official word is out, x86 murders arm on every front...


monstercameron

Well-Known Member
Joined
Oct 19, 2010
Messages
1,001
Website
www.thinkteletronics.com
anandtech.com just released some info on the power consumption between nvidia's so-called super chip and intels z-series(no amd z-series here). Not inly is the dual core chip faster than nvidias quad core, it uses ~20% less power and reaches idle state earlier. Intel is positioning this chip for cheap phones so prices must be competitive with other arm socs vendors and android does run on the device(although there was a bruhahah about it not working with linux because of something to do with sleep states?). So all you x86 haters, what now?

http://www.anandtech.com/show/6529/busting-the-x86-power-myth-indepth-clover-trail-power-analysis

http://www.engadget.com/2012/10/04/motorola-razr-i-review/ --motorola razr i review
 
Yeah, so-called super chip that's over a year old vs the just released Clovertrail and is on a trailing process from what everyone else has out, that's a great way to make a generalization about everything on ARM. Tegra 3 has also shown very poor power consumption vs other ARM SoCs, go look at the iPhone 5 review. From the review you linked it's apparent that a lot of that is from the GPU (which is also much stronger than Z2760's). In fact you can clearly see the difference in GPU power consumption when idle or near idle is playing a huge role but of course Anand's sensationalist article title wants to draw your attention in a different way (nothing new here). Did you also know that Windows RT isn't even yet supporting Tegra 3's companion core, according to nVidia themselves? Who knows what other problems it has with power management.

And did you stop to think of the implications that Intel themselves performed this test and chose the hardware, but a quick skim will make it seem like it's Anand's own?

This says everything:

http://images.anandtech.com/reviews/SoC/Intel/CTvT3/word-cpu.jpg

http://images.anandtech.com/reviews/SoC/Intel/CTvT3/word-gpu.png

This is actually a refreshing comparison and much better than the usual Javascript bullshit (which he then gets into..). Love Javascript all you want but I can tell you what I'm going to be running on my Pandora.

Tired of people who think ARM is some homogeneous hardware platform and central to every power comparison. There's a LOT of stuff involved here.

When it comes to power consumption and performance comparison what you should be looking at is the upcoming big.LITTLE Cortex-A15 + Cortex-A7 SoCs vs perhaps the upcoming dual core + SGX544MP2 Medfields. Silvermont (22nm) is too far away to enter consideration. Could also look at high end Kraits. Tegra 3 is way too old.

And no Intel is not positioning Cloevertrail for phones (much less cheap phones!), Clovertrail is for tablets. Medfield (single CPU core) is for phones. Both of them have poor GPUs. This GPU is very, very far behind what's in your beloved Ontario, so you'll see a very different story trying to play Windows games (which AFAIK is the only reason you care so much about x86)
 
Last edited by a moderator:
@Exophase wasnt talking about gpu, as you can see, the dual core 32nm clovertrail beats quad core 40nm nvidia tegra 3. Plus the sgx 545 @ 533Mhz should be adequate for the pandora according to some of the discussions I have seen on these forums.
 
Nice emphasis on "dual core" vs "quad core", but do you know if any of the tests performed actually used four threads heavily? Of course I don't think quad core is an important consideration either. I'd much rather have a dual core Cortex-A15. I wouldn't dream of considering Tegra 3 for Pandora 2. The whole comparison is just unbalanced. It's really MS's fault for launching SurfaceRT with an obsolete SoC but that's really neither here nor there.

The process difference can't be ignored. Look at the difference in power consumption Samsung got moving Exynos 4 from its 45nm to 32nm, I've heard reports that it was as high as twice as efficient in some cases (Apple A5 to A5r2 also showed huge gains) - balance that against the 20% better CPU power efficiency you're calling here. It's more than just a simple transition because the new processes introduce HKMG. If we were talking about 22nm Atom SoCs vs 28nm TSMC and 32/28nm Samsung that'd be one thing and probably worthy of serious consideration but that isn't what we're talking about. Intel's 22nm mobile parts are coming much later.

Gaming was your entire argument for x86 in the other thread, I don't know what discussions made you think SGX 545 @ 533MHz is adequate when that's substantially weaker than what other SoCs are providing right now. If you're saying that people don't care about 3D for Pandora 2 then what's the point? I mean outside of the agenda you have. You think people here are biased in favor of ARM and I'm sure a lot are but when you make threads with this thread instead of a more honest "Clover Trail better power and performance than Tegra 3" you look just as biased.
 
I never said I wasnt biased...and a sensational thread makes for a better discussion because it will incur a more passionate response. My arguments for x86 hasn't been just gaming but backwards compatibilty with older software aswell as more software(much more x86 compiled s/w compared to arm).
 
Exophase, I agree with you here. I'm not sure why somebody would need more than 2 threads on a tablet or handheld at this point in time. I'm sure it's nice, but with the current generation of UI, I can't imagine wanting to run more than 2 processes at a time. I guess, one could want to run something in the background and something else in the foreground and give them both multiple cores, but I'm not sure I see that as being particularly desirable at this time.

