mannerisms

Ruminations on the electronics industry from David Manners, Senior Components Editor on Electronics Weekly.

Marketing Driving Mobile Multi-Processing, says ST-Ericsson

The recent spat between Samsung and Qualcomm with Samsung announcing an eight core chip-set for mobile phones and Qualcomm’s CEO dismissing it as a publicity stunt, is put in context by a paper from ST-Ericsson written by Marco Cornero and Andreas Anyuru.

“Multiprocessing has been adopted as a marketing tool,” say Cornero and Anyuru, “this is used to pass the oversimplified message that more processors mean better platforms and performance.”

“The performance of single-processors in mobile platforms is still growing and, from a software performance perspective, it is more profitable today to focus on faster dual-core rather than slower quad-core processors,” say the authors.

Multiprocessing is not adopted from choice – it is a necessity imposed when frequency scaling reaches a chip’s heat dissipation limit.

But scaling continues to deliver more transistors, and multiprocessing is the only way to keep exploiting the increase in transistors.

“Although counterintuitive, one microprocessor at a given frequency consumes more power than two cores at half the frequency,” say the authors, “this is due to the fact that a higher voltage is needed for sustaining higher frequencies, combined with the fact that dynamic power consumption grows quadratically with respect to voltage.”

The penalty for adopting multiprocessing is “a fundamental disruption in software performance scaling.”

For multiprocessing, software has to be written to map efficiently on the multiple processors, i.e. it needs to be “parallelised.”

“Too bad that software paralleliation is still one of the toughest challenges in computing in general, for which no satisfying general-purpose solution in terms of productivity has been found yet, despite the initial enthusiasm and for many intrinsically serial applications there are no solutions at all,” say the authors.

“The problem is so difficult that very little progress has been made in the last decade in PC applications,” point out the authors, “after a decade most software, including demanding games, office applications, multimedia playback and Web browsing, still shows a good use of two processors only, with very few applications (like video authoring) being able to use more. The fact is that, except for very few applications, PC software in large hasn’t been parallelised at all since the introduction of multiprocessing a decade ago. “

The paper concludes that: “Like for PC, software scales proportionally with single-processor performance, while depending on the application, it scales less than proportionally or not at all for multi-processors.”

The situation is, of course, completely different in data centres where multi-processing scales well with the parallel workloads.

As there is still headroom for single processor performance improvement in mobile phones,  why have mobile phone vendors moved to multiprocessing?

“The interesting question is: why have mobile platforms jumped into multiprocessing well before reaching the single-core performance saturation?” asks the paper, “we can think of two main reasons: dual-processors can be exploited effectively by modern operating systems, and the fundamental technologies required, such as cache coherency and multiprocessing capable operating systems, were already in place,” say the authors, “the second reason is aggressive marketing.”

For quad-core the justification is much less. “What is less clear is the motivation to move to quad-processors, since the PC experience tells us that even after a decade of multi-processors in the market, using more than two processors is useless for the large majority of the software,” says the paper, “the only technical explanation would be if mobile software were fundamentally more multiprocessing-friendly than on PCs, but this is not yet the case.”

“The fact is that there are no strong technical reasons. The motivation appears to be all in the marketing,” conclude the authors, “like they did with PCs, people will soon realize that the number of cores in a mobile device does not correspond to true value added for the customers. ”

For web browsing “very little or no advantage at all is measured when moving from dual to quad-processors, for lack of sufficient and well-balanced software parallelism needed for exploiting more than two processors,” says the paper, “given the already mentioned small level of effective software parallelism, a faster dual-processor easily outperforms a slower quad-processor.”

“We have performed similar experiments on other important applications, like mobile video games, application start-up time, and even multimedia, obtaining similar results over and over again: marginal or no benefits at all when moving from dual to quad-processors at the same frequency, and 15-20% faster dual always beating slower quad-processors,” says the paper.

Tags: core processors, cores, dynamic power consumption, performance perspective, qualcomm

Related posts

11 Comments

  1. David Manners
    January 24, 2013 15:40

    How true, how very true, david, I think we agree on this.

