Stories
Slash Boxes
Comments

SoylentNews is people

SoylentNews is powered by your submissions, so send in your scoop. Only 17 submissions in the queue.
posted by Fnord666 on Wednesday November 04 2020, @06:03PM   Printer-friendly

Arm Cortex-A78C core supports up to 8 cores per cluster, 8MB L3 cache for always-on laptops

Arm Cortex-A78 CPU core was first introduced in May 2020 with a focus on mobile devices like smartphones and was followed by Cortex-A78AE for automotive and industrial embedded applications in September.

The company has now introduced a new variant with Arm Cortex-A78C supporting up to eight cores per cluster, a larger cache up to 8MB for higher performance, and advanced security features all designed for always-on laptops and other "on-the-go" devices.

[...] All those improvements will provide increased performance in laptops, likely at the cost of higher power consumption, but considering Arm laptop often get over 20 hours of battery life, it may be a worthwhile compromise to lose a couple of hours of battery life for higher performance.

This is being seen as a reaction to Apple's custom ARM SoCs for Macs, which are expected to be announced within a week. A successor to the Qualcomm Snapdragon 8cx could use 8 "big" cores.

Also at Wccftech.

Previously: ARM Announces Cortex-A78 and Cortex-X1


Original Submission

Related Stories

Qualcomm Announces Snapdragon 8cx, an ARM Chip Intended for Laptops 17 comments

Qualcomm announces the Snapdragon 8cx, an 'extreme' processor for Windows laptops

The "X" stands for "extreme." That's what Qualcomm's marketing department wants you to think about the new eight-core Snapdragon 8cx.

It's a brand-new processor for always-connected Windows laptops and 2-in-1 convertible PCs, and from Qualcomm's perspective, it might seem a little extreme. Physically, it's the largest processor the company has ever made, with the most powerful CPU and GPU Qualcomm has devised yet. Qualcomm says it'll be the first 7nm chip for a PC platform, beating a struggling Intel to the punch, and the biggest performance leap for a Snapdragon ever. The company's promising "amazing battery life," and up to 2Gbps cellular connectivity.

The TDP is 7 Watts, and the chip supports up to 16GB of LPDDR4x RAM.

Previously, a "Snapdragon 1000" for laptops was said to be in the works, but with a 12 Watt TDP.

See also: Firefox running on a Qualcomm 8cx-powered PC feels surprisingly decent

Previously: First ARM Snapdragon-Based Windows 10 S Systems Announced
Snapdragon 845 Announced
ARM Aims to Match Intel 15-Watt Laptop CPU Performance
Intel Reportedly "Petitioned Microsoft Heavily" to Use x86 Instead of ARM Chips in Surface Go


Original Submission

ARM Announces Cortex-A78 and Cortex-X1 10 comments

Arm's New Cortex-A78 and Cortex-X1 Microarchitectures: An Efficiency and Performance Divergence

Today for Arm's 2020 TechDay announcements, the company is not just releasing a single new CPU microarchitecture, but two. The long-expected Cortex-A78 is indeed finally making an appearance, but Arm is also introducing its new Cortex-X1 CPU as the company's new flagship performance design. The move is not only surprising, but marks an extremely important divergence in Arm's business model and design methodology, finally addressing some of the company's years-long product line compromises.

[...] The new Cortex-A78 pretty much continues Arm's traditional design philosophy, that being that it's built with a stringent focus on a balance between performance, power, and area (PPA). PPA is the name of the game for the wider industry, and here Arm is pretty much the leading player on the scene, having been able to provide extremely competitive performance at with low power consumption and small die areas. These design targets are the bread & butter of Arm as the company has an incredible range of customers who aim for very different product use-cases – some favoring performance while some other have cost as their top priority.

All in all (we'll get into the details later), the Cortex-A78 promises a 20% improvement in sustained performance under an identical power envelope. This figure is meant to be a product performance projection, combining the microarchitecture's improvements as well as the upcoming 5nm node advancements. The IP should represent a pretty straightforward successor to the already big jump that were the A76 and A77.

[...] The Cortex-X1 was designed within the frame of a new program at Arm, which the company calls the "Cortex-X Custom Program". The program is an evolution of what the company had previously already done with the "Built on Arm Cortex Technology" program released a few years ago. As a reminder, that license allowed customers to collaborate early in the design phase of a new microarchitecture, and request customizations to the configurations, such as a larger re-order buffer (ROB), differently tuned prefetchers, or interface customizations for better integrations into the SoC designs. Qualcomm was the predominant benefactor of this license, fully taking advantage of the core re-branding options.

