Apple’s M1 Ultra surpasses latest Intel and AMD CPUs in benchmark tests

Apple’s latest desktop chip, the M1 Ultra that fuses two M1 Max chips together using an interconnect bus, has spurred comparisons with rival desktop processors. A benchmark test on Geekbench has revealed that Apple’s chip is faster than some of the leading desktop chips.   

According to a Tom’s Hardware report, M1 Ultra generated a score of 24,055 on Geekbench in multi-threaded performance, which puts it slightly behind the AMD Ryzen Threadripper 3990X and way ahead of the Intel Core i9 12900K and AMD Ryzen 9 5950X. 

Ryzen Threadripper 3990X, which is one of the fastest desktop processors, scored 25,133 on the Geekbench test, despite the fact that it has 64 cores, while the M1 Ultra has only 20 cores.  

The other two processors in contention — Intel’s Core i9-12900K and AMD Ryzen 9 5950X, both of which have 16 cores could only muster a score of 17,204 and 16,565, respectively.  

Core is a processing unit made of millions of transistors in a CPU and is responsible for receiving and executing instructions. In the earlier days, most CPUs had a single core that could focus only on one task at a time. Adding more cores has made modern day CPUs more capable at handling multiple tasks with the same efficiency and speed.  

To boost performance further, chip makers came up with a concept called multi-threading or hyperthreading, which splits a core into virtual cores, referred to as threads.   

Apple during the Spring event 2022, had also claimed that the M1 Ultra delivers 90% higher multi-threaded performance than the fastest available 16-core desktop chip with similar power capacity.  

Though the M1 Ultra had an upper hand over two of its closest rivals in the multi-threading performance, the Intel Core i9-12900K stole a march over M1 Ultra with a score of 1,997 in single threaded performance. M1 Ultra scored 1,793, while AMD Ryzen 5950X scored 1,686. 

In a single threaded application, the entire process is executed by a single thread, while in a multi-threaded application, multiple threads handle different tasks simultaneously.