Jump to content
Sign in to follow this  
juani_c

Arduino TRE

Recommended Posts

The Arduino Galileo using an Intel CPU was also announced yesterday.  Here is a post from a pre-release tester about its features http://makezine.com/2013/10/03/10-great-intel-galileo-features/

 

I would think the TRE would have most of the same advantages.

The x86 should have died by 1994.  Seriously, the fact that it still exists at all, today, is 90% a matter of luck.  Why in the name of Zeus would I want an x86 for hardware that has absolutely no legacy attachment to it?  Even worse, ARM is the legacy for this segment.  Just stupid, stupid, stupid on Intel.  

 

Sorry, that was a rant.  ;)

Share this post


Link to post
Share on other sites

The x86 should have died by 1994.  Seriously, the fact that it still exists at all, today, is 90% a matter of luck.  Why in the name of Zeus would I want an x86 for hardware that has absolutely no legacy attachment to it?  Even worse, ARM is the legacy for this segment.  Just stupid, stupid, stupid on Intel.  

 

Sorry, that was a rant.   ;-)

I suppose 8051 should've died a long time ago too, but it hasn't. ;-)

Share this post


Link to post
Share on other sites

I suppose 8051 should've died a long time ago too, but it hasn't. ;-)

 

Respectfully, this is a bad argument (even if it is in jest).  The 8051 is cheap, so it has value in being cheap.  x86 is not cheap, nor has it ever been.

 

Just as a supporting example: a lot of times people like to use the "VHS vs. Beta" analogy to describe a bad technology winning, but it's easy to forget that VHS had longer record-time and was cheaper.  It wasn't worse, it was the better technology in all areas except picture quality.

 

The actual history is that x86 is still around because IBM was arrogant and stupid regarding OS/2, and because in Intel's modern x86 markets, 20k extra transistors hardly matter.  But in things like Atom and Quark, it's just a bad choice -- especially Quark.  The "run software for x86" argument just doesn't hold anymore.  Even if it did, ARM has become the incumbent in these sectors.

 

Sorry for a rant, but this is an interesting topic to me. :)

Share this post


Link to post
Share on other sites

Why in the name of Zeus would I want an x86 for hardware that has absolutely no legacy attachment to it?

The same could be said for ARM's attempt to move into the data center but its not stopping them from trying.  ;-)

Share this post


Link to post
Share on other sites

The same could be said for ARM's attempt to move into the data center but its not stopping them from trying.  ;-)

If ARM can use less energy per DMIPS, there is actually value there.  I'm not sure it can (I have no idea), but heat is a big problem in data centers.

Share this post


Link to post
Share on other sites

If ARM can use less energy per DMIPS, there is actually value there.  I'm not sure it can (I have no idea), but heat is a big problem in data centers.

And Intel feels they can use their undeniable lead in chip fab to make up for any inherent power deficiencies in their chip design.

 

Just as in sports, we have to sit back and watch them play the game to see who wins.

Share this post


Link to post
Share on other sites

And Intel feels they can use their undeniable lead in chip fab to make up for any inherent power deficiencies in their chip design.

 

Just as in sports, we have to sit back and watch them play the game to see who wins.

Intel can do what Intel wants.  The bewilderment I have is less about Intel trying to go into embedded markets, and more about Intel trying to take the x86 into embedded markets.  Nobody cares about x86, outside of the desktop and server market where it's the incumbent.  I'm upset that they aren't trying harder, going x86 is lazy and stupid, and history will repeat itself -- Intel has historically mailed-it-in for embedded, and failed repeatedly.  Ironically, Itanium2 might actually be a good embedded ISA, if scaled-down a bit.  But, they are mailing it in, so don't expect anything that clever.

 

Regarding transistor fab, they have an advantage with cellular baseband, but not in sleepy IoT stuff, which is where Quark is supposed to fit in.  For parts that have low duty cycle, leakage is a bigger problem than switching voltage.

Share this post


Link to post
Share on other sites

@@jpnorair  So do you think Intel made the wrong move selling their XScale PXA to Marvell?

