Shop OBEX P1 Docs P2 Docs Learn Events
We're looking at 5 Watts in a BGA! - Page 37 — Parallax Forums

We're looking at 5 Watts in a BGA!

1313233343537»

Comments

  • tonyp12tonyp12 Posts: 1,950
    edited 2014-04-07 14:42
    >And I like You answer to

    Sapieha, are you still feeling side effects from stuff they put in to you at the Hospital?
    Your English used to be better= din engelska brukade vara b
  • SapiehaSapieha Posts: 2,964
    edited 2014-04-07 14:45
    Hi tonyp.

    Them don't give me that drugs.

    BUT still I are missing some oxygen --- That give problems to thinking


    tonyp12 wrote: »
    >And I like You answer to

    Sapieha, are you still feeling side effect from stuff they put in to you at the Hospital?
    Your English use to be better= din engelska brukade vara b
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2014-05-16 17:24
    cgracey wrote: »
    What we can do at this point:

    A) Continue on the current trajectory.

    B) Pare down the design, jettisoning all kids of features, like hub exec, and get back to something much smaller and maybe faster. The current FPGA implementation could be further honed and later, hopefully, made into a chip using a smaller technology.

    C) Drop the current design to four cogs, which would also reduce cache sizes and hub memory down to 128KB. This would also allow us to shrink the die considerably, as we could change the I/O pin aspect ratio to allow them to fit together more densely, occupying more of what was needed for the core. This would also mean the whole chip would fit on an FPGA.

    D) Retire to an opium den.

    E) Other ideas?

    I think the D option is in order but there is no need to retire to one as "feature creep euphoria" has proven itself a more powerful opiate and the forum has proven itself as a very noisy and crowed den indeed.
  • Bill HenningBill Henning Posts: 6,445
    edited 2014-05-16 17:34
    It should be pointed out that if there had not been a panic over power envelope, we would be getting a shuttle run right about now.
  • bruceebrucee Posts: 239
    edited 2014-05-16 17:43
    It should be pointed out that if there had not been a panic over power envelope, we would be getting a shuttle run right about now.

    And that shuttle run would have supplied all the people hanging out here at the forum.

    But that would also probably be enough parts to cover all Parallax P2hot sales for the life of the product. And would have been a pretty colossal failure. A 5W part with the P2 specs won't find many takers today, maybe in 2000 it might have found some.
  • potatoheadpotatohead Posts: 10,253
    edited 2014-05-16 21:33
    Man, I don't know. Even at 60 - 80Mhz, that design could do a lot. And I do think it would never have hit 5W. Probably 3, running hard at a modest clock.

    Hope we get to revisit it.

    But we are on this one. And I'm liking it overall.

    As far as the time goes, I'm frustrated with waiting. I think everybody is. Chip wants it too. But, I've done other things in the interim.

    Really, I think the world of Chip. And he means well too.

    Here is something I believe: Chip thinks holistically and in layers. I've never met anyone quite like him, and when talking with him, you come to realize he's breaking things down, while at the same time combining them, and out will come some basic insight that is often more potent than one would expect!

    And that means something. It means not ordinary. Simply isn't going to happen, and that's precisely why I'll wait and enjoy the ride getting there. When we do get this next chip, it will exhibit some great characteristics and it will do some darn cool things, powerful things, and it will do them in ways that are as fun as they are potentially unconventional.

    I'll give that full props every single time. There is a whole world of conventional out there. I use it, you guys do too. So what? We know that's there and how it works, why it works. Chip will do something potent, fun, interesting, useful, and that's what I want to have happen. Simple as that.

    Here's an analogy: Ever consider art? I think most of us have. Music, painting, sculpting, whatever. Where there are limits, there is art. For the longest time, the more subtle details of art eluded me. When I began to understand people better, I also began to see minds as distinct things. Some are clean, tight, structured. Others a mess, or beautiful.

    Chip is an artist. This is what he does, and I'm glad he shares it with us.

    Everybody wants it done, and emotions and sentiments run high. Concerns mashed up with expectations, ideas, angst. All through that, Chip is being Chip, trying to make that great thing he sees possible to make.

    Maybe I'm just trying to say I don't see bad intent, or negligence, or anything like that in play. Sometimes it seems so, doesn't it?

    Just know that isn't the case. Everybody wants a good result. I think we are going to get one. That is all.
  • koehlerkoehler Posts: 598
    edited 2014-05-17 00:00
    +1 to everyone above, even potatohead for raw enthusiasm.

    It probably great when you're in a forum and people never stop telling you you're brilliant, you're the sharpest tool in the shed, can I have your baby?........
    Its probably also great when you work for a family business, have carte blanche, and can't be disciplined for not delivering too. Is that criticism or envy?

