A Cortex-X CPU core is added to Qualcomm’s Snapdragon 7+ Gen 2’s premium segment SoC

A Cortex-X CPU core is added to Qualcomm’s Snapdragon 7+ Gen 2’s premium segment SoC

Despite a slightly more eventful 2021–2022 product cycle than Qualcomm may have preferred, 2023 has been a much simpler year for the well-known SoC and cellular modem maker. The company is getting ready to iterate through the next stage of its product stack with the Snapdragon 7+ Gen 2 after introducing the first of their Gen 2 family of parts earlier this year with the flagship-class Qualcomm 8 Gen 2. For the Snapdragon 7+ Gen 2, Qualcomm is trying to give a significant performance bump to the platform with an eye toward what has grown to be Qualcomm’s regular $400 to $600 “premium” market segment, which focuses on flagship-level features with more modest performance and pricing.

This year’s generation of the Snapdragon 7, which is positioned as the successor to last year’s Snapdragon 7 Gen 1, is, generally speaking, more concerned with increasing performance than with introducing new capabilities. There aren’t many new features this year compared to last year’s Gen 1 component, which brought mmWave support and new CPU and GPU architectures, particularly Armv9 architecture CPU cores. Instead of that, the Snapdragon 7 family has received what Qualcomm claims to be one of the largest performance gains ever. This is largely made possible by TSMC’s 4nm process, which is replacing Samsung’s troubled 4nm process. This is similar to the swap Qualcomm made last year for the well-received mid-cycle Snapdragon 8+ Gen 1 component.

Regarding their choice to introduce their initial Gen 2 component as the 7+ rather than the 7, Qualcomm is hinting that this won’t be the first Snapdragon 7 Gen 2 part we see this year. In other words, starting as a Snapdragon 7+ part gives Qualcomm room to introduce a standard Snapdragon 7 part in the future. There’s no reason for Qualcomm to offer a 7+ first unless they had intentions for a part underneath it; otherwise, they could have debuted it as a 7 part similar to the Snapdragon 7 Gen 1, which was always a single chip stack.

The Snapdragon 7+ Gen 2 maintains the same 1+3+4 CPU core configuration as previous generations of the Qualcomm 7 family in terms of CPU organization. The main news here is that Qualcomm switches from employing a slightly faster mid-core to using a more potent CPU design entirely, which results in a considerable performance gain for the top-performing Prime core.

Thus, Qualcomm is using one of Arm’s Cortex-X cores for the Prime core in a Snapdragon 7 component for the first time ever. Although the Cortex-X2 employed here is actually an older Arm design, the Snapdragon 8 Gen 2 and its Cortex-X3 core won’t be infringed upon. Nevertheless the Cortex-X2 represents a huge gain in IPC and clockspeeds over the A710 core utilized for the 7 Gen 1 Prime core (and 7+ Gen 2’s mid-cores). With the IPC advantages of the more sophisticated core, the Prime core’s peak clockspeed has increased from 2.4GHz to 2.91GHz as a result. Overall, Qualcomm claims that the new Prime core would enhance CPU performance for the 7+ Gen 2 by “up to” 50% compared to the 7 Gen 1.

As there is only one Cortex-X2 core, the trade-off is that such a significant performance boost is really only available for single-threaded workloads. The three mid-performance cores are still based on the Cortex-A710 and have a 2% increase in clock speed. As a result, the 7+ Gen 1 won’t significantly improve on workloads with a lot of threads. There should be some benefits from the increased power efficiency of TSMC’s 4nm process, but part of those gains have gone toward making the power-hungry Cortex-X2 practical in terms of battery life.

A faster Adreno GPU is also included in the 7+ Gen 2. There isn’t much information we can give because Qualcomm isn’t giving it a product number or revealing important architectural specifics, as has been the case with its integrated GPUs for a few generations now. Based on the feature overview, it appears that this is not utilizing the more modern GPU architecture from the 8 Gen 2. As a result, it appears Qualcomm has added a larger version of their current GPU and very probably increased its clockspeed.

Nonetheless, Qualcomm claims a tremendous 2x performance gain over the 7 Gen 1, a platform that only offered 20% more over its own predecessor. This raises significant performance expectations for the new SoC. Although they aren’t flagship-class SoCs, Qualcomm still tries to market the Snapdragon 7 series as being a suitable match for gaming devices, particularly in China. As a result, it’s not all that unexpected to see Qualcomm investing so much in GPU performance.

Overall, Qualcomm claims that, at least in terms of “long everyday use,” the 7 Gen 2 has 13% more power efficiency. As shown by the 8+ Gen 1 part from last year, switching to TSMC’s 4nm process should yield significant benefits. Yet, it’s also obvious that Qualcomm has been allocating a sizable chunk of those advantages toward enhancing overall performance.

A 32-bit (dual 16-bit) LPDDR5 memory controller feeds the dragon. The 7+ Gen 2 does not enable faster LPDDR5X memory, in contrast to the Snapdragon 8 Gen 2, therefore the Qualcomm 7 series is stuck with the status quo. In this instance, this means that LPDDR5-6400, or 25.6 GB/second of memory bandwidth, is supported.

Speaking of which, not just the CPU and GPU blocks have experienced significant performance gains. Similar to the GPU’s 2x speed boost, Qualcomm’s Hexagon DSP/AI engine block has also seen considerable performance improvements. Although Qualcomm provided few technical details, it appears likely that this is a significantly enhanced version of the Hexagon block used in the previous 7 Gen 1. This is because neither INT4 nor micro-tiling, two major features of the next-generation Hexagon block on the 8 Gen 2, were mentioned in our briefing.

The triple 18-bit Spectra ISP of the Snapdragon 8 is one feature that is being brought down to the Snapdragon 7. Replacing the 14-bit unit found in past generations of the platform, the 18-bit unit on the 7+ Gen 2 will provide capability for triple exposure computational HDR video capture, as well as enhanced low light photography, which Qualcomm terms their Mega Low Light feature. The ultimate result is that the 7+ Gen 2 can now record 4K video at up to 60fps, doubling the 7 Gen 1’s 4K30 limit, and can capture at greater resolutions when employing zero shutter lag capabilities.

In conclusion, the Snapdragon 7+ Gen 2 will be available very soon. Redmi and Realme are two OEMs planning to ship phones based on the new technology, which will be used in devices later this month, according to Qualcomm.

Disclaimer: If you need to edit or update this news from compsmag then kindly contact us Learn more

About The Author

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top