[...] At the end of the day, what we're getting are two different microarchitectures – both designed by the same team, and both sharing the same fundamental design blocks – but with the A78 focusing on maximizing the PPA metric and having a big focus on efficiency, while the new Cortex-X1 is able to maximize performance, even if that means compromising on higher power usage or a larger die area.

While Cortex-A78 will only improve performance by around 7% from microarchitectural changes alone, Cortex-X1 will improve performance by up to 30% due to a wider design, doubling of most cache sizes, and other changes. Cortex-X1 cores are also expected to reach 3 GHz on a "5nm" node, delivering even more performance. The Cortex-X1 cores could use up to 50-100% more power than Cortex-A77/A78. Cores could be arranged in a 1+3+4 or 2+2+4 setup of Cortex-X1, Cortex-A78, and Cortex-A55 cores.

See also: Arm Announces The Mali-G78: Evolution to 24 Cores


Original Submission

This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
(1)
  • (Score: 1, Insightful) by Anonymous Coward on Wednesday November 04 2020, @07:07PM (2 children)

    by Anonymous Coward on Wednesday November 04 2020, @07:07PM (#1073028)

    how about "always-off", completely the hell off, when you select off?

    • (Score: 2) by takyon on Wednesday November 04 2020, @07:20PM

      by takyon (881) <{takyon} {at} {soylentnews.org}> on Wednesday November 04 2020, @07:20PM (#1073049) Journal

      It needs to be always-on so you can remain connected to social media.

      --
      [SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
    • (Score: 0) by Anonymous Coward on Wednesday November 04 2020, @07:47PM

      by Anonymous Coward on Wednesday November 04 2020, @07:47PM (#1073075)

      Back to the brain reprogramming center with you!

  • (Score: 2) by Azuma Hazuki on Wednesday November 04 2020, @09:32PM (7 children)

    by Azuma Hazuki (5086) on Wednesday November 04 2020, @09:32PM (#1073143) Journal

    I've seen some ridiculous perf-per-watt benchmarks out of Apple A-series silicon. How does the A78C compare? I'd definitely be interested in something small, compact, and powerful if it's at least Haswell level in, say, 1/8th the power envelope.

    But miss me with that "always-on" bullshit. I want a laptop with the battery life of a Moto Power-series phone and similar ability to idle for a week straight, not to be wasting its cycles and battery sending heaven knows what telemetry to hell knows where.

    --
    I am "that girl" your mother warned you about...
    • (Score: 2) by takyon on Wednesday November 04 2020, @09:57PM (5 children)

      by takyon (881) <{takyon} {at} {soylentnews.org}> on Wednesday November 04 2020, @09:57PM (#1073154) Journal

      ARM designs have typically been slower than custom Apple ARM designs. Cortex-X1 should be faster than Cortex-A78C, and even that is unlikely to catch up to the latest Apple cores.

      https://www.androidauthority.com/arm-cortex-x1-vs-apple-1121289/ [androidauthority.com]

      --
      [SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
      • (Score: 2) by jasassin on Thursday November 05 2020, @06:56AM (4 children)

        by jasassin (3566) <jasassin@gmail.com> on Thursday November 05 2020, @06:56AM (#1073315) Homepage Journal

        ARM designs have typically been slower than custom Apple ARM designs.

        I'm confused. ARM licenses their technology, and Apple makes better custom chips. What exactly is Apple paying for? An ARM instruction set basically? Sorry, I don't quite get it.

        --
        jasassin@gmail.com GPG Key ID: 0x663EB663D1E7F223
        • (Score: 2) by driverless on Thursday November 05 2020, @07:21AM

          by driverless (4770) on Thursday November 05 2020, @07:21AM (#1073318)

          What exactly is Apple paying for? An ARM instruction set basically?

          At the moment, yes, while they still need ARM, but give it a few years...

        • (Score: 3, Informative) by TheRaven on Thursday November 05 2020, @04:37PM

          by TheRaven (270) on Thursday November 05 2020, @04:37PM (#1073455) Journal

          What exactly is Apple paying for?

          In the first instance, for a fairly exhaustive compliance test suite so that each generation of their implementation is pretty much guaranteed to be backwards compatible.