    Ken has a list of features that customers reportedly need/want to continue with the Prop.
    The current B/W re-architecture is not on that list, nor is there any evidence it will help the Prop in any way outside of some few forumista's personal wishes, which isn't many sales to justify all the changes inherent.

    Ken explicitly pulled Chip off the forums, so that all the work on the forest could be done without the distractions of the trees.
    Unfortunately, that had to be circumvented with someone's tree idea, and the result is that the bulwark of the Prop appears to be changed, for what appears to be a non-critical but theoretically impressive feature.

    I expect there will be more impacts/issues as we go down this path, and more h/w and snapcount-type s/w widgets are going to be needed, all to support this B/W-monster feature that now is the primary feature of the Prop.

    No worries though, customers are locked-in to the Prop and won't/can't go anywhere else no matter how cheap ARM gets, or how many Cores they add.
    Or at least that seems to be the thought process out in Sacramento.

    This is why artisitic types are artists, and not business owners, and many of the greatest died broke.
  • potatoheadpotatohead Posts: 10,253
    edited 2014-05-17 00:32
    Well, thanks I think.

    Hey, have you run a business? That's not a put down, just an honest question. Secondly, have you worked with niche ones?

    I have, and currently do. I often consult on product development, and some of the things I have been asked to sort out or maximize would likely surprise you.

    Raw enthusiasm indeed! I'm not taking that badly, nor should I.

    Know what being great at service, support and in business dealings overall is worth? Lots. Parallax has that. Always has, and people will often value doing that kind of business at a premium over a more core numbers game, and there are a considerable number of those people in the world too. Enough that I've seen it play out several times well. It's often enough to retire nicely on while feeding families who loved the work. Niches are interesting things, and the dynamics surrounding them aren't always by the book, MBA 101. They rarely are. Fact.

    Back when Chip was developing the P1, and Parallax was running well on Basic Stamps and education, there he was building something. He used to try to show it off. Many didn't understand, some laughed, others wondered what the point of it all was. He's told some of that story here, and I've had him tell it to me during one of our all night hack sessions we used to have at the expo. (we really need another one when this device gets done)

    When it was released, people started to give it a go. Lots of great things happened. And here we are. That thing people laughed at, which didn't add up on any of the MBA metrics as viable, worthy, etc... is now a nice business. Many of us, myself included, took one look at the block diagram and wondered what it was all about. I bought a Demo Board that day. Chip noticed people like me showing up and asked about it. Great dialog. Just know this is a labor of love and money. Not just one or the other. I could say much more, but it has to be over a beer someday Koehler. I like you and all that, despite us butting heads. Know that too. Maybe fly Heater, Ross, et al. in for a romp! (seriously, I would help pay! What a great talk that would be!)

    Please don't call people fools, or imply that. In this case, it is simply untrue. I know you really don't mean that.

    Here's the deal: We are going to get FPGA images and give this design a go. If it really sucks, it's gonna come out and it's gonna get fixed because Chip doesn't release Smile. Neither does Parallax.

    Until then, let the guy work. He's got an impressive history going back to the 8 bit era. My enthusasim is not unfounded.

    And you can call me spud, if that helps. :)
  • koehlerkoehler Posts: 598
    edited 2014-05-18 12:55
    potatohead wrote: »
    Well, thanks I think.

    Hey, have you run a business? That's not a put down, just an honest question. Secondly, have you worked with niche ones?

    I have, and currently do. I often consult on product development, and some of the things I have been asked to sort out or maximize would likely surprise you.

    Haven't run a business, just been responsible for multi-milion dollar capital infrastructure projects from presenting to the Board to budgeting to managing vendors to actual technical implementations. Basically, lots and lots of 'Big Picture' experience compared to 99% of my peers.

    Raw enthusiasm indeed! I'm not taking that badly, nor should I.

    Absolutely! I think you are reading far more into some posts than there is. I'm from NY, if wanted to insult, it would be very clear. ;)
    Know what being great at service, support and in business dealings overall is worth? Lots. Parallax has that. Always has, and people will often value doing that kind of business at a premium over a more core numbers game, and there are a considerable number of those people in the world too. Enough that I've seen it play out several times well. It's often enough to retire nicely on while feeding families who loved the work. Niches are interesting things, and the dynamics surrounding them aren't always by the book, MBA 101. They rarely are. Fact.

    Agreed. Ken/Chip had the opportunity to take a chance with the Prop, and its apparently worked out well enough to help support Parallax.
    Back when Chip was developing the P1, and Parallax was running well on Basic Stamps and education, there he was building something. He used to try to show it off. Many didn't understand, some laughed, others wondered what the point of it all was. He's told some of that story here, and I've had him tell it to me during one of our all night hack sessions we used to have at the expo. (we really need another one when this device gets done)

    When it was released, people started to give it a go. Lots of great things happened. And here we are. That thing people laughed at, which didn't add up on any of the MBA metrics as viable, worthy, etc... is now a nice business. Many of us, myself included, took one look at the block diagram and wondered what it was all about. I bought a Demo Board that day. Chip noticed people like me showing up and asked about it. Great dialog. Just know this is a labor of love and money. Not just one or the other. I could say much more, but it has to be over a beer someday Koehler. I like you and all that, despite us butting heads. Know that too. Maybe fly Heater, Ross, et al. in for a romp! (seriously, I would help pay! What a great talk that would be!)

