The Second Batch Info Thread

[Monk]

if 10,000 is the minimum then perhaps they will just order 12,000 to make it all nice and neat.

Maybe 10,000 is the minimum unit of order, so the minimum increase on 10,000 would take the number up to 20,000?

:shrug

"Craig Rothwell ... is the one guy who took most of the risk of the project onto his shoulders. He did most of the organization with all the companies, the money, etc."

1) what other major risks could there be besides finances?

Dunno. Maybe the risk that someone from the forums would be irked enough by his informative posting style to kneecap him? ED doesn't take so many risks there, his posting style is different

But anyway he said he took MOST of it. That doesn't meant that ED adn MWeston and Faith and whoever else haven't got second or even third mortgages to finance this project - just that Craig has somehow taken MORE risk than them. Possibly visited more bankers and loan sharks than they did (whioh can be risky too).

2) and why would all the developers let him move the money around if most of it wasn't his?

Faith. If you believe that someone can make you a lot of money by managing a little money from you, it can be very tempting to give them AT LEAST a little money. There's a whole bunch of industries out there that depend on people believing that someone else can make them more dough from their money than the owner could by themselves. Investment funds, well actually now I think of it almost every (if not every) financial institute going has this principle as their foundation.

3) it appears that he is the only one with a store (at least what it says on the about us page), so would it be too presumptuous to say that he has the most money to invest, since he will be selling them on his store?

Doesn't ED have a store? :unsure:

 

[mali]

Doesn't ED have a store? :unsure:

Yes, he does and Fatih, too. Aesir knows this, or should know it, as it has been explained a few days ago, or he forgot it, or he ignores it, dunno.

 

[aesir911]

Doesn't ED have a store? :unsure:

Yes, he does and Fatih, too. Aesir knows this, or should know it, as it has been explained a few days ago, or he forgot it, or he ignores it, dunno.

yeah... minor lapse of... knowledge. (http://www.gp2x.de/shop/index.php)

 

[darkborn]

At this rate the second batch will ship before the first.

Nasty one!

 

[michaeljustman]

BTW, aren't the 720 MHz CPUs the same as the 600 MHz, only tested and guaranteed to run at a higher speed?

Correct. They're the same hardware, just "binned" according to test results.

Which means the second batchers are pretty much guaranteed a lower speed chip as any that were capable of 720 mHz would have been sorted out and sold at double the price, whereas the first batch all has 600 mHz rated chips that wouldn't have been sorted.

 

[Aninhumer]

Which means the second batchers are pretty much guaranteed a lower speed chip as any that were capable of 720 mHz would have been sorted out and sold at double the price, whereas the first batch all has 600 mHz rated chips that wouldn't have been sorted.

The discussion has been had many times before.
Basically there are a lot of possibilities about what the release of a new part means for the old one, but no one except TI can do anything better than speculate.

I think we should avoid making sweeping assertions about something we don't really have any information about.

 

[michaeljustman]

Which means the second batchers are pretty much guaranteed a lower speed chip as any that were capable of 720 mHz would have been sorted out and sold at double the price, whereas the first batch all has 600 mHz rated chips that wouldn't have been sorted.

The discussion has been had many times before.
Basically there are a lot of possibilities about what the release of a new part means for the old one, but no one except TI can do anything better than speculate.

I think we should avoid making sweeping assertions about something we don't really have any information about.

I didn't mean to state it so emphatically, but it is a reasonable assumption, but you're right, for all I know the 600 mHz chips may just come off the line and ship whereas other chips off the same line are sent for more rigorous testing to be sold as 720 mHz chips.

 

[Monk]

BTW, aren't the 720 MHz CPUs the same as the 600 MHz, only tested and guaranteed to run at a higher speed?

Correct. They're the same hardware, just "binned" according to test results.

Which means the second batchers are pretty much guaranteed a lower speed chip as any that were capable of 720 mHz would have been sorted out and sold at double the price, whereas the first batch all has 600 mHz rated chips that wouldn't have been sorted.

Simplistically that's what it LOOKS like. That's also what TI relies on it looking like, as they want to sell higher cost 720MHz packages. But it's still guesswork (educated guesswork, but guesswork). There are other interpretations.

