Apple M1 Ultra (64-GPU)
Benchmarks & Specs

In the Apple M1 Ultra, Apple relies on a total of 20 CPU cores. The clock frequency of 3.2 GHz (performance cores) and 2.06 GHz (efficiency cores) is on the same level as that of the normal Apple M1 or the larger expansion stages Apple M1 Pro and Apple M1 Max.

The Apple M1 Ultra with 64 GPU cores is currently the largest and fastest Apple processor. It combines two individual Apple M1 Max processor chips via an interface with a bandwidth of 2.5 TB/s. AMD also uses a similar technology in its current AMD Ryzen processors.

In addition to the CPU cores, the two graphics units are also synchronized via the interface of the two processors in the Apple M1 Ultra. According to Apple, the interface itself consists of 1000 separate lines, which make such a large bandwidth possible in the first place.

The memory bandwidth of the LPDDR5-6400 memory is almost 820 GB/s, which is approx. 10x as fast as current desktop processors from AMD and Intel. Since Apple does not use an external graphics unit in its new products, the memory has to be fast enough not to slow down the GPU units of the Apple M1 Ultra. The processor and graphics unit dynamically share the memory of up to 128 GB.

The Apple M1 Ultra with 64 GPU cores achieves an FP32 processing performance (single precision) of 21.2 TFLOP/s. The strongest Apple iGPU is about 5x as fast as the fastest iGPU of the competition AMD Ryzen 9 6980HX. When it comes to graphics performance, the Apple M1 Ultra does not need to fear comparison with external graphics cards from AMD or NVIDIA. The FP32 performance is roughly on par with an AMD Radeon RX 6800XT or an NVIDIA RTX 3070.

Thanks to the latest 5 nm production and the energy-efficient ARM chip design, the energy consumption of the Apple M1 Ultra is also impressive and is slightly above the latest AMD Ryzen 6000 mobile processors.

• High integrated graphics performance (64 GPU cores)
• Excellent AI performance through 819 GB/s memory bandwidth
• Strong Multi-Core performance (1624 points Cinebench 2024)
• Efficient manufacturing in 5 nm process