Cortex A78 Chipset Review, Features & Performance Benchmark

Cortex A78 Chipset Review, Features & Performance Benchmark


Cortex A78 Part 1

#smartphones    #Samsung    #flagship   

Arm replaces CPU and GPU flagships, and more : Arm has announced flagship processors for phones : Cortex – X1 CPU, Cortex – A78 CPU, Mali – G78 GPU and Ethos – N78 neural network processor ( NPU ).

With a 30 % peak performance increase over Cortex – A77, it offers an even more competitive solution for flagship smartphones and large – screen devices.

The future for CPU performance in Android looks bright even as one major CPU chip producer stands on the brink of closure.

Today, ARM detailed its new high – end CPU core designs that will be used by chipset vendors in the upcoming smartphone flagships.

Today, ARM detailed its new high – end CPU core designs that will be utilized by chipset distributors in the upcoming smartphone flagships.

#program    #chips    #introduced   

Also, for the first time, they ‘ve introduced extra – performance cores i.e. Cortex – X1 core developed out of its new Cortex – X Custom Program.

Since the program is brand new and custom chips are yet to be made, Arm presented its own Cortex – X1 core.

As part of its annual IP mobile platform announcement, Arm introduced a new initiative called Cortex – X Custom ( CXC ) program and the first product from the program, the Cortex – X1 CPU cores.

The company has revealed their next processor cores, the Cortex A78 an expected upgrade to the current A77 and the Cortex X1, a brand – new core altogether.

ARM says that it will supply Cortex X1 cores only to those partners who have been a part of the CXC program from the beginning and that they would all get the same configuration of the chip.

#consuming    #avoiding   

So the real benefit of Cortex – A78 is higher efficiency which should lead to either a more constant performance or longer battery life with Cortex – A78 consuming 50 % less than Cortex – A77 at the same performance.

It is designed for high – end performance at optimal efficiency and should deliver 20 percent greater sustained performance than the Cortex – A77.

This avoids power throttling for applications demanding a lot of performance, which improves the user experience by avoiding lag or frame – rate drops.

Instead of looking at the Cortex A78 in terms of its peak performance output, consider the fact that a Cortex A78 core running at 2.1 GHz consumes up to 50 percent less power than a Cortex A77 core running at 2.3GHz.

#improvement    #machine   

The ARM Cortex – A78 CPU is claimed to bring a 20 % sustained performance improvement over the previous – gen ARM Cortex – A77 CPU.

It brings 30 percent peak performance improvements over Arm Cortex – A77 CPU and 22 % over the just – announced Cortex – A78 core.

ARM says the Cortex – X1 CPU will offer a performance gain of 30 % over the Cortex – A77 design and a 2x improvement at ML performance among other benefits.

The new Cortex – A78 CPU provides a 20 % sustained performance improvement over Arm Cortex – A77 CPU in the same 1 Watt power envelope and more efficient management of computing workloads along with greater on – device Machine Learning ( ML ) performance.

#Mali    #GPU    #graphics    #Valhall   

Like the Mali – G77 GPU, Mali – G78 has Arm’s Valhall architecture, but is said to deliver 25 % more graphics performance.

It is accompanied by the high – end Mali – G78 GPU and the slightly less powerful Mali – G68 GPU.

Designed to be paired with the Cortex – A78, the new Mali – G78 GPU sticks with the Valhall architecture of last year’s Mali – G77 but promises a 25-percent increase in graphics performance.

The company claims the GPU will deliver a 25 % increase in graphics performance compared to the current Mali – G77.

ARM a egalement devoile un GPU milieu de gamme, le Mali – G68, une variante du Mali – G78 plus abordable.

The Mali – G78 supports up to 24 cores and brings along 25 % performance gains over the previous – gen Mali – G77 GPU.

#displays   

As an example, this could mean lowering power efficiency beyond what Arm allows for the normal Cortex – A78 in exchange for class – leading performance on flagships or phones with large displays.

The next generation of mobile SoCs, at least the ones using Cortex A78 cores, seem to promise slightly improved performance, but significantly better power efficiency.

However for devices which need maximum performance, whether for driving sizable displays or something else, it could well fit the bill.

Alternatively, a phone maker could clock the A78 to consume half the power at the same performance as the A77, Williamson said.

