Pi vs BeagleBone-Black

Stuart Richards wrote:

 

mynameisJim wrote:

 

Stuart Richards wrote:

 

My guess is that there are significantly less BCM2835 SoC based non-Raspberry Pi devices in existence than there are Raspberry Pis.  I think there is a strong possibility Broadcom had a significant number of BCM2835 SoCs that they couldn't sell prior to the Pi.

Well I would say see my previous post to this exact statement you made just a couple of posts ago, but it would seem the police state keeps censoring them so here it is again:  That is not the way chip manufacturing works.  You don't build a huge excess, especially if you're continuing with the idea that the chip was a failure.  You build to meet the order and while I keep saying you might end up with just a small buffer to fill capacity, the reality of the situation is that, more often than not, you force the individual to purchase in quantities equal to a production batch quantity.

 

My statements are not identical.  To clarify, when I referred to devices in my first post about this I meant products developed, not units sold.  I was referring to units sold in my subsequent post.  As far as I am aware the Roku 2 line of products is the only one prior to the Raspberry Pi that use the BCM2835 SoC.  I'm sure Broadcom didn't develop the BCM2835 exclusively for Roku and therefore must've been disappointed by the total number of design wins it achieved.

I'm not sure if I'm understanding the difference you're going for (sorry).  To clarify, in both cases aren't you suggesting that there were a "significant number of BCM2836 SoCs that they couldn't sell"? I.e. they had a stockpile of them somewhere?  If the second post is more "Oh darn, we didn't sell as much as we had hoped when we designed this, but at least there isn't a warehouse full of them."  Then until we were getting some numbers on the Roku I would have said I agree that that's possibility, though I don't think it would have affected the Pi.

I mean first off it suggests a stronger connection between the foundation and broadcom than evidence supports.  I get that it's fun to concoct theories and possibilities by sniping comments out of context and ignoring common business practice for this field, but's all about occam's razor.  Either the foundation has been lying about everything for over 2 years, broadcom made a dummy charity and forced one of its employees to get a second mortgage on his home just to sweeten the deception, and 6 people went along with this deception despite the fact that 4 or 5  of them don't have any affiliation with broadcom at all (I'm not sure about Pete Lomas).  OR they're inexperienced and are learning in the school of hard knocks what it takes to get the job done resulting in sometimes conflicting statements.

Also, assuming the less likely deception angle is in fact true, why use a chip that (in the scenario you presented) is failing for an additional 10K in sales?  No one really thought the pi would be as hot an item as its been.  Towards the end before the launch it looks like people were starting to think more than 10k would be sold, but over a million?  From a manufacture standpoint, it doesn't stand to reason to try to boost sales on a failing chip when you don't have an excess of stock just lying around, which a supplier such as Farnell might (If they were allowed to puchase the chip, lol), but a manufacture would not.

Also, I had said what follows previously, but that post mysteriously vanished, ho hum.

I like the idea of the Begeal bone black.  I'm concerned that its primary OS is an embedded OS as opposed to a full fledged OS.  I know that it has a full OS option, but it's not receiving the love attention the embedded one is.  I'm also interested in the on-chip ethernet it comes with and how it ultimately communicates with the CPU.  Like the Pi, one thing that concerns me is the spotty documentation.  I've been able to find things like this https://github.com/beagleboard/am335x_pru_package through google, but it doesn't seem to be giving the full story and I could not find it via the main site.  Finally, I don't like that to get some sort of video option that I have to purchase an addition "cape".  Horses for courses as they say, it definitely fits the bill as a contender for the Model A as it is geared more towards the hobbyist who need more CPU power than the Pi can provide.

Stuart Richards wrote:

 

mynameisJim wrote:

 

Stuart Richards wrote:

 

My guess is that there are significantly less BCM2835 SoC based non-Raspberry Pi devices in existence than there are Raspberry Pis.  I think there is a strong possibility Broadcom had a significant number of BCM2835 SoCs that they couldn't sell prior to the Pi.

Well I would say see my previous post to this exact statement you made just a couple of posts ago, but it would seem the police state keeps censoring them so here it is again:  That is not the way chip manufacturing works.  You don't build a huge excess, especially if you're continuing with the idea that the chip was a failure.  You build to meet the order and while I keep saying you might end up with just a small buffer to fill capacity, the reality of the situation is that, more often than not, you force the individual to purchase in quantities equal to a production batch quantity.

 

My statements are not identical.  To clarify, when I referred to devices in my first post about this I meant products developed, not units sold.  I was referring to units sold in my subsequent post.  As far as I am aware the Roku 2 line of products is the only one prior to the Raspberry Pi that use the BCM2835 SoC.  I'm sure Broadcom didn't develop the BCM2835 exclusively for Roku and therefore must've been disappointed by the total number of design wins it achieved.

I'm not sure if I'm understanding the difference you're going for (sorry).  To clarify, in both cases aren't you suggesting that there were a "significant number of BCM2836 SoCs that they couldn't sell"? I.e. they had a stockpile of them somewhere?  If the second post is more "Oh darn, we didn't sell as much as we had hoped when we designed this, but at least there isn't a warehouse full of them."  Then until we were getting some numbers on the Roku I would have said I agree that that's possibility, though I don't think it would have affected the Pi.

I mean first off it suggests a stronger connection between the foundation and broadcom than evidence supports.  I get that it's fun to concoct theories and possibilities by sniping comments out of context and ignoring common business practice for this field, but's all about occam's razor.  Either the foundation has been lying about everything for over 2 years, broadcom made a dummy charity and forced one of its employees to get a second mortgage on his home just to sweeten the deception, and 6 people went along with this deception despite the fact that 4 or 5  of them don't have any affiliation with broadcom at all (I'm not sure about Pete Lomas).  OR they're inexperienced and are learning in the school of hard knocks what it takes to get the job done resulting in sometimes conflicting statements.

Also, assuming the less likely deception angle is in fact true, why use a chip that (in the scenario you presented) is failing for an additional 10K in sales?  No one really thought the pi would be as hot an item as its been.  Towards the end before the launch it looks like people were starting to think more than 10k would be sold, but over a million?  From a manufacture standpoint, it doesn't stand to reason to try to boost sales on a failing chip when you don't have an excess of stock just lying around, which a supplier such as Farnell might (If they were allowed to puchase the chip, lol), but a manufacture would not.

Also, I had said what follows previously, but that post mysteriously vanished, ho hum.

I like the idea of the Begeal bone black.  I'm concerned that its primary OS is an embedded OS as opposed to a full fledged OS.  I know that it has a full OS option, but it's not receiving the love attention the embedded one is.  I'm also interested in the on-chip ethernet it comes with and how it ultimately communicates with the CPU.  Like the Pi, one thing that concerns me is the spotty documentation.  I've been able to find things like this https://github.com/beagleboard/am335x_pru_package through google, but it doesn't seem to be giving the full story and I could not find it via the main site.  Finally, I don't like that to get some sort of video option that I have to purchase an addition "cape".  Horses for courses as they say, it definitely fits the bill as a contender for the Model A as it is geared more towards the hobbyist who need more CPU power than the Pi can provide.