How Much Will Arm Be A Handicap To The Pandora?

[jlebrech]

Plus games would still have to be ported and we would have floods of noobs up in arms because game X doesnt run when its a windows directx game for example.

 

[cosurgi]

Angstrom will use .ipk, .opk (new iteration of .ipk), and .deb files. Angstrom is the core of several embedded and handheld distributions
I dindn't know that, and it's a nice surprise. What is the compatibility mechanism for .deb? Is it similar to how 'alien' in debian handles .rpm packages, or is it more integrated?

Ideal for me would be if I could "apt-get source && apt-get build-dep" in Angstrom just like in debian, but probably that would be too good to be true.

 

[Creature XL]

Proper OpenGL will probably eventually make it to the omap, but it will most likely take a year or two.

Besides backwards compatibility with x86, ARM cpus really have no drawback.

I am NOT up to date with armv7. However, I know stuff about armv4 (the ones in Gp32 and Gp2x).
No integer division IS a drawback. And no easy to use FPU instructions. I bet even the m68060 has better performance per MHz (2-scalar integer AND FPU at the same time, or almost at teh same time). And it was able to shut off unused components. It was release 1992 so some 16 years befoe the omap.

I too believe the OMAP and ARMingenerell are super CPU's BUT don't say they are the best EVVAAA!

B2T: arm > x86 for embedded systems

 

[darkblu]

Proper OpenGL will probably eventually make it to the omap, but it will most likely take a year or two.

Besides backwards compatibility with x86, ARM cpus really have no drawback.

I am NOT up to date with armv7. However, I know stuff about armv4 (the ones in Gp32 and Gp2x).
No integer division IS a drawback. And no easy to use FPU instructions. I bet even the m68060 has better performance per MHz (2-scalar integer AND FPU at the same time, or almost at teh same time). And it was able to shut off unused components. It was release 1992 so some 16 years befoe the omap.
you do realise you're comparing a 3-generation-old power-efficient cpu to a top-of-the-crop desktop behemoth of the cisc hayday, right? who's to say how a contemporary-lithography '060 would fare against a v7 arm?


QUOTE
I too believe the OMAP and ARMingenerell are super CPU's BUT don't say they are the best EVVAAA!

B2T: arm > x86 for embedded systems


i believe sindbad's original comment was in the context of mobile computing, where arm have traditionally beaten the crap out of the competition (to put it in your terms, 'arm > all'). and even if we assume arm's performance may not have necessarily been the sole leading factor for that, it surely has played a non-detrimental role. anyway, what i'm saying is, let's give credit where credit is due.

ps: /*pet-peeve rant */

basically, classic cisc's* suffer from the the inefficient isa syndrome, where high performance is achieved at the expense of high power draw**. back in the day when zilog introduced (explicit) register bank switching, to help the 8080 isa out of the pit, intel 'knew better'. thirty-plus years later they'so deep in their own shit, so they struggle to set their foot on the exponentially-growing mobile market (the biggest chip-maker on the planet - oh, the irony).

* gnommish register file, hypertrophied op set.

** for that ultra-heavy decoding stage, and all those extra tricks meant to compensate for penalizing the compiler earlier, by presenting it with a sub-par isa at compile time.

 

[Svartalf]

Angstrom will use .ipk, .opk (new iteration of .ipk), and .deb files. Angstrom is the core of several embedded and handheld distributions
I dindn't know that, and it's a nice surprise. What is the compatibility mechanism for .deb? Is it similar to how 'alien' in debian handles .rpm packages, or is it more integrated?


More integrated. Think of it more in terms of a lighter-weight .deb packaging system. .deb as it is is good for desktops, etc. but isn't so hot for mobile and embedded devices because the metadata which enables things like apt to do it's magic makes for a bunch of extra space consumed. There is metadata involved with .opk and you can do some apt-like things with it, but not everything that apt would do for you.

QUOTE

Ideal for me would be if I could "apt-get source && apt-get build-dep" in Angstrom just like in debian, but probably that would be too good to be true.



Heh... Ideally that would be nice. Unfortunately, you don't have that sort of framework on the device and there's reasons why Ubuntu Mobile Edition wouldn't quite fit on the old NAND size and still is an "iffy" proposition on the current amount of NAND. The apt/deb metadata would be one of those reasons. If you did the metadata for what Angstrom typically does, you'd eat 32+ MB of store just for the metadata. With a laptop or a desktop, typically you've got some space to burn and this isn't all that big a deal. It IS on an embedded device, even the Pandora with as much as it's got in the way of resources.

 

[rokdcasbah]

I think it's kinda funny that x86 processors keep getting faster and faster, but Windows keeps getting more and more bloated. What is the point of all these advances in clock and bus speeds and 45nm power saving toaster ovens just to keep up with Microsoft?

In the end, Microsoft and the inherent flaws in the architecture itself have become the x86s' Achilles heel. Hence the reason why people look for alternative hardware and operating systems for varying purposes. Using an ARM processor in an open source, portable device like the pandora is nothing but a strong point IMO.

That's why they call it the upgrade treadmill

OTOH, when people say stuff like, I had a computer 20 years ago and it did the same stuff, I have to think: maybe they're using crappy programs, or they've forgotten the horrors of the old days, like ISA.

 

[Sphinxter]

Angstrom will use .ipk, .opk (new iteration of .ipk), and .deb files. Angstrom is the core of several embedded and handheld distributions
I dindn't know that, and it's a nice surprise. What is the compatibility mechanism for .deb? Is it similar to how 'alien' in debian handles .rpm packages, or is it more integrated?

Ideal for me would be if I could "apt-get source && apt-get build-dep" in Angstrom just like in debian, but probably that would be too good to be true.

You might be pretty happy then except for using, 'ipkg', rather than, 'apt'.

Plus games would still have to be ported and we would have floods of noobs up in arms because game X doesnt run when its a windows directx game for example.

No more than any other linux user is now.

 

[Creature XL]

Proper OpenGL will probably eventually make it to the omap, but it will most likely take a year or two.

Besides backwards compatibility with x86, ARM cpus really have no drawback.

I am NOT up to date with armv7. However, I know stuff about armv4 (the ones in Gp32 and Gp2x).
No integer division IS a drawback. And no easy to use FPU instructions. I bet even the m68060 has better performance per MHz (2-scalar integer AND FPU at the same time, or almost at teh same time). And it was able to shut off unused components. It was release 1992 so some 16 years befoe the omap.
you do realise you're comparing a 3-generation-old power-efficient cpu to a top-of-the-crop desktop behemoth of the cisc hayday, right? who's to say how a contemporary-lithography '060 would fare against a v7 arm?



Erm... the main intention was the missing division instruction IS a DRAWBACK.

And just for information purpose I added that the "disableing functional blocks on a clock-by-clock basis" is something which was do able 16 yars ago.
Then I got carried away, cause I love the 060 (that is on the expansion card for the ATARI Falcon030 which I pre-paid 2 years in advanvce )


EDIT:
Additionally, the 1982 or so developed 68000 had a div instruction and it had embedded versions aswell (like the 060 has Lowpower versions )

 

[darkblu]

Erm... the main intention was the missing division instruction IS a DRAWBACK.

and my remark was about the fact that many things had changed since then, but i was wrong there - v7 has hw division only in the M and R profiles, the A profile (pardora's cortex) is still div-less. my bad. comes to show how far i'm from arm : )