Hacker News new | past | comments | ask | show | jobs | submit login

I found a bunch of ML benchmarks, only one had a comparison to a RTX3070. Seems quite a bit better than "not able to run anything". In particular you can use up to 96GB ram, 4x the 4090. Granted it's m2 max is approximately 3070 in speed for such workloads, at least it doesn't decrease when unplugged.

  ResNet50:
  M1 amx   = 155.3
  M2 max   = 188.8
  M1 ultra = 242

  mobilenetv2
  M1 mmx   = 415.7
  M2 max   = 467.9
  M1 ultra = 634.4

  DistilBERT
  M1 mmx   = 134
  M2 max   = 164
  M1 ultra = 215

  BERTLarge
  M1 mmx   = 20
  M2 max   = 25
  M1 ultra = 30
  
  CIFAR-10: 3 batch sizes
  RTX3070  240, 283, 328
  M1 Max   297, 358, 443
  M2 Max   242, 280, 327
  M1 Ultra 207, 308, 453

  Densenet 121: 3 batch sizes
  M1 Max    57  104  704
  M2 Max    66  124  831
  M1 Ultra  49   99  614



Consider applying for YC's W25 batch! Applications are open till Nov 12.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: