Apple M3
Benchmarks & Specs

With the Apple M3, the manufacturer has introduced its first third-generation processor. The SoC contains a CPU part that has the same features as the Apple M1 or Apple M2 with 4 P cores and 4 E cores. The cores now clock a little higher at up to 4.06 GHz. This also results in slightly higher CPU performance compared to its predecessors.

The jump in performance compared to the Apple M2 is between 10 and 20 percent and mainly comes from the increased clock frequency. However, minor optimizations have further increased the efficiency of the processor. In our Cinebench R23 efficiency benchmark, which compares the multi-core performance achieved with the energy consumption under full load, the Apple M3 currently takes first place.

The increased efficiency is certainly also due to TSMCs new manufacturing process. TSMC now manufactures the Apple M3 in 3 nm, while the Apple M2 was still manufactured in 5 nm. Due to the increased performance, the TDP of the Apple M3 was increased to 22 watts, while its predecessor still had to make do with 20 watts.

Apple has not touched the memory connection compared to the Apple M2. Up to 24 GB of LPDDR5-6400 memory can still be addressed. The maximum bandwidth is 102.4 GB/s. What is new is the management of shared memory. This is now divided more intelligently between the CPU and the internal graphics (iGPU). Apple calls this “Dynamic Cache”. The new memory management should give the Apple M3 significantly better graphics performance.

And although the pure FP32 raw graphics performance of 3.55 TFLOPS has not changed compared to the Apple M2, the Apple M3 achieves really better frame rates in practice. The AI area of the SoC, which is now twice as large at 35 TOPS, also contributes to this. At the same time, the internal graphics now support hardware decoding of the AV1 code and enable ray tracing in hardware.

• 3-nanometer manufacturing
• Strong Single and Multicore Performance
• Integrated Apple M3 (10 Core) graphics
• High energy efficiency