Smartphone makers have a choice between using the new ARM cores to maximize performance, or to deliver better battery life with the same performance as the prior generation.

#software    #company    #Synopsys    #semiconductor   

ARM, the Softbank – owned semiconductor and software design company, has taken the wraps off its latest CPU and GPU stack.

As the world’s 15th largest software company, Synopsys has a long history of being a global leader in electronic design automation ( EDA ) and semiconductor IP and is also growing its leadership in software security and quality solutions.

Inc. ( Nasdaq : SNPS ) is the Silicon to Software(tm ) partner for innovative companies developing the electronic products and software applications we rely on every day.

Today, the company unveiled the Cortex A78 processor cores, the Mali G78 graphics processor, and the Cortex X chip, all of which have the ability to fine – tune the manufacturers of turnkey solutions based on them.

#node    #nm    #process   

Note that these performance increases also include the benefit of the 5 nm process node, over the 7 nm process node used on the current Cortex – A77-based SoCs.

These Cortex – A78 cores will be made out of 5 nm fabrication nodes, enabling greater performance with lower power requirements.

The Cortex A78 cores move to a 5 nm process node which helps improve not only the efficiency numbers but also those of performance and even the area taken up by the core within the processor die.

ARM said that a Cortex – A78 core running at 3GHz would deliver 20 percent more sustained, single – core performance than the Cortex – A77 core running at 2.6GHz, assuming 1 watt per core.

#Apple    #chips    #series    #Qualcomm   

Qualcomm’s Snapdragon chips and Apple’s A – series chips have different structures and performance but they’re all based on Arm’s designs.

The new Cortex – X1 is a direct response to the further customized and optimized chips that you find in Apple’s A – series designs and chips like Qualcomm’s Snapdragon 8xx series.

Apple’s A – series chips are big and wide, as the decode width of all A – series chips since the A11 has been 7-wide, which is wider than even desktop CPU architectures.

For the first time since 2013 and the Apple A7, ARM will be able to get close to Apple’s A – series chips in terms of peak performance.

The new Cortex – X1 chip takes a similar approach to custom ARM – based chips from Apple and Qualcomm in that it seeks to optimize the architecture further for high – performance.

#Ethos    #s   

The new NPU also delivers up to twice the peak performance of Ethos N77, 25 % better performance efficiency, and 40 % greater DRAM bandwidth efficiency.

The company is promising that the Cortex – A78 will deliver 20 percent greater sustained performance over the prior generation; the Mali – G78, 25 percent better overall performance; and the Ethos – N78, 25 percent more performance efficiency.

Finally, there’s the new Ethos – N78 neural processing unit ( NPU ), which provides up to 30 % higher area efficiency and 2x peak performance compared to the first – gen Ethos – N77 NPU.

The Ethos – N78 NPU supports up to 90 different configurations with performance ranging from 1 TOPS to 10 TOPS, and customizable area ( inferences / s / mm2 ), throughput ( inferences / s ) and average DRAM bandwidth ( GB / s ).

#compared    #year    #design   

Compared to last year’s Cortex – A77 design, it features a 20 per cent increase in sustained performance, while remaining within a 1-watt power budget.

It features a 20 percent increase in sustained performance, quot; compared to last year’s Cortex – A77 design, while staying within a 1 watt power budget.

ARM says that this features a 20 % increase in sustained performance and that is compared to the Cortex – A77 design it released last year.

The Cortex – A78’s marquee achievement is a 20 % boost to sustained performance, while being 15 % smaller compared to last year’s A-77 design that was used in chips like Qualcomm’s top – of – the – line Snapdragon 865.

For the same levels of performance ( on SpecInt 2006 ), moving from the A77 to the A78 yield fairly significant improvement in power.

#custom    #program    #ARM    #X   

The Cortex – X1 was designed within the frame of a new program at Arm, which the company calls the Cortex – X Custom Program .

ARM is also introducing a new customized Cortex – X program for its partners to create their own specialized Cortex CPU together with ARM for their specific purposes beyond the Cortex – A78 stock.

ARM has also introduced a new Cortex – X custom program, which will allow partners to shape a Cortex CPU product for specific use cases.

As per ARM, the Cortex – X1 CPU marks the beginning of a new Cortex – X Custom program that will allow partner brands to work with ARM’s team for designing custom CPUs based on their demand.