No, that was a good move because XScale was dead-end technology, so selling it was good business.

 

Embedded is a tougher market than desktop & server.  You need to sell a lot more chips to make the same revenue, and since the barrier-to-entry is low, there is more competition and thus lower profit.  So Intel isn't likely to put in the effort to do something meaningful in embedded markets -- they've tried a few times in the past, and it has never really worked-out.  

 

That's what I mean by "mailing it in."  The the Quark is just recycled Atom technology, and the Atom is recycled Pentium-1 technology.  Intel will never spend money on embedded R&D, because they don't believe they can win at it.  The market for Quark is niche, which Intel knows and they are fine with it.  ARM Cortex-A is going to deliver better performance at lower price and lower power.  Intel knows.  So, as an engineer it pisses me off that Intel -- who has the ability to do something great -- is just happy to address a niche with a mediocre product.  On the other hand, if they are able to make good money in their integrated cellular baseband IC business (which they assembled recently), this is an area where we might see some real innovation from them.

Share this post


Link to post
Share on other sites

Perhaps Intel does not wish to potentially "bleed" money off their revenue. They make very good desktop / server processors. Processors which ARM could never hope to compete with. Power usage, and raw performance.

 

ARM on the other hand does very well in the limited performance / power usage system market. Plus despite the Cortex A being closer to CISC( vs RISC ) now days. ARM does not have all that x86 bloat that comes with desktop class processors.

 

Personally, I'd like to see an ARM processor that was capable of performing as well as an Intel desktop classed processor. However, then It'd be ARM who would be bleeding money trying to keep up with Intel.

 

So what we're really talking here, is apples vs oranges.

Share this post


Link to post
Share on other sites

@@yyrkoon : be careful, you are treading on delicate ground.  It is important to separate Quark from modern, x86 desktop processors.  They are hugely different, as different as Pentium vs. PowerPC.  Anyway, some points of reflection:

 

- Cortex A is not closer to CISC.  CISC is completely dead, in a practical sense.  The x86 isn't really CISC anymore.

 

- "x86 bloat" is about 20k transistors, which is the part that converts legacy x86 ISA into microcode.  Modern compilers tend to bypass this, anyway.

 

- ARM fully intends to move upmarket.  Whether they succeed or not, we will find out soon enough.  Everything they've done so far has been from the Lotus school of design, so I would bet that they can come up with an application-focused chip that will challenge Intel, in one of Intel's markets, and it will utterly crush x86 in that market within 3 years of arrival.

Share this post


Link to post
Share on other sites

@@jpnorair Either way, x86 is not RISC, and some would argue that neither is an ARM Cortex A processor. Depends on which definition you go by I suppose. Quite honestly, I have not taken a close look at the Quark. I suppose I assumed it was x86 at least in a basic sense. Perhaps I assumed wrong. All I am concerned with from a personal standpoint is *if* it can run standard x86 software. If this is the case, then all the issues that come with other architectures should go away. *shrug*

 

I think ARM would do well in low power, where performance is not necessarily the motivating force. Personally I like everything ARM I have used so far, although sometimes it can be a major PiTA to work around software issues. This is of course in the context of Linux. This is no ones fault really, just that x86 has much better support thus far. It *is* getting better however. Debian for instance runs very well on ARM, with minimal software related issues. Most of the "important" packages are there and work, as well as many, many others.

 

Anyway, I think that if ARM tries to compete with Intel head-to-head they will fall by the wayside like Cyrix, then Centuar. However  I do personally think there is plenty of room for ARM to get into the personal/server market on a small scale. Where low power is important, and ( again ) all out performance is not important. Many systems could fall into this category.

 

An important thing to note however is that ARM makes nothing. They sell their IP to others ( many others as far as I know ). nVidia for instance makes( or at least has made a few ) "mean" processors. As have others. But the best performing ARM system I have read about so far(an allwinner processor ), beats an Intel i3 in most benchmark tests. Compared to the last several years, I think this is significant. However, there is a long ways to go, and lots of places to fall on the way up.

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Sign in to follow this  

×
×
  • Create New...