The Apple M3 Max (16 Core) is a system on a chip (SoC) from Apple for notebooks that was launched towards the end of 2023. It integrates a new 16-core CPU with 12 performance cores with up to 4.06 GHz and 4 efficiency cores with 2.8 GHz. There is also a slimmed-down 14-core variant with a 30-core GPU.
Thanks to the higher clock rates and architectural improvements, the processor performance is also significantly better than the M2 Max in benchmarks and can keep up with the fastest mobile CPUs (such as a Core i9-13900HX).
The M3 also integrates a new graphics card with dynamic caching, mesh shading and ray tracing acceleration via hardware. In the top model, all 40 cores of the chip are used and support up to 5 displays simultaneously (internal and 4 external).
GPU and CPU can jointly access the shared memory on the package (unified memory). This is available in 48, 64 and 128 GB variants and offers 400 GB/s maximum bandwidth (512 bit bus).
The integrated 16-core Neural Engine has also been revised and now offers 18 TOPS peak performance (compared to 15.8 TOPS in the M2 but 35 TOPS in the new A17 Pro). The video engine now also supports AV1 decoding in hardware. H.264, HEVC and ProRes (RAW) can still be decoded and encoded. Like its predecessor, the Max chip offers two video engines and can therefore encode and decode two streams simultaneously.
Unfortunately, the integrated WLAN only continues to support WiFi 6E (no WiFi 7), unlike the small M3 SoC thunderbolt 4 is also supported (max 40 Gbit/s).
The chip is manufactured in the current 3nm process (N3B) at TSMC and contains 92 billion transistors (+37% vs. Apple M2 Max). Under load, the CPU part consumes up to 56 watts, the chip can use a total of 78 watts.
The Snapdragon X Elite X1E-80-100 is a pretty fast ARM architecture processor (SoC) for use in Windows laptops. The X1E-80-100 is the second-fastest member of the Snapdragon X series, as of May 2024; as such, it has 12 Oryon CPU cores (3 clusters of equally powerful cores; 12 threads) running at up to 4.0 GHz, the 3.8 TFLOPS Adreno X1-85 GPU, the 45 TOPS Hexagon NPU and an impressively fast LPDDR5x-8448 memory controller.
The faster Snapdragon X Elite X1E-84-100 is thought to be based on the same die with the same NPU and the same number of CPU cores but with a higher GPU and CPU clock speeds.
Architecture and Features
Qualcomm Oryon cores are in part based on Nuvia IP; they most likely make use of the ARM v8.7 microarchitecture. Much like modern AMD and Intel processors, the Snapdragon chip is compatible with USB 4 and thus with Thunderbolt 4 but it remains to be seen if it is is compatible with GeForce or Radeon discrete graphics cards.
The Qualcomm SoC is thought to have at least 12 PCIe 4 and 4 PCIe 3 lanes for connecting various kinds of devices. NVMe SSDs are supported with a throughput of up to 7.9 GB/s; furthermore, most laptops built around the chip are expected to have 16 GB of LPDDR5X-8448 RAM. There is also a 45 TOPS NPU for accelerating AI workloads.
Performance
When all of the 12 cores are under load, they can run at up to 3.4 GHz. The much more desirable 4.0 GHz clock speed is only achievable under single-thread or dual-thread loads.
Given the very similar clock speeds, its multi-thread benchmark scores should be about the same as that of the X1E-78-100 for Ryzen 7 7840U and Apple M3-like performance in most real-world tasks. We'll make sure to update this section once we get our hands on a system powered by the X1E-80-100.
Like any other Windows on ARM platform, the Snapdragon X chip works best with applications and games compiled specifically for ARM processors. If making use of the built-in x86 to ARM emulation mode, a performance penalty of about 20% is to be expected.
Graphics
The integrated 3.8 TFLOPS Adreno X1-85 GPU is DirectX 12-enabled but not DirectX 12 Ultimate-enabled, despite featuring VRS and ray tracing support. The 3.8 TFLOPS of performance that it delivers pale in comparison to the slowest current-generation professional Nvidia Ada graphics card for laptops, the RTX 500 Ada, that delivers up to 9.2 TFLOPS.
