the open pandora 2 must use an amd apu!


You know, it's really hard to want to talk to you when you selectively ignore 90% of what I say :/


To answer your question, no, I don't think the people here enjoy being contrary. There's a difference between liking less mainstream things and liking things because they're less mainstream.
 
Seriously though, I don't really get you. First you say that OPT should use an SoC with an integrated FPGA, which would need mostly new development work to utilize it. Then you say it should use x86 to avoid having to do new development to utilize it. Now you're saying MIPS. I'm starting to think you're just enamored with bucking the status quo, or enjoy being contrary.

People accuse me of being contrary as well. I read once that a living hell for an "INTJ" is courtroom guard because you have to listen to arguments all day and not get your two bits in... :)


The reason why this topic is interesting and frustrating to me is that there seems to have been serious effort put in by the x86 guys to the point where they are about to be competitive in portable devices. The response seems to be a repetition of the dogma and superstitions that we've all known for years. ARM is lower power consumption, x86 are beasts that will burn your legs and club baby seals.


RISC vs CISC is probably the oldest nerd-religious debate there is. IMO: at some point (soon?) the CISC guys will iterate out the drawbacks of their chips and the RISC guys will make chips as powerful as the x86 ones. Eventually they'll meet in the middle and the only difference is support.


AMD has 28nm chips coming next year and Intel has 22nm Atoms on their way likely late next year.


IMO, if x86 catches up before the software gap is closed then ARM will be overtaken very cleanly.


I'm not familiar enough with the currently available x86 and ARM SoCs and their TDP ratings, so I have to assume you and OPT are making good decisions based on a standard requirements/analysis/decision process. I guess what tweaked me to chime in is the answer is "ARM is better for portable" rather than "ARM is better for portable on our timelines, but x86 is making it interesting!"
 
This is a pretty interesting debate, from my perspective what I and even other fairly geeky consumers are interested in from chipset hardware is what software it will run in a capable fashion & what sort of battery life / heat level it will have, not the underlying architecture.


If it came to a choice of x86 that would run perhaps at 75% the speed and 75% of the battery life of the Arm equivalent for a similar price then it would be a highly difficult call. As the software base for x86 via Linux & Windows is a huge plus, my gut feel is that if this 75% & 75% figure could be reached then the sensible option would be to go with x86.


However as has been stated this is not currently the case, nor is it likely to be in the near future. However, you can bet that Intel & AMD will have been / will be, spending millions on making this happen eventually.


A big question for me is when can we realistically expect the P2 to come to market? my guess would be 2014 or 2015 at the earliest based on the understanding that the existing debts & pre orders of OPLtd would have to be cleared before a probably hugely expensive prototyping phase could be entered.


Who knows how the x86 vs Arm debate will look then?


One thing's for sure, it'll be interesting to find out, as a betting man my money would be on Arm, but it does no harm to hedge your bets.
 
Last edited by a moderator:
Speaking of Flame Retardant Suit.

Parallella Computer Specifications


The following list shows the major components planned for the Parallella computer:

  • Dual-core ARM A9 CPU
  • Epiphany Multicore Accelerator (16 or 64 cores)
  • 1GB RAM
  • MicroSD Card
  • USB 2.0 (two)
  • Two general purpose expansion connectors
  • Ethernet 10/100/1000
  • HDMI connection
  • Ships with Ubuntu OS
  • Ships with free open source Epiphany development tools that include C compiler, multicore debugger, Eclipse IDE, OpenCL SDK/compiler, and run time libraries.
  • Dimensions are 3.4'' x 2.1''


Once completed, the Parallella computer should deliver up to 45 GHz of equivalent CPU performance on a board the size of a credit card while consuming only 5 Watts under typical work loads. Counting GHz, this is more horsepower than a high end server costing thousands of dollars and consuming 400W.
 
Last edited by a moderator:
This is a pretty interesting debate, from my perspective what I and even other fairly geeky consumers are interested in from chipset hardware is what software it will run in a capable fashion & what sort of battery life / heat level it will have, not the underlying architecture.


If it came to a choice of x86 that would run perhaps at 75% the speed and 75% of the battery life of the Arm equivalent for a similar price then it would be a highly difficult call. As the software base for x86 via Linux & Windows is a huge plus, my gut feel is that if this 75% & 75% figure could be reached then the sensible option would be to go with x86.


