Die Shrunk Omap3: Lower Power, Higher Speeds, Better Graphics

[lambertus]

TI has announced four additions to its OMAP3 family of ARM SoCs (system-on-chips). The new OMAP36xx series switches to a 45nm production process, delivering speeds up to 1GHz, a 25 percent reduction in power consumption, and 75 percent better graphics performance, says TI.

That's an impressive boost in graphics speed and battery runtime. It's probably way too late for Pandora to switch to the new OMAP 3640 SoC .... is it?

http://www.linuxdevices.com/news/NS8848023892.html

Update: Apparently ppl do not get the humor I tried to put in the subtitle and post. So to clarify: I don't want Pandora to change last minute, just to report that the processor used in the Pandora has evolved to a next level. Maybe interesting for next Pandora releases...

 

[lulzfish]

Yes, it is increidbly way too late.

If the Pandora was a huge x86 desktop system, you could just switch out the chips, because most CPUs share sockets within their generation.

However, as far as I can guess, the Pandora's chips change pinout slightly every time, and the tiniest change in one pin or in dimensions will require a full re-confoobulating of the PCB, which it is far, FAR too late for.

 

[Zarneth]

That's an impressive boost in graphics speed and battery runtime. It's probably way too late for Pandora to switch to the new OMAP 3640 SoC .... is it?
Too late to switch?

Long answer: Not at all. Assuming you don't mind waiting an extra 4-12 months and paying an increased price to cover all the extra work to redo the board. Of course by that time there'll be something even better that's just been released. It's a miracle they slipped in the memory upgrade a while back as is.

Short answer: Yes it's too late. It was too late about 3-6 months back.

Even shorter answer: Go away.
(Sorry, couldn't resist )

 

[Kyo]

I can out-shorten that:

yes.

 

[lambertus]

That's an impressive boost in graphics speed and battery runtime. It's probably way too late for Pandora to switch to the new OMAP 3640 SoC .... is it?
Too late to switch?

[...]

Even shorter answer: Go away.
(Sorry, couldn't resist )
I did read the article before posting, so I know it takes a few months before the samples start appearing and that it therefore is too late change the current Pandora design. Doesn't mean it wouldn't be cool to see this chip appearing in future Pandora releases...

There's no need to react the way you did.

 

[arrrgh]

Doesn't mean it wouldn't be cool to see this chip appearing in future Pandora releases...

If you mean in the successor to the Pandora, then yes, that would be great. But if you mean to just slip it in in a future batch, then that would be a very bad idea.

 

[Kramy]

I heard there were licensing issues with 36xx SoCs; lots of binary blobs, or something? Not suitable for most hobby/linux projects?

 

[Mutilator]

There will always be a faster chip coming out. The pandora specs aint exactly weak and can do a grand job with emulators etc. Maybe they will do a version 2 pandora like the gp2x had the f200 etc. but wont be untill they start to get their investments back.

 

[ttcircus]

There will always be a faster chip coming out. The pandora specs aint exactly weak and can do a grand job with emulators etc. Maybe they will do a version 2 pandora like the gp2x had the f200 etc. but wont be untill they start to get their investments back.

LOL. wont be until they actually finish the first one.

 

[Zeno Arrow]

However, as far as I can guess, the Pandora's chips change pinout slightly every time

Yeah right... pinouts on this new chip will not change, it's still part of the same processor family (OMAP3) it's just a more power efficient design (due to 45nm process used). It's not something we can swap in just yet because the new chips are just reaching the engineering sample stage, they won't be ready for mass production for possibly 1-2 years.

I'd have no problem with a F200-style Pandora upgrade down the line (Pandora v1.5), simply because it could offer increased battery life without splitting the community in terms of software (other than the software requiring a very high overclock).

 

[joa]

Until we hear from Craigix I'd say it's way too late. I think they're being a bit optimistic with the increase in performance. Aren't they always?

