Change In Mhz Ratings..

[Exophase]

I think if any significant number of the chips operate at 720MHz over the full temperature range, which is actually rather extreme, then probably most of them will operate at 720MHz at room temperature and lower. If you really want your Pandora to overclock maybe you should play it in a freezer :B

 

[Tor]

TI probably just started [1] speed-sorting these chips. Which they've done before. Nokia got speed-sorted (and more expensive) OMAP2 SoCs for their N800/N810 tablets so that they were guaranteed to run at 400MHz.

What it could mean for us is that now that they sort the things there may be a lower rate of OC-able chips in the standard 600MHz set.

[1] Guesswork by me.

 

[MonkeyChops]

about checking every board for oc ability,
are they (op) really testing every board anyways?

I thought they were doing extensive tests on the first 100 and as long as the failure rate was low they were just going to spot check the mp boards.

to be honest, I am really looking forward to the oc ability of the pandora. When I was first interested in the pandora and i saw the 600 mhz, I was like, meh. Then I read that they were being overclocked to 900 and I started to get excited, then I pre-ordered.

600 is more than enough for most of what the pandora will do, I'm sure, but I'll be pushing my pandora to the limits. I'm very interested in the possblity of dc, ds, and psp not to mention dosbox. All of which would benefit greatly from higher clock speeds.

Its defiantly not a deal breaker for me but I will certainly be disappointed if I can't overclock. I'm hoping to get close to the 900 mark with both units I bought.

 

[openthebox]

I always viewed the OCability as a bonus feature, so I'm not too hung up on it.

With that said, how does one (especially a noob like me) go about testing the OCability?
Is the test as simple as:
1) If it did not shut down the unit immediately.
2) If the running apps does not behave eratically.
3) If the unit does not lock up.
4) If no burning smell come out of the unit.

 

[Exophase]

You'd probably want to run a test that stresses every level of the system for several hours. Something that streams data from RAM to NEON (while L2 cache is on) and does extensive operations on it could be effective. Probably also have DSP testing, but separate.. especially since right now nothing is even using it. And the two clocks are async anyway.

 

[sold]

I was totally looking forward to OC'ing my Pandora. It was a "bonus" feature that was being talked about like it was definite up until now. I don't expect the guys to take a loss, but I do think the team should at least look at the cost of the soc now compared to when they bought the first lot, and if the price is the same/less, moving to the speed tested boards couldn't hurt. This is a product for enthusiasts, so the ability to overclock would be a major selling point because some of us want the extra juice we can squeeze out of our Pandoras.

By my understanding the speed tested chips are stable at 720mhz, but the 720mhz guaranteed parts are being sold at a slight increase in price (keeping in mind they had gotten cheaper over the last year) and the ones that don't run at 720 are being sold cheaper.

If the manufacturing process is the same, then this is a complete load of bullshit. The product that TI are selling is the same, if they wanna bump up their specs for the sake of marketing so their soc appears to be more "powerful" so they can remain competitive, that should not be at any cost to their customers or the consumer.

They are choosing to push their product a little harder and are guaranteeing a slightly higher speed, but it is sounding more and more like they haven't changed a thing. This "grading" thing stinks. We are not buying a freaking steak here. This isn't something that is grown naturally where differences in quality are to be expected due to variables (environment/genetics/etc.) where it is justifiable that product can be graded and sold accordingly. This is a high-tech precision part manufactured in an environment that is designed to eliminate variables which create undesired effects and TI should offer better quality control than what it sounds like they are offering currently imho.

AND they should be giving the OP team the speed sorted boards for a fair price out of principle.

Nothing against the OP team or the project. Still loving your work. What TI are trying to do here sounds a little shifty to me, and before anyone replies "but all hardware manufacturers have done something like this in the past" or something to that effect, that still doesn't make it right.

If I've got something wrong here, please feel free to clarify Sorry about the long post

 

[Vorporeal]

TI only (originally) guaranteed the chips for 600MHz. They produced the chips in such a way that they'd all reach 600MHz stably at standard voltages. When you're producing something that small though, even precision, computer-guided robots don't do things exactly the same way every time. Bits of dust or other particulate matter can interfere (even in a very controlled, "sterile" atmosphere). Things aren't totally perfect. Because of this, some chips can reach 900MHz stable, others can't even hit 720MHz stable. TI is just making another SKU by which you can buy chips knowing that they'll hit at least 720MHz.

Unrelated to sold's post:

Thinking about the whole issue, I noticed some strange stuff. Why have all the boards tested by the devs reached high speeds? (Most if not all have reached 720MHz+, right?) If they purchased a totally random assortment of chips, you'd think that some of them wouldn't have passed 720MHz if TI is actually making two separate lines, one that is 600-720MHz and another that is 720MHz+. You could call it luck, but it might not be. If they really (by luck) didn't get any under-720MHz ones, it would seem like most of the chips surpass 720MHz, and there wouldn't be that many thrown into the 600MHz bin. I wonder if TI is just making a higher price-point where you're GUARANTEED that you'll get 720MHz, but the cheaper chips aren't cheaper because they're known not to hit 720MHz, they might just not have been tested and therefore aren't guaranteed to get that high. It's unlikely, but definitely possible.

