The ARM has unveiled its next generation CPU and GPU designs and is re -branding something at the same time – meet the ARMC1 CPU and G1GPU, which will create the arm Lomax Compute Sub System (CSS). “Lumax” is the new branding that will use the arm for mobile -based designs (for others, “NIVA” will be used for PC, “Zina” and so on for automotive).
This generation of CPU and GPU designs improves a large scale in the burden workload (AI, Ray tracing), while also provides solid improvement to the low shiny workload.
Note that the arm has also developed the interior chip set, which we will add to a separate post.
Let’s look at the CPU first and we will start with a cheating sheet to help you find you new branding-Cartax-X and Cartax-A are no more!
- Cortex-X9XX → Arm C1-ULTRA
- (New) arm C1-Premium
- Cortex-A7xx → Arm C1-Par
- Cortex-A5xx → Arm C1-Nano
The new C1 core is first built on Armv9.3 architecture. All of them support ARM’s Scale Matrix Extension 2 (SME 2), which accelerates the workload of AI, but also improves more normal things – such as SME2 to regulate HDR video is 10 % more efficient.
Arm C 1 Ultra Focuses on increasing IPC (IPC). In single threaded tasks, the C1 ultra-Cartax-X is 925 to 25 % faster.
The arm has also introduced a new rank of CPU core, Arm C1-Premium. Its purpose is a sub-flagship design-it has a 35 % smaller surface area of ultra (which means the chips will be cheaper), which provides it with the best class area performance (ie silicon’s performance per mm).
Arm C1-Pro The core will be used as a large cover for permanent work loads and medium -range designs in high performance chips. Compared to the Cartax-A725, it can provide 16 % more performance in things like gaming. As an alternative, it can be more effective for video playbacks, web browsing and 12 % more power for social media.
Arm C1- nano The changes are almost fully focused on the power performance-this small core is efficiently efficiently from 520 to 26 % more powerful. It improves modest performance and also has a 2 % smaller core area.
The CPU Corps will be arranged by the new C1-Dynamiq Unit (DSU). It is responsible for all the core L3 caches, power management for cores and more. The new DSU enables up to 26 % of power saving compared to the previous DSU-120.
An arm C1 CPU cluster can be formed with at least 1 CPU cover and a maximum of 14 core. Chaspus designers can combine three basic varieties, choosing between ultra, premium, pro and nano.
ARM claims that a C -1 CPU cluster industry’s leading benchmark of real -world work burden brings an average of 15 % speed -up for things like gaming and video streaming. In addition, it uses an average of 12 % less energy for things like video playback, web browsing and social media compared to the previous generation CPU.
As we have mentioned above, the new C1 core SME2 extends to improve mass performance for AI. We are talking for the 4.7X low delays for the Wishes Base (an automatic speech identification model, ie text from text), the 4.7X for the Google Jima 3 model 4.7x for high AI performance and stability AI 2.8X Audio Generation (a text of audio models that can produce background audio, music and more).
The app developers will get better performance on the Next gene hardware as the SME2 support itself is connected to the AI Framework from ARM, Google, Microsoft, Alibaba and Meta.
According to a recent study, 83 % of the overall games play on mobile. This is a lucrative business in terms of sports and hardware. Arm says it has sent more than 12 billion GPUs to date – and they are still the most powerful.
Sleeve Mali G1 Ultra The second-generation ray tracing unit (RTUV2) introduces, which doubles the performance of the ray tracing compared to the Amlas-G-925 GPU.
Note that there is more to offer game scenes compared to ray tracing, so you can expect to see a 40 % more frame rate in sports using hardware ray tracing. In addition, RTUV2 is now a separate module and brings numerous reforms of power performance, including a simple – when the system can sleep when it is useless.
Putting the Ray Tracing aside, Finanian G -One Ultra can reach 20 % more rusty performance than the G925. It can also be 9 % more powerful. GPUs can also be used for AI and thanks to the new GP -16 Matrix computing path, the G1 Ultra AI -in -to -faster is 20 % faster.
There are some more solid numbers in popular sports:
| Arena Breakout | +25 % |
| Hanki Star Rail | +19 % |
| Gennashi Effect | +17 % |
| Foretenite | +11 % |
| Sports demo (Mori) inside the house | +26 % |
ARM Mali-G1 GPUs can be formed between 1 and 24 sheder core. The G -1 GPUSRM accuracy supports Super Resolution (ASR), which is the company’s worldly advanced tech (Think about DLSS). It is already supported by unrealistic engine 5 and is integrated into Fortnite.
Of course, you don’t yet have a smartphone with C1 CPU and G1 GPU. Arm says the next generation’s hardware will be “in consumer devices in the near future”.
Source 1 | Source 2