Its gaming performance is somewhat underwhelming. While slightly faster than the aging Iris Xe (96 EUs), the Adreno is not fast enough to outpace the Radeon 780M meaning there is little point in comparing it with full-fat discrete graphics cards. Baldur's Gate 3, a triple-A title released in 2023, is pretty much unplayable at 1080p / Low. GTA V, a game that saw the light of day in 2015, is playable at 1080p / High.
The Qualcomm iGPU will let you use up to 3 UHD 2160p monitors simultaneously. 2160p120 integrated displays are supported, as are the popular AV1, HEVC and AVC video codecs (both decoding and encoding). The hardware does not support the VVC codec; that feature remains a Lunar Lake exclusive for now.
Power consumption
The X1E-80-100 is most likely going to be less power-hungry than its more powerful brother (the latter can be set to consume up to 80 Watts). Expect to see anything between 20 W and 45 W under long-term workloads because that's what Ryzen HS chips normally consume.
The SoC is built with a 4 nm TSMC process for decent, as of H1 2024, energy efficiency.
The Apple M3 is a system on a chip (SoC) from Apple for notebooks that was introduced in late 2023. It integrates a new 8-core CPU with 4 performance cores with up to 4.06 GHz and 4 efficiency cores running at up to 2.75 GHz. Apple claims that the CPU is up to 20% faster than in the old Apple M2 (3.5 GHz).
Due to the higher clock speeds and architecture improvements, the processor performance is also significantly better than the M2 in benchmarks (see e.g. Geekbench below) and can keep up with the fastest CPUs in short single-core tests (like the Raptor Lake i9-13950HX).
The M3 also integrates a new graphics adapter with dynamic caching, mesh shading and ray tracing acceleration. According to Apple, it is 20% faster than the GPU in the M2. The chip integrates again 10 GPU cores, but the cheaper variant only offers 8 cores (e.g. in the entry iMac). Furthermore, the GPU only supports 2 displays (an additional 6K60 display to the internal one).
Both GPU and CPU can access the unified memory on the package together. It is still available in 8, 16 and 24 GB variants and offers the same 100 GB/s maximum bandwidth (unlike the Pro models that feature a reduced memory bandwidth).
The integrated 16-core Neural Engine has also been revised and now offers 18 TOPS peak performance (versus 15.8 TOPS in the M2 but 35 TOPS in the new A17 Pro). The video engine now supports AV1 decoding in hardware. H.264, HEVC and ProRes (RAW) can still be decoded and encoded.
Unfortunately, the integrated wireless network module only supports Wi-Fi 6E (no Wi-Fi 7) and due to the support of only a single external monitor, the chip also has to make do with no Thunderbolt 4 (Thunderbolt 3 / USB 4 support only for up to 40 Gbit/s).
The chip is manufactured on the current 3nm TSMC process (N3B most likely) and contains 25 billion transistors (+25% vs. Apple M2). The 3nm process should also contribute to the excellent efficiency of the chip. Under load, the M3 CPU consumes approximately 20 Watt.
Average Benchmarks Apple M3 Max 16-Core → 100%n=16
Average Benchmarks Qualcomm Snapdragon X Elite X1E-80-100 → 81%n=16
Average Benchmarks Apple M3 → 82%n=16
- Range of benchmark values for this graphics card - Average benchmark values for this graphics card * Smaller numbers mean a higher performance 1 This benchmark is not used for the average calculation
v1.28
log 02. 06:03:37
#0 checking url part for id 15113 +0s ... 0s
#1 checking url part for id 17575 +0s ... 0s
#2 checking url part for id 15110 +0s ... 0s
#3 redirected to Ajax server, took 1719893017 +0s ... 0s
#4 did not recreate cache, as it is less than 5 days old! Created at Mon, 01 Jul 2024 05:15:43 +0200 +0s ... 0s
#5 composed specs +0.011s ... 0.011s
#6 did output specs +0s ... 0.011s
#7 getting avg benchmarks for device 15113 +0.001s ... 0.011s
#8 got single benchmarks 15113 +0.004s ... 0.016s
#9 getting avg benchmarks for device 17575 +0.004s ... 0.019s
#10 got single benchmarks 17575 +0.004s ... 0.023s
#11 getting avg benchmarks for device 15110 +0.001s ... 0.024s
#12 got single benchmarks 15110 +0.005s ... 0.029s
#13 got avg benchmarks for devices +0s ... 0.029s
#14 min, max, avg, median took s +0.028s ... 0.057s