The Apple M2 is a 8 core processor with a clock frequency of 0.66 GHz (3.50 GHz). The processor can compute 8 threads at the same time. The Apple M1 clocks with 0.60 GHz (3.20 GHz), has 8 CPU cores and can calculate 8 threads in parallel.
The performance values of the processor's AI unit. The isolated NPU performance is specified here, the total AI performance (NPU+CPU+iGPU) can be higher. Processors with support for artificial intelligence (AI) and machine learning (ML) can process many calculations, especially audio, image and video processing, much faster than classic processors.
Graphics (iGPU) integrated into the processor not only enable image output without having to rely on a dedicated graphics solution, but can also efficiently accelerate video playback.
A photo or video codec that is accelerated in hardware can greatly accelerate the working speed of a processor and extend the battery life of notebooks or smartphones when playing videos.
Up to 24 GB of memory in a maximum of 2 memory channels is supported by the Apple M2, while the Apple M1 supports a maximum of 16 GB of memory with a maximum memory bandwidth of 68.2 GB/s enabled.
The Apple M2 has a TDP of 20 W. The TDP of the Apple M1 is 18 W. System integrators use the TDP of the processor as a guide when dimensioning the cooling solution.
The Cinebench 2024 benchmark is based on the Redshift rendering engine, which is also used in Maxon's 3D program Cinema 4D. The benchmark runs are each 10 minutes long to test whether the processor is limited by its heat generation.
The Multi-Core test of the Cinebench 2024 benchmark uses all cpu cores to render using the Redshift rendering engine, which is also used in Maxons Cinema 4D. The benchmark run is 10 minutes long to test whether the processor is limited by its heat generation.
Cinebench R23 is the successor of Cinebench R20 and is also based on the Cinema 4 Suite. Cinema 4 is a worldwide used software to create 3D forms. The single-core test only uses one CPU core, the amount of cores or hyperthreading ability doesn't count.
Cinebench R23 is the successor of Cinebench R20 and is also based on the Cinema 4 Suite. Cinema 4 is a worldwide used software to create 3D forms. The multi-core test involves all CPU cores and taks a big advantage of hyperthreading.
Geekbench 6 is a partial load benchmark for modern computers, notebooks and smartphones. In the single-core test, only the fastest CPU core is measured. The test run simulates the performance in practice.
The practical Geekbench 6 multi-core benchmark tests the system's performance under partial load. The processor's maximum power consumption is far from being used up.
Geekbench 5 is a cross plattform benchmark that heavily uses the systems memory. A fast memory will push the result a lot. The single-core test only uses one CPU core, the amount of cores or hyperthreading ability doesn't count.
Geekbench 5 is a cross plattform benchmark that heavily uses the systems memory. A fast memory will push the result a lot. The multi-core test involves all CPU cores and taks a big advantage of hyperthreading.
The theoretical computing performance of the internal graphics unit of the processor with simple accuracy (32 bit) in GFLOPS. GFLOPS indicates how many billion floating point operations the iGPU can perform per second.
Some of the CPUs listed below have been benchmarked by CPU-monkey. However the majority of CPUs have not been tested and the results have been estimated by a CPU-monkey’s secret proprietary formula. As such they do not accurately reflect the actual Passmark CPU mark values and are not endorsed by PassMark Software Pty Ltd.
Blender is a free 3D graphics software for rendering (creating) 3D bodies, which can also be textured and animated in the software. The Blender benchmark creates predefined scenes and measures the time (s) required for the entire scene. The shorter the time required, the better. We selected bmw27 as the benchmark scene.
Efficiency of the processor under full load in the Cinebench R23 (multi-core) benchmark. The benchmark result is divided by the average energy required (CPU package power in watts). The higher the value, the more efficient the CPU is under full load.
The performance values of the processor's AI unit. The isolated NPU performance is given here, the total AI performance (NPU+CPU+iGPU) can be higher.
Processors with the support of artificial intelligence (AI) and machine learning (ML) can process many calculations, especially audio, image and video processing, much faster than classic processors. The performance is given in the number (trillions) of arithmetic operations per second (TOPS).
In the Blender Benchmark 3.1, the scenes "monster", "junkshop" and "classroom" are rendered and the time required by the system is measured. In our benchmark we test the CPU and not the graphics card. Blender 3.1 was presented as a standalone version in March 2022.
When comparing the Apple M2 and Apple M1, we find that not much has changed at all. The Apple M2 can be seen as a slight further development of the Apple M1, in which Apple has mainly screwed on the clock frequencies of the CPU cores.
The clock frequency of the Apple M2 is now up to 3.5 GHz, where the Apple M1 stopped at 3.2 GHz. The smaller efficiency cores now clock at 2.8 GHz instead of 2.06 GHz on the Apple M1. The clock frequency of the integrated graphics has also been increased slightly: from 1.3 GHz to 1.4 GHz in the Apple M2.
At the same time as increasing the clock, Apple also increased the number of graphics cores. Instead of 8 GPU cores, 10 GPU cores are now installed as standard. The number of SM processors (execution units) increases from 128 to 160. The Apple M2 now has 1280 texture shaders installed, previously there were 1024.
In addition to the change to the graphic itself, it can now also reserve significantly more memory for itself: it has access to the entire memory, which can now be up to 24 GB (16 GB was the maximum in the Apple M1). The memory in the Apple M2 is now LPDDR5-6400 with a maximum memory bandwidth of 102GB per second.
The Apple M1 only comes with LPDDR4X-4266 memory to 68 GB per second. The number of memory channels is identical. With the level 2 cache, Apple has given the M2 4 MB more, and the Apple M2 now has a total of 20 MB L2 cache.
Due to the similar CPU design, Apple had to increase the TDP (thermal heat dissipation of the processor) from 15 to 22 watts in order to accommodate the increased clock frequencies in the CPU and GPU part as well as the slightly larger GPU with 10 GPU cores.
The Apple M2 is manufactured at TSMC using the 5 nm process. This is the same manufacturing process used in the Apple M1. The Apple M processors are still based on a chiplet design, which makes scaling to larger configurations (Pro/Max/Ultra) relatively easy.