For example, they may bin some to 720MHz - but what if the results of the tests are that none fail? If all the SOCs tested pass at the 720MHz rating, then they likely all would, in which case the SOCs sold as 600Mhz would be exactly as fast as the ones GUARANTEED to run that quickly - they just aren't guaranteed to do so. That would mean second batch "600MHz" SOCS are now slower than first batch "600MHz" ones.

For another tweak, how about the poossibility that yeilds have improved, so that newly-produced chips are more reliable/stable/run faster/do your washing up for you, which is why TI have felt confident enough to laebl any of them as 720MHz. If the product coming off the line is higher quality in ANY way for ANY reason that has lead to TI having the confidence to mark some as faster, then in theory second batch "600MHx" SOCs could, on average, run faster (with stability) than first batch "600MHz" ones.

MWeston was, IIRC, going to ask "our man in TI" what the actual scoop was, but AFAIK never managed to get a concrete answer (I certainly haven't read a definitive "what's what" about the reasoning behind the binning). The most obvious, reasonable, logical conclusion is that 720MHz SOCs are, or should be, faster than 600MHz SOCs... but companies are rarely logical, and marketing is never, EVER reasonable... and marketing is what the extra product line of 720MHz SOCs is about.

My personal guess is that the first batch SOCs will turn out to be more overclockable than second batch ones. If I believed otherwise, after all the dicking about OpenPandora have done, I'd have cancelled my order months ago and waited to see if they survive long anough to produce a second batch, then buy from that or batch 3, if not Pandora 2. A Pandora-like product based on GHz processors, possibly multi-core, is increasingly of more interest to me.

 

[CCmachined]

Craig probably needs to adjust for years of inflation in there from the initial price...

Okay, so we're looking at about £250~260 for a less overclockable device?

Bah, I supported the project for 15 months. Oh well, as is life. Not really anyone's fault.

 

[mali]

CC_machine, what you read is speculation and nobody knows how overclockable the 2nd batch Pandoras are. I'll let you know when I get mine

 

[peca]

I didn't mean to state it so emphatically, but it is a reasonable assumption, but you're right, for all I know the 600 mHz chips may just come off the line and ship whereas other chips off the same line are sent for more rigorous testing to be sold as 720 mHz chips.

Carefully, size does matter! It is not mili Hz but mega Hz, so not mHz but MHz.

I curious why some people doesn't care about sentences of multipliers and units. The same stand for bits and bytes (b vs B) . Then they feels confused when reads that Pandora has 2 Gb of RAM.

 

[limetang]

Craig probably needs to adjust for years of inflation in there from the initial price...

Okay, so we're looking at about £250~260 for a less overclockable device?

Bah, I supported the project for 15 months. Oh well, as is life. Not really anyone's fault.

No, you're probably looking at £275+ for a device with currently unknown overclockability.

 

[Usoppe]

Its possible, too, that the first batch IS AS overclockable as the second batch. My theory:

TI just needed to collect some 720 MHz capable OMAP3s of the production before they introduce the 720 MHz OMAP3s.
And at a certain amount of 720 MHz OMAP3s they introduced the product. So it could be possible, too, that they have classified the OMAP3s from its first production on.

Its more like if you have an optimistic or pessimistic/realistic view on things.

But I wouldnt put my hand in fire and say "first production OMAP3s are more overclockable than the ones in the second batch of Pandoras".

We have no clue about that, no comment, nothing. Thats fact! So we will see in near future how good or bad each Pandora is for overclocking

 

[Dunny]

I'm fairly certain that a long while ago, Craig (or OP member) had stated that all of the SOCs we needed for the first batch had already been bought and paid for. This was quite a while before the issue of binning reared its ugly head, so surely - given that they already have the SOCs:

1. The overclockability of the first batch is uncertain, and an unknown number will reach speeds as high as (or higher than) 720MHz
2. The second batch will be only rated at 600MHz and overclocking is unlikely.

D.

 

[Monk]

I'm fairly certain that a long while ago, Craig (or OP member) had stated that all of the SOCs we needed for the first batch had already been bought and paid for. This was quite a while before the issue of binning reared its ugly head, so surely - given that they already have the SOCs:

1. The overclockability of the first batch is uncertain, and an unknown number will reach speeds as high as (or higher than) 720MHz
2. The second batch will be only rated at 600MHz and overclocking is unlikely.