          The main thing that they're getting of value is to outsource a large amount of the ecosystem cost. Apple develops their OS, but their toolchain and assembly fast paths in a load of libraries that they (and third-party software on their platform) use are not Apple-only. This is a big cost reduction for them, maintaining your own ISA is estimated to cost at least one or two billion dollars.

          --
          sudo mod me up
        • (Score: 2) by TheRaven on Monday November 09 2020, @10:19AM

          by TheRaven (270) on Monday November 09 2020, @10:19AM (#1075081) Journal
          Oh, one more thing: the Arm partnership agreement is a broad patent cross-licensing agreement that covers patents held by any of the members, for the purpose of implementing Arm ISAs. Anything Samsung, Qualcomm, nVidia, or whoever has developed that is covered under that agreement is available for Apple to use in Arm cores, but it would not be available for implementing other cores. This also gives some good defensive cover against Intel. I don't know if there's a cross-licensing deal formally between the Arm partnership and Intel, but if not then I am reasonably confident that Intel is almost certainly violating at least one patent held by the Arm ecosystem and so would kick off a load of counter-suits if they tried to sue anyone for implementing Arm. If they did the same against an Apple ISA then none of the defensive patents would apply and Apple would have to find some Apple-owned patents that Intel infringed (much harder).
          --
          sudo mod me up
        • (Score: 2) by takyon on Saturday November 21 2020, @01:45AM

          by takyon (881) <{takyon} {at} {soylentnews.org}> on Saturday November 21 2020, @01:45AM (#1080049) Journal

          AFAIK, Apple has a "perpetual license" for the ARM architecture for being early [wikipedia.org], and has better terms than the other members.

          --
          [SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
    • (Score: 3, Interesting) by takyon on Thursday November 05 2020, @09:01PM

      by takyon (881) <{takyon} {at} {soylentnews.org}> on Thursday November 05 2020, @09:01PM (#1073541) Journal

      This is from today's AnandTech 5950X and 5900X review:

      Page 3 [anandtech.com]

      Being an x86 core, of the difficulties of the ISA is the fact that instructions are of a variable length with encoding varying from 1 byte to 15 bytes. This has been legacy side-effect of the continuous extensions to the instruction set over the decades, and as modern CPU microarchitectures become wider in their execution throughput, it had become an issue for architects to design efficient wide decoders. For Zen3, AMD opted to remain with a 4-wide design as going wider would have meant additional pipeline cycles which would have reduced the performance of the whole design.

      Page 4 [anandtech.com]

      I do hope that these designs come in a timely fashion with impressive changes, as the competition from the Arm side is definitely heating up, with designs such as the Cortex-X1 or the Neoverse-V1 appearing to be more than a match for lower-clocked Zen3 designs (such as in the server/enterprise space). On the consumer side of things, AMD appears to be currently unrivalled, although we’ll be keeping an eye open for the upcoming Apple silicon.

      Page 9 [anandtech.com]

      https://images.anandtech.com/graphs/graph16214/119125.png [anandtech.com]

      In the performance per clock uplifts, measured at peak performance, we’re seeing a 20.87% median and 24.99% average improvement for the new Zen3 microarchitecture when compared to last year’s Zen2 design. AMD is still quite behind Apple’s A13 and A14 (review coming soon), but that’s natural given the almost double the microarchitectural width of Apple’s design, running at lower frequencies. It’ll be interesting to get Apple Silicon Mac devices tested and compared against the new AMD parts.

      [...] What I hope to see from AMD in future designs is a more aggressive push towards a wider core design with even larger IPC jumps. In workloads that are more execution bound, Zen3 isn’t all that big of an uplift. The move from a 16MB to a 32MB L3 cache isn’t something that’ll repeated any time soon in terms of improvement magnitude, and it’s also very doubtful we’ll see significant frequency uplifts with coming generations. As Moore’s Law is slowing, going wider and smarter seems to be the only way forward for advancing performance.

      https://twitter.com/andreif7/status/1324431700663930889 [twitter.com]
      https://twitter.com/andreif7/status/1324436277970829315 [twitter.com]

      Obviously, there are things that can be improved, and advantages that could be leveraged to perform better when compared to Apple/ARM. One of the things rumored for Zen 4 is a nice big (gigabytes) L4 cache (e.g. HBM) stacked on top of the I/O die. Ultimately, decreasing the distance between cores and memory is one of the best ways to improve performance, and putting towers of cache around is an intermediate step. The L3 cache could be layered up [soylentnews.org], for instance.

      --
      [SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(1)