Why Qualcomm's Use of ARM Makes Sense

Today, Qualcomm (NASDAQ: QCOM  ) has what is likely the most robust product stack in the mobile silicon game. In addition to a suite of applications processors (with or without integrated modems), the company also offers discrete modems as well as RF Front Ends. There is no semiconductor company today that offers the breadth and depth of Qualcomm's mobile product offerings as well as its revenue base. This has traditionally afforded Qualcomm the R&D budget to custom tailor just about every one of its IP blocks, including the CPU.

Qualcomm's 64-bit stack is all ARM CPU IP
Since Apple (NASDAQ: AAPL  ) made "64-bit" practically a requirement for smartphones thanks to its implementation of the ARMv8 64-bit instruction set in its A7 system-on-chip, the race has been on to get that marketing checkbox checked. Qualcomm, whose former chief marketing officer had previously dismissed 64-bit as a gimmick, very deftly put together an entire 64-bit product stack from the very low end Snapdragon 410 intended for sub-$150 phones to the Snapdragon 810 intended for phablets and tablets.

The interesting thing here is what while Qualcomm has typically used its own processor cores at the high end and mid-range while implementing ARM's (NASDAQ: ARMH  ) designs at the low end, Qualcomm has gone full-on ARM across the stack from the very lowest end to the highest end. The latest-and-greatest Snapdragon 810 sports four ARM Cortex A57 cores running at up to 2GHz as well as four ARM Cortex A53 cores for low power tasks in a big.LITTLE configuration.

Does this matter?
In Qualcomm's press release, the company claimed that it was still hard at work on a custom 64-bit core. While there is little reason to doubt that the company is doing so, there's a real philosophical question that needs to be answered here: Should Qualcomm even bother investing that much in a high-end custom core? While some will say, "Yes, absolutely," I take a contrarian point of view and say that no, it's not actually necessary.

Qualcomm hasn't differentiated much with Krait...
Qualcomm's line of custom CPUs is called Krait, with various numbers following (e.g., Krait 200, Krait 300, Krait 400, and so on). These processors have usually been very good, but neither the performance nor the power consumption of these cores has been dramatically better than those using stock ARM cores. In fact, the ARM Cortex A15 seemed to offer similar to superior performance to Krait 400 in a number of tests in a similar power envelope when the A15 was implemented by a company with a skilled physical design team, like Samsung's.

Further, a mobile system-on-chip and, in particular, a smartphone-oriented system on chip is so much more than a CPU and a graphics processor. In fact, see below for a rough representation of the kinds of blocks you'll typically find in a top-end smartphone processor:

Source: Qualcomm. 

Take a close look. In a typical high-end system-on-chip, you have the following blocks:

  • Memory controller
  • DSP/sensor hub
  • Multimedia engine
  • Image signal processors (i.e. camera)
  • Modem
  • Display processing
  • GPU
  • GPS
  • CPU

In lower-end ones, things like connectivity will be integrated into the chip as well. However, the point is that the CPU is just one part of the entire equation, and it's the integration of all of these IP blocks and delivery of product on time and on target to the customer that matters. If Qualcomm can do a strong physical implementation of the "stock" ARM cores and deliver a highly integrated, fully featured system-on-chip, then neither the OEMs nor end users will care if Qualcomm could have squeezed an extra 5-10% or so of performance/watt from just the CPU block.

Foolish bottom line
While some may get hung up on Qualcomm using "stock" ARM cores, the reality is that ARM's cores are actually quite good and if implemented in a well-balanced system-on-chip surrounded by leadership IP, then the chip is likely to perform well technically and financially. And given that most of Qualcomm's chip volume has included stock ARM cores anyway, this doesn't seem to stop Qualcomm from being the most profitable mobile chip vendor and the most valuable chip company on the planet at a market capitalization of $137.52 billion, ahead of the second-place company at $134.31 billion. 

The biggest thing to come out of Silicon Valley in years
If you thought the iPod, the iPhone, and the iPad were amazing, just wait until you see this. One hundred of Apple's top engineers are busy building one in a secret lab. And an ABI Research report predicts 485 million of them could be sold over the next decade. But you can invest in it right now... for just a fraction of the price of AAPL stock. Click here to get the full story in this eye-opening new report.


Read/Post Comments (4) | Recommend This Article (1)

Comments from our Foolish Readers

Help us keep this a respectfully Foolish area! This is a place for our readers to discuss, debate, and learn more about the Foolish investing topic you read about above. Help us keep it clean and safe. If you believe a comment is abusive or otherwise violates our Fool's Rules, please report it via the Report this Comment Report this Comment icon found on every comment.

  • Report this Comment On April 22, 2014, at 10:44 AM, raghu78 wrote:

    Ashraf

    A15 is much more power hungry than Krait. here is an article comparing Krait vs A15

    http://www.anandtech.com/show/6536/arm-vs-x86-the-real-showd...