#enabled    #energy   

That kind of performance boost was enabled because the X1 is both bigger and more power – demanding than the A78.

Again, extra performance hasn’t meant a raid on energy usage, and the new part is 10 per cent more energy efficient.

With performance approaching desktop designs, but with only a fraction of the power required, Arm is enabling new levels of compute performance and interactivity that was once impossible half a decade ago.

It supports up to 24 cores and is said to have 10 % better energy efficiency to enable longer battery life.

Since the previous gen, you have decent performance increases alongside, Arm claims, a smaller and lower power consumption CPU core, too.

Last year the launch of the Cortex – A77 CPU was trumpeted as offering up to 35 per cent more performance over the previous generation, so the bar is being raised without sacrificing power.

#peak    #decode    #offers   

ARM says the Cortex – X1 offers 30 percent more peak single – core performance than the previous Cortex – A generation — a bit more than the 20-percent improvement offered by the Cortex – A78 in general.

The Cortex – X1 has a 30 % increased peak efficiency than the Cortex – A77 and is 23 % sooner than the new Cortex – A78 at integer calculations.

The Cortex – X1 has a 30 % higher peak performance than the Cortex – A77 and is 23 % faster than the new Cortex – A78 at integer calculations.

The DynamIQ cluster of 4x Cortex – A78 and 4x Cortex – A55 cores provides 20 % sustained performance improvements over the 4x Cortex – A77 and 4x Cortex – A55 cluster.

#years    #Austin    #family   

Based on the same Austin – family CPU architecture that was first introduced in the Cortex – A76 three years ago.

The Cortex – A78, codenamed Hercules, is the latest iteration of its Austin – family CPU architecture that was first introduced in the Cortex – A76 three years ago.

Two years later, and here we are, with the Cortex – A78 representing the third iteration of Arm’s new Austin – family CPU microarchitecture, which had started from scratch with the Cortex – A76.

In a sense, we sort of saw the start of this last year with the introduction of the Neoverse – N1 CPU, already having some more notable microarchitectural changes over its Cortex – A76 mobile sibling.

#customers    #product    #company    #design    #targeted   

While the first product in the CXC program is targeted toward multiple unnamed customers, future products could be targeted toward individual customers.

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.

Here, Arm proclaims that it allows customers to customize and differentiate their products more; but the real gist of it is that the company now has the resources to finally do what some of its lead customers have been requesting for years.

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.

#terms    #area    #balance    #Arm   

At the end of the day, making everything bigger doesn’t necessarily strike the right balance in terms of performance, power, and area.

The A78, in tandem with some of the other changes, was said to have shrunk slightly, striking a better power / area balance.

With the A78, the Austin team revisited and scrutinized many of the buffers that grew in terms of power and area ROI.

Every once in a while following a number of generations, designers have to look back at some of the long – term modifications that have been made and scrutinize them in terms of the performance benefits that they delivered versus the area and power costs.

To that end, as you’ll see, the Arm team revisited the various components of the microprocessor, shrinking some buffers at times and enlarging others, in order to yield a better balance of performance and power.

#percent    #ISO    #FinFet    #nm   

When compared with Cortex A77 at 2.6GHz 7 nm FinFet, the Cortex A78 at 3GHz and 5 nm FinFet and based 1W per core comes about 20 percent faster.

Cortex A78 gives you seven percent more performance compared to Cortex A77 in ISO comparison, with four percent less power and five percent smaller area.

FinFet has 50 percent lower energy consumption and the same performance when compared to the Cortex A77 clocked at 2.3GHz in 7 nm FinFet.

In an ISO – process comparison, we can see that the Cortex – A78 is roughly 5 % smaller than the Cortex – A77.

In terms of ISO – process and ISO – frequency, Arm is reporting a modest 7 % improvement on SPECint 2006.

#cache    #MOP    #MB   

It can have up to 1 MB of L2 cache and up to 8 MB of L3 cache – double what the new Cortex – A78 supports.

In terms of cache sizes, the X1 supports 64kB L1 and up to 1 MB L2 cache, while the DynamIQ cluster has been upgraded to now support 8 MB of L3 for ultimate performance.