    I agree, it was a niche they were able to expand and/or create on some level.
    Please don't call people fools, or imply that. In this case, it is simply untrue. I know you really don't mean that.

    Now you are sort of putting words into my mouth.

    I just reread my previous post, and I'm not calling anybody a fool, at all.
    My comment which mentioned you by name perhaps should have ended with a :) ?

    If there was an implication, it was simply that being sort of a family run business with 2 brother's in the lead is certainly odd, and seem to have benefits and drawbacks. Others have been a bit more blunt about it.

    Here's the deal: We are going to get FPGA images and give this design a go. If it really sucks, it's gonna come out and it's gonna get fixed because Chip doesn't release Smile. Neither does Parallax.

    See, here is where it gets 'interesting'.
    Parallax is unusual in the way it does business, its principal personalities, etc, etc.
    No one wants to be rude or offensive.
    However, because of that, there is a definite cult of personality thats been built up, such that criticism seem to be take by some as a direct insult to Chip, Ken, Parallax, etc.

    As I said in another thread, Chip and Ken seem more than able to handle that without taking personal offense.
    I see no need for anyone to protect them from criticism, they're adults.

    No company has everything go right, and neither has Parallax, of which the Javelin and Parallax-Semi pop into mind.
    Also, the P2hot pops into mind. IIRC, it was 5W just for the logic, and as several people suggested it would end up closer to 7-8 once memory, etc were added. Which IIRC again, was infact the case.
    Even allowing for lower speed, or average running without every flip-flop/ALU unit toggling, it still sounds like it would have been a 3000-4000ma uController.
    Chip is smart*, and he also let it get to that point, so here we are again with an even newer rearchitecting of the P2.

    *Again, this isn't to criticize Chip, as much as to point out that somewhere in Chip's process things got away from him, and this is arguably why its taking years to get an updated product out.

    If one wants to be honest with oneself, one has to call a spade a spade. These are process failures which have cost Parallax millions of dollars, time, market, etc. Parallax seems to have been able to absorb this without failing, so in the end it may not be as fatal as it would for another company.

    I've moderated some of my views, because after re-reading Ken's post of several weeks ago, it now seems clear that Ken is happy to remain focused on the education markets they have, and some of the more niche markets with entrepreneurs, artists, tinkerers, etc.


    Until then, let the guy work. He's got an impressive history going back to the 8 bit era. My enthusasim is not unfounded.

    And you can call me spud, if that helps. :)

    Again, not saying your enthusiasm is unfounded. You have every right to think as you wish.
    In my case and experience, one sets realistic goals to reach, and engineers to them.
    If things change or come up, you address them as necessary.

    However, as the old saying goes 'Burn me once, shame on you, burn me twice...." is apropos here.
    This applies to the current process being followed. Failing to address that will most likely see the same issues crop up.
    At some point, one can hope that the $$ will continue to roll in long enough for a final product to be delivered, despite the process issues.
  • potatoheadpotatohead Posts: 10,253
    edited 2014-05-18 13:34
    Basically, lots and lots of 'Big Picture' experience compared to 99% of my peers.

    This explains a lot. I'll understand you much better now. Thanks. Not a bad thing, but a difference in perspective I didn't pick up on earlier.
    Now you are sort of putting words into my mouth.

    Yeah. My bad. Again, I understand a few things better now that I know where you are coming from.
    However, because of that, there is a definite cult of personality thats been built up, such that criticism seem to be take by some as a direct insult to Chip, Ken, Parallax, etc.

    Yes. This is how niches work. Apple computer is a shining example of how that can all work. They are big, but they are still playing with some similar rules, less so now that Jobs is gone. The two single most important things to realize is that it's not about volume, but margin. The second one is sales, service, support being excellent, as much value added as possible. Obviously, asking for that value is important too. (Apple is particularly adept at managing the shift from high margin lower volumes to higher volumes, while keeping impressive margins anyway. Few do this well. None that I've seen do it as well.)

    The dynamic is this: Exemplary sales, service, support and value added creates the framework for "the cult" that you speak of. And it is a little cult like, and that's not a bad thing! Having set that up, people recognize value differently, which creates the margins needed to be sustainable and take risks at a smaller scale that would normally require outside investment and or credit.