My worry is that this appears in future pandora's and emulators are optimised for the faster chips. thoughts?

 

[MDave]

Until we hear from Craigix I'd say it's way too late. I think they're being a bit optimistic with the increase in performance. Aren't they always?

My worry is that this appears in future pandora's and emulators are optimised for the faster chips. thoughts?

Simple. Lock out the increased speed, and then you gain an even bigger increase in battery life. Sony did something similar with the PSP, releasing higher spec PSP's but were still locked down to old specs for games.

 

[Tor]

I have a suggestion: Let's wait until the fastest possible processor is out.
..
..right?

 

[OrR]

There's no need to react the way you did.

I disagree. Even if you were joking, this is thread number 9267352 of this nature so it's not funny anymore, just annoying.

 

[Svartalf]

did read the article before posting, so I know it takes a few months before the samples start appearing and that it therefore is too late change the current Pandora design. Doesn't mean it wouldn't be cool to see this chip appearing in future Pandora releases...

Heh... All you gain is a bit of clock speed and a lowering of power consumption for all intents and purposes...

I'd rather we utilized the next batch of tech that follows that one on TI's roadmap. By the time we're doing a successor unit, if we're still at this particular form of insanity, TI will have the OMAP4 rolling out. The OMAP4 uses TWO Cortex-A9's (much, much faster per clock than our A8 and we're already faster than any other ARM NOT using an A8/A9...) in an SMP configuration and a much faster, more capable GPU with an estimated consumption that's in the same general class we're getting right now.

Take something just shy of the XBox-360 in power. Put it in the palm of your hand.

QUOTE

There's no need to react the way you did.



There's no need, but much reason that you got the unfortunate response you did.

There's been an incessant stream of unrealistic requests for things, much like you have, mainly from people that don't "get it" on new consumer electronic device production that wouldn't back down from their position when it was politely explained WHY their pet change was unrealistic.

 

[Mithrildor]

Not getting in the Pandora I think. Its to late for the first batch. And the OP team told us they never want to do the same as with the GP2X F200

 

[greendots]

I just curious... If this is just a die shrink wouldn't it mean that all the pins function the same. Maybe just a smaller foot print on the pcb? It might be too late for the first batch but how is it unrealistic for second or third batch onward if the pricing is acceptable?

 

[grobi]

Imagine we would wait for the first IBM PC, delayed due to problems with the 256kb RAM bank. Intel would have announced the 8088 but our design sticks with 8086. If we waited for 8088, the 80186 would be around and so on.

Most of the enhancements with the XT as well as with the pandora will be on the software side. Let the pandoras live up to its potential first before growing its hardware to the next level. Assuming a backward compatibility, giving the community the chance to build applications and games is much more important than the next big cpu from TI.

grobi

 

[Bronek]

I have a suggestion: Let"s wait until the fastest possible processor is out.

sounds like a plan!

Imagine we would wait for the first IBM PC, delayed due to problems with the 256kb RAM bank. Intel would have announced the 8088 but our design sticks with 8086.

if my memory serves me right, 8088 came before 8086 . Won't argue on this, though

 

[hch]

Imagine we would wait for the first IBM PC, delayed due to problems with the 256kb RAM bank. Intel would have announced the 8088 but our design sticks with 8086. If we waited for 8088, the 80186 would be around and so on.
grobi

not that it matters, but the 8088 is just a 8086 with 8 external bus lines only. therefore it was selected (as the cheaper model) to be used in the original ibm pc...

now for the new omaps... it's just a die shrink! this is the kind of upgrade that e.g. sony does with the ps3 without any anouncement. i am pretty sure that 2 years or so down the road, when omap 35xx will be obsoleted by omap36xx possible new pandora batches will have to use this chip. for now, however, there is this little sentence at the end of the press release: "According to TI, the OMAP3610, OMAP3620, OMAP3630, and OMAP3640 will begin sampling during the third quarter of this year." well, that's it for now i'd say.