Arm Announces the Cortex-X3, Cortex-A715 CPU Cores and Immortalis-G715 GPU

[TheLostSwede]

This time last year, I wrote about how digital experiences had never been more important, from personal to business devices - they helped us stay connected and entertained at a time when we needed it most. Compute continues to define our experiences in the modern world, and now these experiences are becoming even more visual.

Smartphones are at the center of our connected lives. From gaming to productivity, through video calling, social media or virtual environments, it is the device that provides us the connection to everyone and everything, in real time. For developers, making these immersive real-time 3D experiences even more compelling and engaging requires more performance. Arm sets the standard for performance and efficient compute, and our latest suite of compute solutions for consumer devices will continue to raise the threshold of what's possible in the mobile market, shaping the visual experiences of tomorrow.




Optimizing visual experiences with Arm Total Compute Solutions
Arm's Total Compute strategy has always been anchored in developer access, security and compute performance. We set out to deliver uncompromised performance leadership for all consumer device markets and we're constantly pushing the boundaries on what our partners can create with optimized system design and implementation.

Arm's 2022 Total Compute Solutions (TCS22) offer different levels of performance, efficiency and scalability, and bring together all elements of the Total Compute strategy. The combination of Arm IP launched in TCS22 will offer up to 28% more performance and up to 16% power reduction across a range of workloads, such as gaming, where benefits will include longer play time. We continue to expand the dimensions of performance beyond general-purpose workloads to workloads requiring specialized processing, propelling mobile technology, not just on the GPU but across CPUs and System IP too.

Gaming performance unleashed with Arm Immortalis and Mali GPUs
While mobile device innovation has provided phenomenal growth in mobile gaming, the rise of highly visual AAA gaming experiences demand performance that can keep up. This year, we are launching a brand-new flagship GPU called Immortalis, built on the heritage of Mali, the world's most shipped GPU, but configured and enhanced to deliver the ultimate mobile 3D experiences. Immortalis-G715 is the first Arm GPU to offer hardware-based ray tracing support on mobile, delivering more realistic and immersive gaming experiences.

In addition, we are launching the new premium Arm Mali-G715 GPU, which includes variable rate shading - a graphics feature available across all new GPUs - to deliver significant energy savings and a further gaming performance boost.

Last year we introduced Mali-G610 which has been popular for providing sustained gaming performance in premium devices. This year the Arm Mali-G615 brings premium mobile use cases, capabilities and features to a wider audience of developers and consumers sooner.

The latest Arm GPUs are the most performant to date, with a 15% performance improvement compared to the previous generation, demonstrating Arm's focus on ensuring that the latest flagship and premium smartphones deliver the AAA gaming experiences that end-users are demanding. The new GPUs also build upon the highly efficient Arm Mali-G710 GPU, with 15% energy efficiency improvements to deliver more game time 'on-the-go'.



For more technical details on Immortalis-G715 and the new Mali GPUs, visit our blog.

Armv9 CPUs: premium sustained performance and efficiency for next-generation devices
2021 saw the first of the Armv9 CPUs and this year we are meeting the demand for performance and efficiency with a new expanded CPU cluster. The new CPUs are all designed to push peak performance and deliver exceptional sustained performance and efficiency.

We are continuing to increase single-threaded performance with the new Arm Cortex-X3 which targets a range of benchmarks and applications, delivering a 25% performance improvement compared to the latest Android flagship smartphone and a 34% performance improvement compared to the latest mainstream laptops.

Arm Cortex-A715 focuses on efficient performance, delivering a 20% energy efficiency gain and 5% performance uplift compared to Cortex-A710, reaching the significant milestone of matching the performance of Cortex-X1.

The importance of efficient performance makes Cortex-A715 the CPU cluster workhorse of "big.LITTLE"-based configurations, with the technology now the most commonly used heterogeneous processing architecture for consumer devices worldwide.

Last year we launched the Armv9 based Arm Cortex-A510, our first high efficiency "LITTLE" core in four years, and this year's updated version maintains performance while delivering a 5% power reduction. Meanwhile, the newly updated DSU-110 supports 50% more cores in CPU clusters compared to last year's generation, enabling scalability across different tiers of consumer devices.



For more technical details on the new Armv9 CPU cores, visit our blog.

Mobile gaming runs on Arm
Compute performance has allowed us to develop ways to connect more authentically, closing the gap created by distance. As technology is evolving, visual experiences allow us to feel more present in a shared moment, or feel the transcendence of entering another world. TCS22 is shaping the visual experiences of tomorrow. We are engaging more people than ever in exploring 3D worlds and augmenting the world around us with virtual visual content. Mobile gaming is driving the evolution of real-time 3D technology and the growth of visual experiences will see it deployed even more widely in the future.

Developing on Arm allows creators to target energy-efficient high-performance computing solutions. That means longer play time for the next wave of innovative applications and visual experiences. I can't wait to see what the developers of tomorrow will create. The best visual experiences live forever and I'm confident they will be on Arm.

