3 ms·
The M6 doubled the neural engine from 16 to 32 cores. I would expect that the M7 doubles that again to 64 from 32? That would make sense. I believe that the
by bhouston 2mo ago
The M6 doubled the neural engine from 16 to 32 cores. I would expect that the M7 doubles that again to 64 from 32? That would make sense.
I believe that the CPUs are actually limited by ram bandwidth more than the neural engine right when it comes to LLM processing?
Maybe the M7 introduces something new to get around the current ram bandwidth problems on the non-Ultra chips.
- wmf 2mo agoLPDDR6 is coming.
- bhouston 2mo agoI understand that will boost read rates to around 14 Gbps as compared to the current 10 Gbps for LPDDR5X, so a 40% improvement.
- wmf 2mo agoThe bus is also 50% wider so the bandwidth is double.
- bhouston 2mo agoOr they could make use of LPDDR5X-PIM? That would be such a killer feature and competitive advantage.
- nomel 2mo agohttps://www.tomshardware.com/pc-components/dram/hot-chips-2026-samsung-makes-lpddr5x-smart-with-logic-unit-in-memory-lpddr5x-pim-is-3-01x-faster-than-lpddr5x-in-ai-inference-with-8x-the-bandwidth https://www.tomshardware.com/pc-components/dram/hot-chips-20... Super cool!
- bigyabai 2mo agoApple's biggest bottleneck for real-world inference is prefill processing. They need a better GPGPU architecture, which is what I'm expecting M7 to reveal.
- tedd4u 2mo agoAgreed, seems like the M5 has already made steps in that direction, with 4x prompt processing / prefill performance vs. M4. [1] And token generation also got a 10% boost. The data below is only for Pro & Max but I think the base M5 got the same relative boosts vs. M4 base. Chip BW (GB/s) GPU Cores Q4_0 Prompt Q4_0 Gen M4 Pro (20c) 273 20 439.78 50.74 M4 Max (40c) 546 40 885.68 83.06 M5 Pro (20c) 307 20 ~1500 to 1700 ~56 M5 Max (40c) 614 40 ~3000 to 3500 ~92 [1] https://www.hardware-corner.net/m5-pro-m5-max-local-llm-4x-faster-prefill-202061108/ https://www.hardware-corner.net/m5-pro-m5-max-local-llm-4x-f...
- JumpCrisscross 2mo agoELI5?
- tedd4u 2mo agoLook at the "Q4_0 Prompt" column. That's the tokens-per-second processing the initial prompt/system prompt. This is where the "neural accelerators inside each GPU core" is seen most prominently.
- giantrobot 2mo agoThe M5 GPU added better matmul and dot product support in hardware (IIRC) that really boosted performance of those kernels over previous GPUs. Current transformer models rely heavily on matmuls. Prefill (processing the current context) is mostly limited by raw GPU processing power where token generation (predicting the next token from the current context) is mostly limited by memory bandwidth. While an older M-series might run a matmul or dot product kernel just fine the M5 can run them much faster.