On top of this, ARM’s new design has also boosted its MOP cache throughput by 33 %, delivering 8Mops per cycle.

Up to 6 MOPs may be fetched from the pre – decoded MOP cache or up to 4 instructions may be fetched from the L1 cache and be decoded each cycle.

The X1 doubles SIMD throughput, can dispatch 5 instructions or 8 Mops per cycle, and offers up to 1 MB of L2 and 8 MB of L3.

Cortex A78 Cpu Part 2

Samsung Exynos 990 uses Cortex – A76 and Cortex – A55 CPU cores as well as the Mali – G77 GPU.

Now for this year, ARM has announced the ARM Cortex – A78 CPU architecture and the Mali – G78 GPU, the successors of the Cortex – A77 CPU, and the Mali – G77 GPU respectively.

The report states that the Exynos 992 could use ARM’s new Cortex – A78 CPU and Mali – G78 GPU architectures for improved performance.

Interestingly, the report states the Exynos 992 will use the all – new ARM Cortex – A78 CPU and Mali – G78 GPU designs.

Qualcomm’s Snapdragon 865, for example, uses partially custom versions of the ARM Cortex – A77 and Cortex – A55 designs for its Kryo 585 CPU, while Samsung’s Exynos 990 uses Cortex – A76 and Cortex – A55 CPU cores, along with a GPU Mali – G77.

#percent    #compared   

Cortex A78 gives you seven percent more performance compared to Cortex A77 in ISO comparison, with four percent less power and five percent smaller area.

The company is promising that the Cortex – A78 will deliver 20 percent greater sustained performance over the prior generation; the Mali – G78, 25 percent better overall performance; and the Ethos – N78, 25 percent more performance efficiency.

When compared to the new Cortex – A78, the ARM Cortex – X1 offers 22 % higher integer ( single – thread ) performance.

Compared to the Cortex – A77, the Cortex – X1 offers up to 30 percent higher integer performance and up to 100 percent higher performance in ML ( machine learning ) and AI ( artificial intelligence ) tasks.

#power    #designs   

These CPU, GPU and NPU designs respectively offer the highly desirable duo of upgrades in increased performance for the same or lower power draw. 6

ARM is promising up to a 25-percent performance boost in its new flagship – class Mali – G78 GPU, while using less power than its predecessors.

With performance approaching desktop designs, but with only a fraction of the power required, Arm is enabling new levels of compute performance and interactivity that was once impossible half a decade ago.

ARM is introducing a series of new chip designs for the coming year that the company says will bring stronger CPU, graphics, and AI performance and reduced power consumption.

ARM wants its core designs to be used across a wider variety of applications, and that has forced ARM to invest in new core IP and designs which are less restricted when it comes to power and die area.

#graphics   

Essentially, ARM is naming the GPU as Mali – G68 if it has six or lower cores, and as Mali – G78 if it has seven or more GPU cores.

The Mali – G78 supports up to 24 cores and brings along 25 % performance gains over the previous – gen Mali – G77 GPU.

Like the Mali – G77 GPU, Mali – G78 has Arm’s Valhall architecture, but is said to deliver 25 % more graphics performance.

The Mali – G78, which will deliver a 25 % increase in graphics performance relative to Mali – G77, brings support for up to 24 cores.

The company claims the GPU will deliver a 25 % increase in graphics performance compared to the current Mali – G77.

#Asynchronous   

And as usual, these high – performance CPU cores can be combined with lower – performance, more energy – efficient CPU cores in the same chip.

Going by ARM’s performance claims, processors that use Cortex – X1 CPU cores would be able to offer performance that’s in the same ballpark as that of Apple’s A13 Bionic SoC.

Moving on to the Cortex – X1, which is a new custom line of CPU cores that are all about performance.

The A78 may also be mixed with a high – performance Cortex – X1 core in order to provide certain workloads with an additional boost in single – core performance.

Additionally, one or more of the A78 cores may be swapped out for a Cortex – X1 core in order to achieve even higher performance.

#chips   

The future for CPU performance in Android looks bright even as one major CPU chip producer stands on the brink of closure.

Qualcomm’s Snapdragon chips and Apple’s A – series chips have different structures and performance but they’re all based on Arm’s designs.

For years, tech reviewers and users have bemoaned the fact that Apple’s CPU architectures are multiple years ahead of ARM’s Cortex – A series.

In terms of CPU core architectures, from 2021, every major mobile chip vendor of note will use ARM’s stock CPU IP ( as Samsung System LSI has given up on its Exynos M custom cores ).

ARM announced the new Cortex – A78, which is a CPU core that is expected to be used by virtually of the big chip makers, which means that we will see it powering Android phones in 2021.

#Qualcomm    #smartphones    #Apple   

The Cortex A7x cores just could not match Apple’s engineering, and now, ARM’s answer to that is the Cortex – X1 core.

The upcoming Qualcomm Snapdragon 875 will probably incorporate both the Cortex – X1 as well as the Cortex – A78 as part of its Prime Core and Performance Cores .

We can definitely expect Qualcomm to further tweak the performance of the Cortex A78 cores, but it would be far more exciting to see products that make use of ARM’s Cortex – X1 core.

We expect to see Cortex – A78 CPU cores in a lot of mid – range and high – end smartphones over the next few years.

The Cortex – A78 cores will be coupled with the high – performance Cortex – X1 which will be responsible for driving heavy tasks like gaming or other high – performance applications, a similar approach to the custom ARM – based chips developed by Apple and Qualcomm.

#program    #Custom    #partners   

This week, the brand debuts three major products : the Arm Cortex – X Custom Program, Arm Cortex – A78 CPU, and the Arm Mali – G78 GPU.

ARM has also introduced a new Cortex – X custom program, which will allow partners to shape a Cortex CPU product for specific use cases.

ARM is also introducing a new custom Cortex – X program for its partners to create their own specialized Cortex CPU in conjunction with ARM for their specific purposes beyond the Cortex – A78 stock.

Cortex – X1 is a result of the CXC program in which ARM makes custom CPU cores in collaboration with its partners.

The Cortex – X Custom Program allows for customization and differentiation beyond the traditional roadmap of Arm Cortex products, enabling our partners with a solution for providing the ultimate performance for specific use cases.

#software    #company    #processor    #design   

In the smartphone industry, ARM designs its Cortex CPUs, Mali GPUs, and Ethos machine – learning processors, then licenses them to companies like Qualcomm.

ARM, the Softbank – owned semiconductor and software design company, has taken the wraps off its latest CPU and GPU stack.

Today, the company unveiled the Cortex A78 processor cores, the Mali G78 graphics processor, and the Cortex X chip, all of which have the ability to fine – tune the manufacturers of turnkey solutions based on them.

Inc. ( Nasdaq : SNPS ) is the Silicon to Software(tm ) partner for innovative companies developing the electronic products and software applications we rely on every day.

As the world’s 15th largest software company, Synopsys has a long history of being a global leader in electronic design automation ( EDA ) and semiconductor IP and is also growing its leadership in software security and quality solutions.

#Exynos    #Samsung   

Notably, Samsung is in a strong position to adopt the Cortex – X1 + Cortex – A78 as part of the next flagship Exynos SoC, which will succeed the Exynos 990.

Interestingly, both Exynos by Samsung and Kirin by Huawei ended up using Cortex A76 for the 2020 phones, probably due to its time to market plans.

Cortex X1 is the first CPU in this new lineup and we expect it to replace Samsung’s own Mongoose cores in next year’s Exynos chipsets.

Most of the current flagship processors such as Snapdragon 865 and Samsung’s Exynos 990 are based on the 7 nm process.

Lately, Samsung has been facing a lot of criticism for the performance difference between the two variants of its flagship device.

#cache    #instruction   

Cortex – X1 has four 128-bit NEON units, 64kB L1 cache, up to 1 MB L2 cache, and up to 8 MB L3 cache.

The A78 TLB consists of dedicated L1 TLB for instruction cache ( ITLB ) and another one for data cache ( DTLB ).

For execution efficiency, Arm also doubled the available L1 data cache, the L2 cache, and L3 cache while increasing the L2 Translation Lookaside Buffer ( TLB ) by 66 % to 2k entries.

It can have up to 1 MB of L2 cache and up to 8 MB of L3 cache – double what the new Cortex – A78 supports.