However as has been stated this is not currently the case, nor is it likely to be in the near future. However, you can bet that Intel & AMD will have been / will be, spending millions on making this happen eventually.


A big question for me is when can we realistically expect the P2 to come to market? my guess would be 2014 or 2015 at the earliest based on the understanding that the existing debts & pre orders of OPLtd would have to be cleared before a probably hugely expensive prototyping phase could be entered.


Who knows how the x86 vs Arm debate will look then?


One thing's for sure, it'll be interesting to find out, as a betting man my money would be on Arm, but it does no harm to hedge your bets.

X86 is a horrible inefficient mess these days. That software edge is very quickly being removed. ARM is the future.


A X86 Pandora2 would be such a white elephant... but one that could warm your pockets pretty well.


Well it's modular based so I bet someone will try to cram an X86 in there anyway....
 
Last edited by a moderator:
X86 is a horrible inefficient mess these days. That software edge is very quickly being removed. ARM is the future.


A X86 Pandora2 would be such a white elephant... but one that could warm your pockets pretty well.


Well it's modular based so I bet someone will try to cram an X86 in there anyway....

The software edge is not quickly being removed. It seems to be happening very very slowly. Sure Linaro is doing great work with gcc, and Linus has been pulling more and more ARM specific modules with recent kernels. However if you want the GPU to work properly you're likely dealing with blobs (which constrain you to the supported kernel versions), same if you want to use a DSP or other accelerator.


A year ago I would have agreed with you 100% on the power/watt based on what was available and how things were trending. However it seems to be changing very quickly. Did you read that Anandtech article I posted? Sure benchmarks are an inexact science, but they actually paint a decent advantage for the Clover Trail Atom. As for warming your pocket "Intel claims that in its lowest platform sleep state (S0i3) the SoC's power consumption is below 2mW." So even if that article is skewed, and it likely is, isn't it worth at least looking into?


You should understand I'm totally with you on this. I *WANT* this device. (I want my 1Ghz P1 as well, but that's a different story) I want to help make it a success from the software side. Hacking on something so unique and useful would be a world of fun. I just think that all else being equal, the software will be better, more varied, etc on x86.
 
They can wrap their dead end X86 chips up however they like (yesterdays newspaper would be fitting) but they are the past, inefficient and desperate.


What is the software that you seem to think is missing from ARM devices?
 
Graigix said:
They can wrap their dead end X86 chips up however they like (yesterdays newspaper would be fitting) but they are the past, inefficient and desperate.
there is a difference between being desperate and the need to adjust to different terms

Craigix said:
What is the software that you seem to think is missing from ARM devices?
GIMP and XnView comes to mind


Steam? hmh


anyway I favour the next Pandora to be based on OMAP5
 
Last edited by a moderator:
Graigix said:
They can wrap their dead end X86 chips up however they like (yesterdays newspaper would be fitting) but they are the past, inefficient and desperate.
there is a difference between being desperate and the need to adjust to different terms

Craigix said:
What is the software that you seem to think is missing from ARM devices?
GIMP and XnView comes to mind


Steam? hmh


anyway I favour the next Pandora to be based on OMAP5

What? How is GIMP missing? GIMP was one of the very first things I booted on the first prototype Pandora.


Also XnView comes up on google as being available for arm as long ago as PocketPC?
 
Last edited by a moderator:
They can wrap their dead end X86 chips up however they like (yesterdays newspaper would be fitting) but they are the past, inefficient and desperate.


What is the software that you seem to think is missing from ARM devices?

A kernel that I can take from kernel.org, compile and successfully boot with most hardware supported already?


GPU drivers that supports real OpenGL with actual and regular support from manufacturer (or maybe even open-source like in-house intel GPUs)?


And printing support for cheap printers which provide x86-only linux 'drivers'? Recently bought one samsung wifi printer that refused to work with my arm server aka router/nas.


As I said earlier unfortunately low-power x86 is not there yet (in terms of actual power consumption and drivers) and won't be until late 2013 - 2014 son it's not going to pandora 2 anyway but I don't like that fanatic ARM rhetoric
 
They can wrap their dead end X86 chips up however they like (yesterdays newspaper would be fitting) but they are the past, inefficient and desperate.


What is the software that you seem to think is missing from ARM devices?