    "The Cortex A15 data is honestly the most intriguing. I'm not sure how the first A15 based smartphone SoCs will compare to Exynos 5 Dual in terms of power consumption, but at least based on the data here it looks like Cortex A15 is really in a league of its own when it comes to power consumption. Depending on the task that may not be an issue, but you still need a chassis that's capable of dissipating 1 - 4x the power of a present day smartphone SoC made by Qualcomm or Intel."

    Apple and Qualcomm knew they needed a custom core for better power efficiency. Thats why they stayed away from the A15.

    http://www.anandtech.com/show/7739/arm-cortex-a17

    " On a given process node, the Cortex A17 can occupy around 20% more area than a Cortex A9 or a marginal increase over a Cortex A12 design. Running the same workload, ARM expects the Cortex A17 to be 20% more energy efficient than the Cortex A9 (race to sleep), but I'd expect higher peak power consumption from the A17. The Cortex A17 name was deliberately chosen as ARM expects to be able to deliver similar performance to the Cortex A15 (in mobile apps/benchmarks, likely not in absolute performance), but in a much smaller area and at a lower power. I can't help but wonder if this is what the Cortex A15 should have been from the very beginning, at least for mobile applications."

    For Apple and Qualcomm, there is always value in a custom designed core tuned for best perf/watt. These companies have massive resources and need to invest heavily in differentiating their products if they wish to continue to dominate the markets. Qualcomm dominates smartphone market and its in their best interests to invest R&D in their own custom ARMv8 cores.

  • Report this Comment On April 22, 2014, at 12:08 PM, TMFAeassa wrote:

    raghu78

    Be careful in assuming this is a micro-architecture deficiency on the part of the A15. Remember that Krait ran at relatively low clocks on 28nm LP and only started showing some performance "teeth" on 28nm HPm (HKMG).

    Also be careful in conflating implementation with micro-architecture. A good 28nm HKMG physical implementation of A15 with a focus on low power may have been more successful than the 32nm Samsung A15 in the market. Indeed, the Exynos 5 Octa does very well and clocks nicely within the int'l version of the Galaxy S5.

  • Report this Comment On April 22, 2014, at 1:24 PM, raghu78 wrote:

    Ashraf

    Did you read anand's statement that A15 should have been A17 right from the start. the power efficiency of A15 was not good. Also just to give you what product differentiation is possible with own core. You should know that Cyclone blows every ARM core out there - Krait, A15, A17,A57. It also destroys Silvermont and Jaguar. Apple went with 2 beastly cores. Thats the ideal approach. finish the task fast and race to sleep, Apple was able to do it because of an extremely ambitious custom core.

    Apple is the company Intel and Qualcomm should be watching. Within the next 2 gens Apple will have a SOC which will be every bit capable as Intel's high end SOCs. Apple needs just one more killer feature - a DVFS system like Intel's with aggressive turbo. Clock for clock Cyclone is on par with ivybridge. multithread performance is slightly higher on intel due to HT. Apple could bring a similar feature in due time. Apple A9 vs Broxton is the fight I am waiting for. It will show whether Intel has the balls to make their small core so powerful that it will encroach their high end cores sales and gross margins.

    I expect Apple to go all out with the iphone 6 and next gen ipad air delivering a product which knocks the wind out of Samsung and Qualcomm.

  • Report this Comment On April 22, 2014, at 3:19 PM, TMFAeassa wrote:

    raghu78

    Agreed that Cyclone is the most sophisticated mobile core out there today, but understand that Apple's constraints are far different than Intel's/Qualcomm's/ARM's.

    Apple need not play the "quad core" marketing nonsense, and it can custom tailor the thermal solution of the iPhone/iPad to handle the chip. Further, since Apple controls iOS, there's a lot of efficiency to be exploited there.

    Be very careful in comparing Apple (which is designing chips for just its own product) to general purpose guys like QCOM/INTC. I have no doubt that Apple will continue to put out extremely impressive SoCs with an emphasis on CPU and GPU performance.

    Qualcomm and Intel need to make sure their SoCs are affordable by the OEMs, integrate as much as possible given the die area allocated, and they need to develop and manage a whole suite of SKUs and the associated software.

Add your comment.

Sponsored Links

Leaked: Apple's Next Smart Device
(Warning, it may shock you)
The secret is out... experts are predicting 458 million of these types of devices will be sold per year. 1 hyper-growth company stands to rake in maximum profit - and it's NOT Apple. Show me Apple's new smart gizmo!

DocumentId: 2921503, ~/Articles/ArticleHandler.aspx, 10/23/2014 4:43:39 AM

Report This Comment

Use this area to report a comment that you believe is in violation of the community guidelines. Our team will review the entry and take any appropriate action.

Sending report...


Advertisement