hope for having a P2 in the near-term (say, under a year) is in "vain." Chip has all but wrapped up the design (unless timing or power issues gum up the works). Yes, the LUT stuff has probably added a week to the release, but one or two refinements came as a result. And finishing up the last-remaining details always takes a bit of time. Now as to whether you should share your P2 ideas, of course! I'll admit that I was tempted to say, "No, don't share them if they could significantly delay the P2." But Chip is committed to finishing soon and knows how to parse what can and cannot go into the current design. So, it would be a pity for you to suppress a potentially great idea just because it could delay things. But if your ideas involve totally reworking the busses and so on, just be prepared for Chip to say that the ship has already sailed in terms of bringing on such "cargo." I will admit that I'm somewhat shocked when I see bold ideas like going from 16 cogs down to 8 at this late stage (such a change might be enough to set me adrift), as I feel that it's better if suggestions dovetail with the current design. But you never know when someone's idea might be the way to go or might lead to something else that's good.
I still have the last "Hot" image on my DE2 - I'm ready to test that one also!!!
If you can afford to wait a year or so, then it might be worth it. However if your income depends on winning contracts now then keep going with the P-1 and maybe a ARM or something else.
Product loyalty is one thing, but when that loyalty impacts ones ability to make a living then it should be re-evaluated.
We will most likely test the Smile out of it, bug fix, etc...
It's possible that if the market for your products can stand the expense then putting the whole show, processor and whatever peripheral logic, into an FPGA might be a very good way to go.
FPGA's are not about to go away. There are multiple vendors. At worst if one drops out you have to redesign a board for a different FPGA.
Provided you are not doing anything crazy demanding you have an infinite shelf life for your design effort.
Seems there is more than one answer to this, as you have many angles in this question.
For Hardware that you have to ship tomorrow, then yes, use P1, or P1V, or XYZ.
- but for Software Development, P2 is easily alive enough right now, to test code, as you have already been doing.
Software has long gestation times, and getting code tested now, is vitally important to the critical path of proving the FPGA code.
Tachyon seems a great way to get rapid code coverage, and with a bonus reference point of a wealth of P1 Tachyon code, to confirm operation.
Ken
And we have vetted many ideas. Doing more will cost what we have or pose power and or routing and gate costs.
Ripping it up wont be worth it. Neither will a significant synthesis risk.
I'm presently working on a commercial project which uses the Propeller.
My main source of income is from Propeller programming.
I think the P1 is still alive an well.
I think that it has been 2 years (or more than 1 year and half) that Ken *completely gave up* on P2 development. Two years ago he was involved and he gave some commitment dates and announcements. But given the fact that all due dates and milestones were continually delayed ... he just gave up.
As for commercial products, I know of a few recently new products that use the Propeller as of today and of the 3 embeddd products that WBA is working on, 2 are Propeller based.
A simple 2/3/4 cog 8bit design by Chip would have been a killer for me
Oh yes. One of my first real work projects used five Intel 8085's. Two of which were sharing RAM. The whole thing could only keep up with the process it was controlling with the help of a bunch of counter/timers (How do you calculate standard deviation using counters?)
A Propeller would have done all that in it's spare time.