View at TechPowerUp Main Site | Source

 

[defaultluser]

if they're "leading in efficiency", then why did Google tensor still get owned by the corresponding Qualcomm in everey other device??

 

[mplayerMuPDF]

if they're "leading in efficiency", then why did Google tensor still get owned by the corresponding Qualcomm in everey other device??

there are many factors involved, including the manufacturing node and the core strategy/layout and then there are also DSPs/specialized ASICs as well as software (support/optimization) that have to be taken into account. It seems to me that so many people who are talking (often shit) about smartphone SoCs don't have a clue what they are talking about really. I would not be surprised if that also correlates with fanboyism and buying a new phone every year for no reason.

 

[defaultluser]

there are many factors involved, including the manufacturing node and the core strategy/layout and then there are also DSPs/specialized ASICs as well as software (support/optimization) that have to be taken into account. It seems to me that so many people who are talking (often shit) about smartphone SoCs don't have a clue what they are talking about really. I would not be surprised if that also correlates with fanboyism and buying a new phone every year for no reason.

Google's Tensor inside of Pixel 6, Pixel 6 Pro: A Look into Performance & Efficiency

www.anandtech.com

The exact same foundry process and node as Qualcomm was forced to use, because they couldn't book TSMC until this year?

While the CPU performance is within expected performance for having different core lineups, the beefcake Mali x20 Still wasn't enough to overtake Adreno. on the exact same Samsung process node!

 

[mplayerMuPDF]

Google's Tensor inside of Pixel 6, Pixel 6 Pro: A Look into Performance & Efficiency

www.anandtech.com

The exact same foundry process and node as Qualcomm was forced to use, because they couldn't book TSMC until this year?

While the CPU performance is within expected performance for having different core lineups, the beefcake Mali x20 Still wasn't enough to overtake Adreno. on the exact same Samsung process node!

And as your article points out, they are still using "old" A76s, which means that it is already going to perform worse than a hypothetical ARM reference design. You can't blame ARM here. Personally I don't care about getting the latest and greatest (especially in a phone) but if you are going to study benchmarks so you can hate one SoC/core/manufacturer/designer/whatever for being "bad", then you at least need to get your facts and analysis straight.

 

[Minus Infinity]

Hopefully used in the Samsung Galaxy Tab S9 models next year. S9+ would be a nice alternative to iPad Pro.

 

[Crackong]

I am a little bit confused.
Since ARM is launching the new CPU core 'design' instead of the actual physical CPU core itself.
Where are those performance numbers came from?
Did they manufactured an actual CPU form the new core designs and tested them?
Or did they ran the tests on a virtualized system with a virtualized CPU with virtualized 'New CPU cores' ?

If they manufactured an actual CPU to test them, how to account the performance differences between different manufacture processes from various fabs ?

 

[R0H1T]

I am a little bit confused.
Since ARM is launching the new CPU core 'design' instead of the actual physical CPU core itself.
Where are those performance numbers came from?
Did they manufactured an actual CPU form the new core designs and tested them?
Or did they ran the tests on a virtualized system with a virtualized CPU with virtualized 'New CPU cores' ?

If they manufactured an actual CPU to test them, how to account the performance differences between different manufacture processes from various fabs ?

Yes from what I remember ARM tests these cores with TSMC.

Arm and TSMC Demonstrate Industry’s First 7nm Arm-based CoWoS® Chiplets for High-Performance Computing

Arm and TSMC today announced an industry-first 7nm silicon-proven chiplet system based on multiple Arm® cores and leveraging TSMC’s Chip-on-Wafer-on-Substrate (CoWoS®) advanced packaging solution.

www.arm.com

So their claims aren't totally bogus, though you might not get a (physical) chip for everything they've announced today.

 

[Hofnaerrchen]

I am a little bit confused.
Since ARM is launching the new CPU core 'design' instead of the actual physical CPU core itself.
Where are those performance numbers came from?
Did they manufactured an actual CPU form the new core designs and tested them?
Or did they ran the tests on a virtualized system with a virtualized CPU with virtualized 'New CPU cores' ?

If they manufactured an actual CPU to test them, how to account the performance differences between different manufacture processes from various fabs ?

These numbers are based on simulated CPUs. After years of creating CPUs they probably refined the algorithms to do so. It's done everywhere this way these days. Why do you think everybody needs higher performance chips? It's primarily for science.... consumer products are a byproduct - even if you can make a fortune from them.

 

[R0H1T]

ARM does produce some reference design (chips) but again not for everything all the time.

 

[jonup]

if they're "leading in efficiency", then why did Google tensor still get owned by the corresponding Qualcomm in everey other device??

Because google built tensor cores on 4 gen older cortex cores than the cortex cores Qualcomm builds the snapdragon on. I hope you understand they use the same Arm's architectures. Just the tensor cores are built on the same Arm's architecture as SD855 or SD865. Tensor cores are so irrelavant that I can remmember which obsolete SD are equivelent to them.