Apple's desktop and notebook processors traditionally lead the pack in single-thread workloads, as industry-leading single-thread performance has been the company's focus for a long time. However, Apple's M5 Max processors not only outperform rivals by a huge margin in single-thread workloads, but beat all of them — including the 96-core AMD Ryzen Threadripper Pro 9995WX — in multi-thread workloads in the Geekbench 6 benchmark. However, when it comes to GPU compute performance, not everything is that rosy for the M5 Max.

Single-thread and multi-thread champion

According to recent Geekbench 6 results , Apple's 18-core M5 Max not only beats its direct predecessor, the 16-core M4 Max, in single-thread (4,353 points) and multi-thread workloads (29,644 points), but also the 32-core M3 Ultra that is supposed to be Apple's unbeatable multi-thread machine.

Furthermore, Apple's new flagship CPU beats AMD's 96-core Ryzen Threadripper Pro 9995WX in single-thread (which is not surprising) and multi-thread workloads in Geekbench 6. It should be noted that while most Threadripper Pro 9995WX CPUs score around 26,000 GB6 points in multi-thread workloads, there is one example when this processor hits 30,170 points, which is a bit ahead of M5 Max's 29,644 points.

There is a major catch here as the Geekbench 6 multi-thread benchmark is a brief, bursty test intended to mimic common consumer tasks such as archive compression, PDF processing, and image editing. Its short runtime and bursty nature prevent it from fully stressing ultra-high-core-count processors like the Ryzen Threadripper Pro 9995WX.

Furthermore, many of the suite’s multi-threaded subtests scale efficiently only to roughly 8 – 32 threads, which leaves much of such CPUs' parallel capacity idle, but which creates an almost perfect environment for Apple's CPUs that feature a relatively modest number of cores, but which evolve noticeably in terms of per-core performance from one generation to another. Also, keep in mind that Geekbench 6 is a synthetic benchmark that reflects the potential of the tested hardware but may not reflect its performance in real-world applications.

Apple's M5 Max processor in its maximum configuration packs six 'super' performance (SP) cores featuring increased front-end bandwidth (i.e., wider decoder?), enhanced branch prediction, and a new cache hierarchy to deliver unbeatable single-thread performance as well as 12 new performance (P) cores designed to deliver power-efficient multithreaded performance in professional applications, up from 16 cores (12P + 4E cores) offered by the M4 Max. We do not know details about microarchitectures of Apple's 'super' performance and performance cores, though the 12% single-thread performance difference between M5 Max's SP and M4 Max's P cores is evident.

As for the memory subsystem, the M5 Max features up to 128 GB of LPDDR5X-9600 memory connected to the host via a 512-bit interface, offering 614 GB/s of bandwidth, up 12% from M4 Max (546 GB/s). For now, no workstation processor can match the memory bandwidth of M5 Max or M4 Max. Efficient cache and memory subsystems are crucial for single-thread performance, so this part of the M5 Max also played a significant role in its performance boost compared to the predecessor.

Not quite a GeForce RTX 5090

In addition to its revamped CPU subsystem, Apple's M5 Max also boasts a new GPU that is based on a PowerVR-derived microarchitecture developed by Apple. As it turns out, a big integrated GPU and plenty of memory bandwidth can deliver serious GPU compute oomph: the M5 Max scores 232,718 points on the GeekBench 6 GPU compute benchmark when using the Metal API. Apple's previous-generation M4 Max scores up to 204,453 points in the same tests. Evidently, the new GPU is better than the predecessor, but not that significantly.

When compared to non-Apple GPUs, the one inside the M5 Max easily beats the iGPU inside the Ryzen AI Max+ 395, which scores 133,447 points when unconstrained by thermals. When it comes to discrete graphics cards, Apple's flagship iGPU is ahead of Nvidia's GeForce RTX 5070 ( 207,061 points , Vulkan ), but trails the GeForce RTX 5070 Ti ( 253,890 points, Vulkan) and has no chance against the GeForce RTX 5090. Still, building an integrated GPU that delivers compute performance comparable to one of the best graphics cards is a breakthrough.

Follow Tom's Hardware on Google News , or add us as a preferred source , to get our latest news, analysis, & reviews in your feeds.