Nice hyperbole but you're not backing it up. How about a fact? If your decision making process consists of relying on superstitions then you must have some really intelligent and patient people to work with. I'm not an embedded hardware guy, though I have worked with some more packaged solutions (XScale PXA270 boards) building OS/software images. When I see the following, I wonder if the OPT team has as well.


Idle (Screen On) Web Browsing HD Video Playback


ASUS Transformer Pad Infinity 2.7W 3.4W 3.1W



Apple iPad 2 2.5W 2.6W 2.5W


Apple iPad (2012) 4.3W 4.5W 5.9W


Intel Clovertrail FFRD 2.3W 2.8W 3.0W


Why not respond with "We've seen those and in our use case, they just won't work because of x" or "Hey, that's interesting, thanks for the info".


As for the software side, I should have been more clear. There are fewer developers optimizing software to work on ARM as there are for x86. That doesn't mean that what has been done hasn't been done well, nor does it mean that quick ports aren't valid. I have no doubt that much of the most popular software required on the P1 and P2 will make it there. BUT, if you're running x86 then it's a moot point since ALL of the software is available there and optimization is as simple as running gcc with the right options.


The other side (and what I mentioned) is the driver question. If you're running vendor provided binaries then you'll eventually be stuck at an old revision of a kernel. "Sorry, can't run any wireless-z USB modules because you need kernel 3.7 for that, TI stopped supporting anything past 3.2, and the backport is just too much of a PITA." That is guaranteed to happen.


I am aware that the Atom I refer to has an SGX GPU with somewhat dubious open source support, and is similar to the one in the P1, however this GPU and its variants is used ALL over the place and proper GPL linux support is coming along.
 
Last edited by a moderator:
The reason why there is a lack of interest in debating this is because we all already know the advantages of ARM and that this conversation has been done to death.


No one can be bothered to go though it yet again.


ARM is just all about elegance which we love, it also gives unbeatable battery life and it's the future.


X86 has just had its day. Just like z80 and 68k before it.


Android will soon be the most popular OS in the world, almost every device, laptops included, will move to ARM.


Welcome to the future.

https://www.youtube.com/embed/vvHTN0Yi1t4?feature=oembed
 
Last edited by a moderator:
The reason why there is a lack of interest in debating this is because we all already know the advantages of ARM and that this conversation has been done to death.


No one can be bothered to go though it yet again.


ARM is just all about elegance which we love, it also gives unbeatable battery life and it's the future.


X86 has just had its day. Just like z80 and 68k before it.


Android will soon be the most popular OS in the world, almost every device, laptops included, will move to ARM.


Welcome to the future.

You can just say that you don't care about linux on pandora (I mean some random distribution, not the one that some kernel genius spend a lot of time to make it work) and this will end the discussion, mostly.


I already have 2 andriod devices (phone and tablet) and don't really care how well next pandora or laptop will run android that will rule the world with powerful 10+w arm processors that will surely let a device of the future work more from the battery than x86 cpu with the same tdp because arm is so superior.


If we stop talking about then suddenly arm SoCs become not that great and have to prove they are good enough compared to x86.


for now intel is moving towards highly integrated SoCs that can switch between various power states with low latency (the thing that is norm in current arm SoCs) so I believe that in the end it will be only instruction set difference.


And developers are lazy, they will optimize mostly only where it's absolutely needed no mater what architecture their code compiles for
 
The reason why there is a lack of interest in debating this is because we all already know the advantages of ARM and that this conversation has been done to death.


No one can be bothered to go though it yet again.


ARM is just all about elegance which we love, it also gives unbeatable battery life and it's the future.


X86 has just had its day. Just like z80 and 68k before it.


Android will soon be the most popular OS in the world, almost every device, laptops included, will move to ARM.


Welcome to the future.

https://www.youtube.com/embed/vvHTN0Yi1t4?feature=oembed

I guess we'll see. I don't have a horse in the race, just a spectator. ARMH was the only stock I ever purchased (sold it after it quadrupled for me, no need to be greedy!) because I was as sure as you are of the advantages and the forthcoming dominance. Now I'm not so sure, those stats are interesting if they hold up, and surely AMD and Intel wouldn't be investing so heavily if they didn't think they were on the right track.


Keep your eyes peeled around CES time!


Oh, and as for Android, you are aware of course that Android is just a layer on top the Linux right? Everything I said about binary driver blobs is true and is a huge issue for proper open source development on the ARM platform.
 
Back
Top