To feed the beast, Arm increased the Branch Target Buffer ( BTB ) by 50 % up to 96 entries, increased the instruction catch to handle 5 instructions and the Macro – op ( MOP ) cache to 8 MOP, and doubled the MOP cache capacity to 3k entries.

 

Cortex A78 Features Part  3

#CPU    #Mali    #GPU   

It may also feature two Cortex – A78 CPU cores, two Cortex – A76 CPU cores, and four Cortex – A55 CPU cores alongside ARM’s unannounced Mali MP20 GPU.

Samsung Exynos 990 uses Cortex – A76 and Cortex – A55 CPU cores as well as the Mali – G77 GPU.

The Exynos octa – core processor reportedly features two Cortex – A78 CPU cores, two Cortex – A76 CPU cores, and four Cortex – A55 CPU cores.

Now for this year, ARM has announced the ARM Cortex – A78 CPU architecture and the Mali – G78 GPU, the successors of the Cortex – A77 CPU, and the Mali – G77 GPU respectively.

Interestingly, the report states the Exynos 992 will use the all – new ARM Cortex – A78 CPU and Mali – G78 GPU designs. 0

#Exynos    #feature    #tipster    #shutting   

According to the tipster, the Exynos 1000 may feature an octa – core configuration, where 4 cores will be performance – based, while the staying cores will be focused around efficiency.

According to the tipster, the Exynos 1000 might feature an octa – core setup, where 4 cores will be performance – based, while the remaining cores will be focused around efficiency.

Because shutting down its customized core center means the Exynos 1000 will not rely on customized high – performance Mongoose cores, the upcoming chipset might take advantage of ARM’s Cortex – A78 cores.

Since shutting down its custom – made core facility indicates the Exynos 1000 will not rely on custom – made high – performance Mongoose cores, the upcoming chipset could take benefit of ARM’s Cortex – A78 cores.

#percent    #compared    #CPU   

Cortex A78 gives you seven percent more performance compared to Cortex A77 in ISO comparison, with four percent less power and five percent smaller area.

It brings 30 percent peak performance improvements over Arm Cortex – A77 CPU and 22 % over the just – announced Cortex – A78 core.

The first CPU designed as part of the Cortex – X program is the Cortex – X1 CPU, which brings a 30 % increase in peak performance over the current Cortex – A77 CPU.

The ARM Cortex – A78 CPU is claimed to bring a 20 % sustained performance improvement over the previous – gen ARM Cortex – A77 CPU.

When compared to the new Cortex – A78, the ARM Cortex – X1 offers 22 % higher integer ( single – thread ) performance.

#displays    #run    #phones   

These benefits result in better performance per watt, making the design of the A78 well suited for both power- and performance – hungry 5 G phones, as well as foldable and other devices featuring larger displays.

Smartphone makers have a choice between using the new ARM cores to maximize performance, or to deliver better battery life with the same performance as the prior generation.

Most interesting is how Arm can run the core at different speeds, depending upon where the bottlenecks may be, as any processor is only as fast as its slowest – running component.

As an example, this could mean lowering power efficiency beyond what Arm allows for the normal Cortex – A78 in exchange for class – leading performance on flagships or phones with large displays.

#Mali    #GPU    #graphics    #supports   

It is accompanied by the high – end Mali – G78 GPU and the slightly less powerful Mali – G68 GPU.

Like the Mali – G77 GPU, Mali – G78 has Arm’s Valhall architecture, but is said to deliver 25 % more graphics performance.

The Mali – G78, which will deliver a 25 % increase in graphics performance relative to Mali – G77, brings support for up to 24 cores.

The company claims the GPU will deliver a 25 % increase in graphics performance compared to the current Mali – G77.

Because the Mali – G78 supports up to 24 cores, on paper, it offers 25 % better graphics performance compared to the Mali – G77 that’s in the Exynos 990.

#Custom    #program    #X   

As per ARM, the Cortex – X1 CPU marks the beginning of a new Cortex – X Custom program that will allow partner brands to work with ARM’s team for designing custom CPUs based on their demand.

ARM has also introduced a new Cortex – X custom program, which will allow partners to shape a Cortex CPU product for specific use cases.

ARM is also introducing a new custom Cortex – X program for its partners to create their own specialized Cortex CPU in conjunction with ARM for their specific purposes beyond the Cortex – A78 stock.