I'd personally rather have either more battery life or a faster processor for the money. Now, for things like Raspberry Pi or the Beagleboards that are mostly just kept plugged in, that probably isn't true. But, then again, they have a power cord so it's not as important if they have good battery life.

I saw these articles posted elsewhere and it seemed a lot like Intel trying to preserve it's instruction set so that it can maintain it's current monopoly. Considering that everybody except for Intel has deals to work with ARM or already is, I have to maintain a certain skepticism about whether Intel is really on the right side of history here.

I never said I wasnt biased...and a sensational thread makes for a better discussion because it will incur a more passionate response. My arguments for x86 hasn't been just gaming but backwards compatibilty with older software aswell as more software(much more x86 compiled s/w compared to arm).
No, passion is exactly the opposite of a better discussion, unless of course you're trolling. In general passion and quality discussion aren't like peanut butter and chocolate.
 
the second i opened this thread, i knew there was gonna be a back and forth between monstercameron and Exophase....wasnt disappointed 
 
Last edited by a moderator:
Yeah, of course the intels beat the crap out of the dated socs like Tegra 3...


But I want to see real comparisons against A-15 with big-little technology, too.
 
The Tegra 3 was released on November 9, 2011.

http://en.wikipedia.org/wiki/Tegra

If you're going to talk about possible tech for a Pandora II, then you're talking about the Tegra 4.

  • Wayne
  • Processor: Quad ARM Cortex-A15 MPCore + low power companion core
  • Improved 72 GPU cores with support for DirectX 11+, OpenGL 4.X, and PhysX
  • 28 nm HPL[35]
  • About 20 times faster than Tegra 2 and 6 times faster than Tegra 3[36]
  • Hardware VP8 encoding en decoding
  • To be released in Q1 2013
 
The Tegra 3 was released on November 9, 2011.

http://en.wikipedia.org/wiki/Tegra

If you're going to talk about possible tech for a Pandora II, then you're talking about the Tegra 4.

  • Wayne
  • Processor: Quad ARM Cortex-A15 MPCore + low power companion core
  • Improved 72 GPU cores with support for DirectX 11+, OpenGL 4.X, and PhysX
  • 28 nm HPL[35]
  • About 20 times faster than Tegra 2 and 6 times faster than Tegra 3[36]
  • Hardware VP8 encoding en decoding
  • To be released in Q1 2013
yeah because pandora launched with bleeding edge hardware, tegra 4 isnt even announced yet...also the speculated numbers that nvidia want are just too high for a device like the pandora.
 
yeah because pandora launched with bleeding edge hardware, tegra 4 isnt even announced yet...
Is that sarcasm?

The omap 3530 was cutting edge when the pandora was designed - we only got it because TI wanted a guinea pig to beta test their products
 
^ which also shows that no matter what people discuss in this subforum, by the time the P2 is designed it's already old news and by the time the P2 comes out it will be recognised as obsolete.

This subforum is so useless it's funny :)
 
^ which also shows that no matter what people discuss in this subforum, by the time the P2 is designed it's already old news and by the time the P2 comes out it will be recognised as obsolete.

This subforum is so useless it's funny :)
Obsolete?  Mmm... - no.  The existing Pandora 1Ghz is not obsolete.  There is -nothing- on the market with it's blend of hardware and software capabilities that can fulfill the same collection of roles.

Having one (or more) components that have been surpassed by those used in other devices is, frankly, irrelevant.  Until someone puts one in a handheld full-function laptop with Linux OS, game controls, full sized SD slots and audiophile grade headphone amp, the Pandora will still be the current edge of technology.

Even the Pandora 2 may not make the Pandora obsolete for some of it's user base.
 
Hopefully they can pull off the modular design (ie. Pandora CPU core board + peripheral board) for Pandora 2.

That way the SoC can be updated by swapping a module only. That would open a whole new set of options (in addition to easy upgrades).
 
^ which also shows that no matter what people discuss in this subforum, by the time the P2 is designed it's already old news and by the time the P2 comes out it will be recognised as obsolete.

This subforum is so useless it's funny :)
You don't know that. Pandora's hardware got dated so quickly because Pandora took 1.5 years longer than anticipated before it started shipping. It's a good faith assumption that the next one they build won't necessarily take so long and have so many problems.

Originally Craig was talking about releasing something late 2013, although now I doubt that'll happen. But if it did SoCs released early next year would be the right target to look at.
 
Just out of curiosity I had my bro run the Sunspider test on his ATIV Smart PC which got 771ms, while my Exynos 5 dual chromebook which I just got scored 687ms (lower is better).

Seems to me that things could go either way, between Intel VS ARM. Things are going to get interesting once more A15 ARM SoC come out next year.
 
Last edited by a moderator:
Hopefully they can pull off the modular design (ie. Pandora CPU core board + peripheral board) for Pandora 2.

That way the SoC can be updated by swapping a module only. That would open a whole new set of options (in addition to easy upgrades).
Yeah, a product without a planned obsolescence would be really cool. 
 
Back
Top