  2. david
    January 24, 2013 15:36

    Doing things just because you can is hardly a strong motivation.
    And I speak as a semiconductor HW veteran with 28 years experience.
    The issue is not how you can put more cores down on silicon.
    It is how you program them!
    “The semiconductor industry threw the equivalent of a Hail Mary pass when it switched from making microprocessors run faster to putting more of them on a chip—doing so without any clear notion of how such devices would in general be programmed” – Dave Patterson [1]
    Check out the Swedsoft report [2] on multicore SW development which led to the establishment of the Swedish multicore initiative
    They identify that 50% of their exports are SW critical, and dependent on multicore software development which is 2-3x more expensive, but only 1% of programmers have the necessary skills.
    [1] http://spectrum.ieee.org/computing/software/the-trouble-with-multicore
    [2] http://www.swedsoft.se/Swedsoft_SRA_2010.pdf.

  3. David Manners
    January 24, 2013 15:16

    Yes david, ST seem to operate in a parallel universe but, on STE’s point about the usefulness of multi-cores, I would point you to Linley Gwennap’s comments on PCs in a recent Microprocessor Report: “Performance is increasing at just 10% a year for desktops and 16% for laptops – a far cry from the good old days of 60% annual performance increases.” So neither multi-core nor process has done the trick in PCs, whereas, as you say, process can do the trick in mobile. I assume ST have a good reason for persisting with FD-SOI despite their imminent jettisoning of STE. I don’t know what that reason is, but it might turn out to have superior characteristics in volume production to finfet, and the French government could be willing to underwrite it as their shot at La Gloire.

  4. david
    January 24, 2013 13:52

    As usual ST have got it completely wrong in terms of marketing strategy.
    Don’t forget they’re not used to this and have a history of doing captive foundry work for the likes of Nokia.
    This has meant they don’t have a good understanding of the market outside of what their customers tell them and their focus is on customer relationship management.
    As a result they got completely blind-sided by the collapse of Nokia.
    They’re trying to undermine Big.Little when really what they have to offer is a power efficient SOI process.
    Presumably this is a stalking-horse for foundry business for people looking for an alternative to FinFETs?
    A much better strategy would have been to say we can deliver excellent compute efficiency in terms of GOPS/W irrespective of the number of cores.
    Instead they’ve decided to take everybody on in a battle they can’t win and will inevitably lose.
    The likelyhood is that they’ll be dead very soon anyway as ST and Ericsson will no longer bankroll them … doh!

  5. January 23, 2013 12:10

    2 is better than 1, 4 is better than 2 and 8 is even better. If MTK can announce an 8-core SoC for mobile, so can Samsung. Welcome to the CPU number game. This is not about system engineering, but about getting consumers to choose one black brick over another. Qualcomm can win this game, but they need to be smart – they have close to 8 cores already (including the DSPs), though not in a symmetric multiprocessing configuration. So they can join the marketing hoopla rather easily. Of course the number game will peter out one day (ex., does anyone know how many trillion colors are there on their flat TV anymore ?)….

  6. Bitter
    January 22, 2013 20:47

    Yes, this bug is holding back my trolling too..

  7. David Manners
    January 22, 2013 08:16

    Oh God I’m sorry about that, SEPAM, I keep falling foul of that myself, it’s incredibly annoying. My apologies. I will try and prevail upon IT to get it changed.

  8. SecretEuroPatentAgentMan
    January 22, 2013 06:27

    Yes, I was trying to add a comment on Qualcomm but get the error messages about too many comments.

  9. David Manners
    January 21, 2013 21:43

    Absolutely, SEPAM, the uncracked problem is spreading a load evenly between general purpose cores. Which is, presumably, why – as you wisely pointed out on an earlier occasion – Qualcomm’s Krait has, and I quote from your comment: “2 Krait CPUs, 1 GPU, 3 Hexagon DSPs and one multimediaprocessor.”

  10. SecretEuroPatentAgentMan
    January 21, 2013 21:37

    > Too bad that software paralleliation is still one of the toughest challenges in computing in general
    Hmm, well this is true for symmetric multiprocessing. Some bypass the problem by taking an asymmetric approach: dedicating cores to specific tasks such as decoding JPG, decompressing and more.

  11. Bitter
    January 21, 2013 15:21

    Such a cute announcement by STE, perhaps they should work on delivering their SOI quad core products than making up excuses for delivering yesteryears technology tomorrow.