D.

Yes to #1, but not to #2.

The original SOCs that OpenPandora already bought were certified to IIRC 600MHz. They have them all for batch 1. They have none for batch 2.

They have said they are going to buy 600MHz certified SOCs for batch 2. They will actually be the same part number, the same spec, the same everything.

What we know about 720MHz binned/certified untis:- They are certified to reach at least 720MHZ

What we don't know about the binning:
1. We don't know if any units fail the 720MHz test.
2. We don't know that new "600MHz" SOCs are worse at overclocking than batch #1 ones
3. We don't know that new "600MHz" SOCs are better at overclocking than batch #1 ones

Really. Seriously. There are reasons/logical arguments that can go etierh way. My personal guess is that new 600MHz SOCs are - or at leas tMIGHT - be less good at overclocking than batch #1 ones - but it is a guess. We have NO data that confirms that, nothing but guesswork and interpretation based on what we ASSUME TI bodies have as their reasons for creating the new product marking and pricing.

We have no evidence, actually, that the 720MHz SOCs will be any different to the first batch 600MHz SOCs, but they have a PAPERWORK difference - they are GUARANTEED to reach 720MHz, which ours are not. That's the only difference we have firm documentation on, and while we can infer more - it's inferance only at this stage.

 

[Dunny]

2. We don't know that new "600MHz" SOCs are worse at overclocking than batch #1 ones
3. We don't know that new "600MHz" SOCs are better at overclocking than batch #1 ones

Really. Seriously. There are reasons/logical arguments that can go etierh way. My personal guess is that new 600MHz SOCs are - or at leas tMIGHT - be less good at overclocking than batch #1 ones - but it is a guess. We have NO data that confirms that, nothing but guesswork and interpretation based on what we ASSUME TI bodies have as their reasons for creating the new product marking and pricing.

Myself, and of course, since we're discussing something for which we have absolutely no evidence for then logically I'd say that the 2nd batch will likely be unstable at 720MHz. Why? Because if TI knows they can charge more for something that will perform better with no extra work other than testing, then all the 600MHz chips will be tested and binned. Anything that remains stable at 720 will fetch a higher price, so it's worth the effort of testing them all.

So I would stand by my original assessment that the first batch will be of uncertain overclock ability, and the 2nd batch will be known-unstable at speeds > 600MHz.

Edit:

Actually, we don't know how TI are doing the testing. If they test at 720 and bin them in "can clock, can't clock" then we have guaranteed 720s and "can go up to somewhere between 600 and 720". So in that case, hypothetically of course, the 2nd batch should be overclockable by an unknown amount which varies from unit to unit, up to 720MHz. The 1st Batch will be able to overclock, and again this will vary greatly from unit to unit, anywhere from 600MHz upwards to an unknown ceiling.

We have no evidence, actually, that the 720MHz SOCs will be any different to the first batch 600MHz SOCs, but they have a PAPERWORK difference - they are GUARANTEED to reach 720MHz, which ours are not. That's the only difference we have firm documentation on, and while we can infer more - it's inferance only at this stage.

Absolutely. But isn't this fun?

D.

 

[craigix]

It's worth knowing that we are yet to come across a chip which can't reach 720mhz.

 

[Vorporeal]

It's always possible that they know that most chips can reach 720MHz (and only a few fail at those speeds), so they bin chips when an order is placed for the 720MHz chips. They might not do pre-binning if the failure rate at 720MHz is very low.

 

[hitbyambulance]

in summary: if anyone planning to order from the second batch is really really worried about this, they can just place an order with the first batch.

 

[WizardStan]

Myself, and of course, since we're discussing something for which we have absolutely no evidence for then logically I'd say that the 2nd batch will likely be unstable at 720MHz. Why? Because if TI knows they can charge more for something that will perform better with no extra work other than testing, then all the 600MHz chips will be tested and binned. Anything that remains stable at 720 will fetch a higher price, so it's worth the effort of testing them all.

On the other hand, they may find they have a surplus of 720 chips and have to sell them rated at 600 because there's still a market for them, a la Intel.