Cortex – X1 is a result of the CXC program in which ARM makes custom CPU cores in collaboration with its partners.

Arm has created a new Cortex – X Custom ( CXC ) program to allow CPU manufacturers closer cooperation with Arm’s engineers to deliver chips tailored to the needs of the customers.

#Ethos    #efficiency   

The new NPU also delivers up to twice the peak performance of Ethos N77, 25 % better performance efficiency, and 40 % greater DRAM bandwidth efficiency.

The company is promising that the Cortex – A78 will deliver 20 percent greater sustained performance over the prior generation; the Mali – G78, 25 percent better overall performance; and the Ethos – N78, 25 percent more performance efficiency.

Ethos – N78 is a neural networking processor that compared with the Ethos – N77 delivers greater on – device ML capabilities and up to 25 % more performance efficiency.

It has the same architecture as Cortex – A77, with modified micro – architecture to increase performance / W and performance / area.

ARM also has a dedicated machine – learning core, the Ethos – N78, which it’s optimized for more efficient data movement.

#chips    #designs    #year    #company   

However, when ARM announces its next – generation cores each May, as it typically does, the chips featuring the latest ARM designs don’t usually arrive in shipping devices until the following year.

We won’t bore you with the technical details, just wanted to point out how important it is when Arm, the company that owns the IP behind the ARM architecture, releases new chip designs.

But it’s unclear how much the company gains by developing custom CPU cores rather than licensing designs from ARM — so it’ll be interesting to see how the company’s next – gen chips stack up against the competition in the coming years.

It is shaping up to be a promising year for System on Chip technologies, with the 5 nm process becoming commercially available by the end of the year.

#Samsung    #smartphone    #Exynos   

Samsung tends to use those custom CPU cores for its flagship smartphone chips — like the Exynos 9820 that powers Galaxy S10 smartphone models sold outside of the United States.

Lately, there has been a lot of discussion around Samsung’s Exynos chipsets and how they perform worse than their Snapdragon counterparts.

Samsung’s next – gen flagship Exynos processor could perform better as the company is switching from Mongoose CPU cores to stock ARM CPU cores.

Arm’s previous – gen Cortex – A77 design is an established staple of high – end flagship smartphones, featuring in the likes of the Qualcomm Snapdragon 865, MediaTek Dimensity 1000 5 G SoC, and Samsung Exynos 980 5 G processors.

Interestingly, both Exynos by Samsung and Kirin by Huawei ended up using Cortex A76 for the 2020 phones, probably due to its time to market plans.

#improvements    #branch    #accuracy    #prediction    #Timeliness   

Chris Abernathy, distinguished engineer, Client Line of Business mentioned improvements in branch prediction, better accuracy, improved better target process as well as many microarchitecture improvements.

Arm claims their microarchitecture is very branch prediction driven so the improvements here add a lot to the generational improvements of the core.

Other architectural tweaks include; front – end improvements in branch prediction and cache, out – of – order core improvements to efficiency, and back – end improvements – especially with a view to bandwidth. 21

Naturally, the branch predictors themselves have also been improved in terms of their accuracy, which is an ongoing effort with every new generation.

Arm seemingly already has some of the most advanced prefetchers in the industry, and here they claim the A78 further improves the designs both in terms of their memory area coverage, accuracy and timeliness.

#cache    #cycle    #bandwidth   

Cortex – X1 has four 128-bit NEON units, 64kB L1 cache, up to 1 MB L2 cache, and up to 8 MB L3 cache.

On Cortex – X1, the Neon engine gets two additional pipes, doubling its compute capacity over Cortex – A78, then Cortex – X1 supports 64kbyte L1 and up to 1Mbyte L2 cache.

It can have up to 1 MB of L2 cache and up to 8 MB of L3 cache – double what the new Cortex – A78 supports.

Cache bandwidth is substantially higher, with doubled available bandwidth to both L1 and L2, as well as the already – mentioned doubling of L2 capacity.

The interface bandwidth from the LD / ST queues to the L1D cache has been doubled from 16 bytes per cycle to 32 bytes per cycle, and the core’s interfaces to the L2 has also been doubled up in terms of both its read and write bandwidth.

 

 

Amanda Fox

Add comment

Your Header Sidebar area is currently empty. Hurry up and add some widgets.