    Most of this effort has been directly funded. It may take longer, cost more, whatever, but it can also be done in a sustainable way. I'm not excusing the amount of time in this case. Things need to get done. However, I'll cite a manufacturing conglomerate I worked with for years. Same basic approach. All cash business, no credit anywhere, private. Setting up new processes could be done in a month. This guy would take a year, or sometimes just buy a company with the process already running. Cash of course. It took me quite some time to understand the difference and how opportunity costs are valued and they don't always mesh the same way you are used to. (and me, as I have to do it too sometimes) When I worked for them 20 years ago, I was wanting a laser to improve on some short run metrics. Was frustrated that a standard ROI didn't work. The guy sat me down, opened the books and showed me how he does things. Let's just say that was one of the most educational 2 hours I ever spent with anyone.

    Today, that company can make nearly anything, and it operates debt free too. Notable. I'll bet it's very well capitalized and his sons will have an impressive legacy to work with.

    Right now, I'm working with another little company that is investing in some pretty niche and advanced manufacturing processes and doing some basic materials engineering while they are at it. It's gonna take them a few years, and they've got some in-house, proven business that funds it. Everybody takes their living expenses out, the rest gets plowed right back into the enterprise, and the owner will also bank a sliver of that, just in case. Being capitalized well enough to weather some variances is important.

    The R&D could be done quicker, a bit more formal, but it would be shared, and or done on credit, etc... The difference is owning a process and having it in house, then able to compete on that, or sell it in various ways with nearly all the profit or payoff being made available instead of being diluted to pay down banks and investors. Interestingly, that it's done in house, niche style, adds value! There is a perception about it that people will pay for. Amazing to me, but 100 percent real.

    Now I'm there to formalize things a little, and put those goals out there in a more structured way, as you say. But there are limits and a core philosophy that will still differ from the "standard" MBA type approach. BTW, I'll be replacing one of those MBA people. No joke.

    Some of those dynamics are in play here. Not all. But maybe that helps explain how you and I clash! :) And I like you Koehler.

    I've not taken anything as an insult, nor have they I am sure.

    What you are seeing is significant differences in how things are valued and perspective.
    These are process failures which have cost Parallax millions of dollars, time, market, etc. Parallax seems to have been able to absorb this without failing, so in the end it may not be as fatal as it would for another company.

    "failures" in the context of the above is a highly relative and arbitrary thing.

    One could say Chip has bootstrapped himself onto chip design. Secondly, we've all bootstrapped onto the FPGA, and learned something about how to work together for common interests, the idea being we can exploit the product together. (which is why things need to get done)

    If you look at availability of funds, whether or not one seeks capital, loans, investment, there is the reality of there being money to spend on these kinds of things, but not necessarily all at one time, or quickly. Something to think about isn't it?

    In any case, "process" needs to be put into context.
  • Heater.Heater. Posts: 21,230
    edited 2014-05-18 14:21
    koehler,
    [...cult of personality...]
    If you want to put it like that. Which I don't think you should as it trivializes something deep.

    I might say we have a culture of admiring engineering talent and tenacity. Read this to see what I mean.

    "Why the Propeller Works": http://forums.parallax.com/attachment.php?attachmentid=45375&d=1170990878
    That story reads like an epic struggle for the little guy to realize his dream and do the impossible at all odds.

    Notice how that P1 chip was designed "polygon by polygon", prior to that the logic was all designed the old fashioned schematic diagram way. The testing was done by hand with a cast off ion beam machine on the cheap.

    For the P2 Chip has updated his skills to designing with Verilog and had to learn about all the new process technologies and design tools. And so on. All without the backing of being part of some huge semiconductor design corporation.
  • potatoheadpotatohead Posts: 10,253
    edited 2014-05-18 14:23
    Precisely and succinctly said Heater.
  • Personally, I have always imagined the Prop2 being almost exactly like a Prop1. Of course, more cogs and faster clock, and also more hardware abstraction (a.k.a UART, ADC). It just seems ridiculous to make this too much of a workhorse and not enough of a useful chip.
  • The current design spec is much different from what appears on this old thread.

  • Whew!!! I saw this thread at the top of the list and panicked! False alarm...

    @moderators : could this thread get locked or something?
  • Heater.Heater. Posts: 21,230
    jacksons,

    A lot of us agree.

    It has been suggested many times over the years that Parallax should put out a 64 pin Prop, with the speed gains and larger memory made possible by switching to an HDL design and smaller semiconductor process. Many would have been very happy with that, even with only 8 COGs.

    The idea being that such a Propeller upgrade could have been out years ago. Without the endless diversions we have had getting us to the current design. Which is still at least year away from being on my bench.






  • Yes, you guys are right. Closing this thread as P2 has moved on somewhat.

    Anyone with new questions or ideas, please do not hesitate to start a new thread. Always great to read what the many minds are thinking!!!
This discussion has been closed.