 

[Exophase]

But TI only just started binning recently, so it's unsurprising that prior chips would have good overclocking..

 

[Vorporeal]

Yes, but you'd think there'd still be variability with the old chips. They went from all of the chips in one pile to good chips in one, bad in the other. You'd expect the dev team to have some of the bad chips, considering that the existing dev boards had chips that were taken from the one big pile.

 

[Monk]

I'm concerned that this will mean that the 600MHz ones are going to be the bin losers that won't even clock at 720MHz at all, putting us in a worse situation than where we started.

I hope I'm misunderstanding you- but are you saying: we can't overclock the Pandora to over 600MHz?

As has always been the case (if not always obviously) any given Pandora may - or may not - be overclockable. Regardless of which batch it comes from.

They were delivered ages ago, and they (so far) seem quite happy to clock to >720mhz. As said though, it is still nice to optimise code.

The fact that it's not been possible to guarantee overclocking, or how high ALL Pandoras will overclock, is why some of the... sceptics among us have tended to call for "Without overlocking" porting and videos of ports from Quakes to DOSBoxes. Because 600MHz is all that's guaranteed, for a Pandora program 600MHz should be the maximum clock rate you NEED to run it (albiet with some stuff turned down, or off).

IMHO.

Ah, so it was some kind of semi-bluff to make it look good. Not saying it might only be overclockable by .001%.
That's business I guess and I should have looked it up myself to know about it. I don't think it's very honest though

edit: I also found this interview with Craig Rothwell: http://www.pocketgam...news.asp?c=8667 and I quote:
It is, without doubt, the most powerful portable games system ever and at the time of writing the most powerful mini computer, too. It has openGL 2.0 3D hardware, a 600-900Mhz CPU, media coprocessors, an 800x480 LCD and 128MB of RAM. The killer app with the hardware is that it is using an ARM Cortex A8 CPU, meaning astonishing battery life – over ten hours – which blows away similar mini PCs based on the old x86 chipsets. You will be able to play games like Quake III with ease and at the cutting edge even Doom III (if the engine goes open source soon, of course).

Wasn't there that university session advertised with a 1GHz clock rate as well? But this is Craig... something which should be in his sig line. Mandatory.

I am...dissapointed that we won't be having the official clock rate of the Pandora raised to 720MHz. That would have been nice. But look at it this way - it probably would have caused a delay of a month or two...

 

[Monk]

Not at all. That's why I said to see if it's a good deal, first. If, by the time the second batch rolls around, the 720s are the same cost as the 600s are now (or possibly even cheaper) then you lose nothing in that batches upgrade. There's even the possibility they may scrap the 600s entirely, discovering that almost all the chips coming off the line are capable of 720 mhz, with the few that don't not reaching high enough quantities to sell in bulk.

If they are separating the parts, then there will always be a price difference.

This seems a little off to me - are you saying that, no matter how cheap the 720 ones ar ein the future, you will always use the cheapest possible product?

Yes, but you'd think there'd still be variability with the old chips. They went from all of the chips in one pile to good chips in one, bad in the other. You'd expect the dev team to have some of the bad chips, considering that the existing dev boards had chips that were taken from the one big pile.

AFAIK It is not certain that there ARE any bad chips. Order some that are certified for 720 and TI will test them to 720 for you, for the extra cost. All others get sold as 600's... so they might actually have a 0% failure rate at 720. We may never know.

 

[sold]

^ That is exactly what us pissing me off.
It is basically quality control that they are charging for.
Fuck that. They should be testing ALL the chips to verify that they work properly.

I like the Pandora. I like it so much that I am hanging in there without the money I spent OR the device, and I am strapped for cash (yeah, I know I'm not the only one), and now there seems to be more delays and more crappy news. For me, 600mhz isn't gonna cut it for dosbox, which makes me a little upset. And getting upset makes me less patient. And the prospect of getting a Pandora that doesn't overclock makes me less inclined to resist the urge to be impatient.

Also, isn't it true that only 105 boards have been manufactured so far? If so we could be looking at the bargain basement chips in the first batch.

If the tested chips are the same price as the 600mhz ones were 12 months ago, I don't understand why the OP team wanna risk upsetting their customers by making this one big expensive lucky dip that could effect some peoples end user experience with the hardware.

EDIT: Not meaning to sound negative or pushy. Just getting a little discouraged, unfortunately.

 

[Exophase]

I'm pretty sure they already bought the chips, guys. Ones that weren't binned.

It's really petty to complain about the guaranteed clock speed being too low when it has the clock speed as advertised when you purchased it. If 600MHz isn't enough for DOSBox then maybe you should shift your complaints to ARM DOSBox being too damn slow rather than the CPU being too slow, because it has a hell of a lot of room for improvement.

