What hardware (CPU/GPU/memory) and network was used for this? What quantization for GLM 5.2? How much tuning of the split was needed?
replyThe lab features two Mac Studios: an Apple M3 Ultra (32 CPU cores, 80 GPU cores, 256 GB unified memory) and an Apple M1 Ultra (20 CPU cores, 48 GPU cores, 128 GB unified memory), both connected via 1Gbit Ethernet.
replyWe use a customized Q2 quantization that preserves sensitive tensors at Q8.
To reduce compute time per layer, we are developing a custom GLM DSA Metal graph.
While we are not yet approaching MTP, we plan to port our existing MTP implementations from versions 4.7 and 5.1 to 5.2.
Since GLM's MTP acceptance rate is very high for a single predicted token, we are exploring token prediction techniques to widen the predicted tokens and utilize parallelism for verification.
Equivalent M3 machines no longer for sale from Apple (only up to 96 GB) but can be had on eBay for around $14,000 each
replyIt's notable that they're so valuable because they feature 800Gbps of memory bandwidth. About twice what's available on the top end of M5, and exactly what makes llm inference fast.
reply> because they feature 800Gbps of memory bandwidth. About twice what's available on the top end of M5
replyOuch, about half of the memory bandwidth of a dedicated GPU though :/ Running LLMs on Apple hardware still doesn't make any sense to me.
M5 max has 614GB/s, you mean the m4?
replyOh, I was looking at every M5 except for the 40-core M5 Max. They have 460.
replyThat sounds cool, but it's still pretty meaningless without information about what your home lab looks like. A few DGX Sparks wired up with their fancy super fast network is much different than a few laptops on wifi.
reply