 

[sold]

True. But overclocking was a quick fix that would've kept me happy at least
Also, overclocking was spoken of as a given up until this most recent development.
I'm not gonna change my mind about purchasing the Pandora because of this, it just made me mad that TI can start separating their chips at this stage, and I thought it would cause more problems for the OP devs. Which in turn means more issues for the community in the future. Such developments can be nothing but detrimental.

Also, if the chips have already been payed for but haven't yet been received, there is no way of knowing what they are gonna get now.

Does anyone know what is actually going on? If you don't mind me asking...

 

[Squidge]

Even if the Pandora had come out 6 months ago, there would have still been no guarantee that any particular Pandora could be overclocked. The only guaranteed speed has ever been 600Mhz+ (ie, 600Mhz with the possibility of overclocking to a faster speed).

If 600Mhz isn't enough for you, then the Pandora isn't suitable for you and you shoudn't have pre-ordered.

The only reason to purchase the more expensive guaranteed 720Mhz chips is if you want to put that on your marketing literature. Since Pandora has only ever been described as 600Mhz, there's no need to do that.

 

[Monk]

Also, if the chips have already been payed for but haven't yet been received, there is no way of knowing what they are gonna get now.

Does anyone know what is actually going on? If you don't mind me asking...

This to me is the crux of the reasoned/reasonable concerns. We cannot - CAN NOT - demand more than 600MHz from our Pandoras because - hyperbole aside (mentioning no Craigs

) the TI recommended 600MHz is all that's ever been GUARANTEED by OP in their official literature. Yes, I know that many times people have been blasted on the forum for suggesting that you cannot guarantee all Pandoras will be overclocked, but that wasn't (IIRC) by OpenPandora themselves. OP sat back and let others make claims on their behalf without correcting them, for the main part, so at worst it's an error of obmission rather than of an active willful act.

The upshot of that is, I believe, that OpenPandora have in effect "plausible deniability" with regard to people having the impression that you can rely on greater speeds than 600MHz. They can problably even point to one or two posts where a team member specifically states that high clock rates are not guaranteed. It's a bit of a "buyer beware" problem with the forum - we tend to be a bit enthusiastic about the Pandora, and listen gleefully when someone says "Oh, Doom III ought to be easily within range of what this machine can do" and pay far less attention when they moderate their view to "Doom III with everything turned down to minimum is probably going to really push the boundaries on this machine".

That aside, the chips have been bought from what we're told. Ordered, paid for, bought - but WERE they delivered? There is wiggle room, perhaps, to wonder if TI "delivered" to themselves, and kept the Pandora SOCs in a warehouse on OpenPandora's behalf... OR that they decided to keep moving stock and would provide OpenPandora's SOCs at manufacture time, right off their own production line. If the latter is the case then, rather than getting old "not filtered" SOCs we would now be getting newer "filtered" SOCs.

I am still not ENTIRELY clear on whether non-72'd SOCs would actually have failed testing at the higher clock rate, or whether they would simply be untested ones - in which case that would make them no worse than the original batch we thought had been delivered (IMHO). But if they were never physically delivered AND are the result of failing to operate at 720MHz, that would sadden me

 

[craigix]

We could also consider that TI know they all work at 720mhz, and are just extracting more money out of them.

 

[Monk]

We could also consider that TI know they all work at 720mhz, and are just extracting more money out of them.

Yup. Business is a funny thing, and increasingly it seems that business people are extruciatingly devious.

Why whould TI (or anyone) sell the same thing - no difference at all - at two different prices? Why not just sell at the higher price for higher profits?

Because some sales would be lost to cheaper competitors. So, in theory, provided it is not too widely known that both "products" are in fact identical, TI (or another business) could capture customers from two markets with one manufacturing process. Neat!

Obviously they benefit most from selling at the higher price - so it's people likely to buy the guaranteed 720MHz-capable produce that they wouldn't not want learning that the unbadged product is JUST as capable. It's slightly - slightly - to their favour to have buyers of the 600MHz unit thinking that they are getting a 720MHz product at a 600MHz price. Which has traditionally included us - we've been thinking that we're getting a >=900MHz product for a 600MHz price when we're actually getting a 600MHz+ product for a 600MHz price. It's all to do with attitude and managing expectations. The same product in all instances, but just the impression of what it will achieve is enough to alter the impression of the final Pandora.

 

[Enverex]

That's assuming the 600Mhz chips wont in the future be the binned 720 chips. (basically if they test them all for 600, then do extra testing for 720's then lots of the 600s will be binned 720s, if they are doing just a single 720 test and never bothering with a 600 test and just getting rid of them if they don't make the 720 then it'll be like it was originally, the 600's will be purely random).

So lets hope it's the latter, that way the 600's wont be "guaranteed not 720".

 

[Squidge]

This might not even be TI doings. If your willing to pay extra, they'll perform any extra tests you wish before shipping, slapping a slightly different part number on the chip to state they have been through the stated tests. Typically after a short while they will normally open up another part number for the general public to see if anyone else is interested. If they are - more money for them.

Or it might even be to catch up with the 36x line which starts @ 720Mhz and goes upto 1 Ghz.

Whatever it is, we are still just guessing and look at the amount of posts...