Skip to content

Commit d9ac4ff

Browse files
tonyjohnchentonyjohnchen
authored
Merge pull request #245 from weikuo0506/update-other-training-readmes
docs: Update A3 Mega, A3 Ultra, and A4 training support matrices
2 parents 31ec5b4 + 1f69e7f commit d9ac4ff

1 file changed

Lines changed: 37 additions & 16 deletions

File tree

‎README.md‎

Lines changed: 37 additions & 16 deletions
Original file line numberDiff line numberDiff line change
@@ -20,32 +20,53 @@ Welcome to the reproducible benchmark recipes repository for GPUs! This reposito
2020

2121
Models | GPU Machine Type | Framework | Workload Type | Orchestrator | Link to the recipe
2222
----------------- | --------------------------------------------------------------------------------------------------------- | --------- | ------------- | ------------ | ------------------
23-
**GPT3-175B** | [A3 Mega (NVIDIA H100)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-mega-vms) | NeMo | Pre-training | GKE | [Link](./training/a3mega/gpt3-175b/nemo-pretraining-gke/README.md)
24-
**Llama-3-70B** | [A3 Mega (NVIDIA H100)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-mega-vms) | NeMo | Pre-training | GKE | [Link](./training/a3mega/llama3-70b/nemo-pretraining-gke/README.md)
25-
**Llama-3.1-70B** | [A3 Mega (NVIDIA H100)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-mega-vms) | NeMo | Pre-training | GKE | [Link](./training/a3mega/llama3-1-70b/nemo-pretraining-gke/README.md)
26-
**Mixtral-8-7B** | [A3 Mega (NVIDIA H100)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-mega-vms) | NeMo | Pre-training | GKE | [Link](./training/a3mega/mixtral-8x7b/nemo-pretraining-gke/README.md)
23+
**GPT3-175B** | [A3 Mega (NVIDIA H100)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-mega-vms) | NeMo (25.07) | Pre-training | GKE | [Link](./training/a3mega/gpt3_175b/nemo-gke/nemo2507/recipe/)
24+
**Llama-3-70B** | [A3 Mega (NVIDIA H100)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-mega-vms) | NeMo (25.07) | Pre-training | GKE | [Link](./training/a3mega/llama3_70b/nemo-gke/nemo2507/128gpus-bf16/recipe/)
25+
**Mixtral-8-7B** | [A3 Mega (NVIDIA H100)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-mega-vms) | NeMo (25.07) | Pre-training | GKE | [Link](./training/a3mega/mixtral_8x7b/nemo-gke/nemo2507/recipe/)
2726

2827

2928
### Training benchmarks A3 Ultra
3029

3130
Models | GPU Machine Type | Framework | Workload Type | Orchestrator | Link to the recipe
3231
------------------ | ----------------------------------------------------------------------------------------------------------- | --------- | ------------- | ------------ | ------------------
33-
**Llama-3.1-70B** | [A3 Ultra (NVIDIA H200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-ultra-vms) | MaxText | Pre-training | GKE | [Link](./training/a3ultra/llama3-1-70b/maxtext-pretraining-gke/README.md)
34-
**Llama-3.1-70B** | [A3 Ultra (NVIDIA H200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-ultra-vms) | NeMo | Pre-training | GKE | [Link](./training/a3ultra/llama3-1-70b/nemo-pretraining-gke/README.md)
35-
**Llama-3.1-405B** | [A3 Ultra (NVIDIA H200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-ultra-vms) | MaxText | Pre-training | GKE | [Link](./training/a3ultra/llama3-1-405b/maxtext-pretraining-gke/README.md)
36-
**Llama-3.1-405B** | [A3 Ultra (NVIDIA H200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-ultra-vms) | NeMo. | Pre-training | GKE | [Link](./training/a3ultra/llama3-1-405b/nemo-pretraining-gke/README.md)
37-
**Mixtral-8-7B** | [A3 Ultra (NVIDIA H200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-ultra-vms) | NeMo | Pre-training | GKE | [Link](./training/a3ultra/mixtral-8x7b/nemo-pretraining-gke/README.md)
32+
**Llama-3.1-70B** | [A3 Ultra (NVIDIA H200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-ultra-vms) | MaxText | Pre-training | GKE | [Link](./training/a3ultra/llama3-1-70b/maxtext-pretraining-gke/README.md)
33+
**Llama-3.1-70B** | [A3 Ultra (NVIDIA H200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-ultra-vms) | NeMo (24.07) | Pre-training | GKE | [Link](./training/a3ultra/llama3_70b/nemo-gke/nemo2407/recipe/)
34+
**Llama-3-70B** | [A3 Ultra (NVIDIA H200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-ultra-vms) | Megatron-Bridge (26.02) | Pre-training | GKE | [Link](./training/a3ultra/llama3_70b/megatron-bridge-gke/nemo2602/)
35+
**Llama-3-70B** | [A3 Ultra (NVIDIA H200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-ultra-vms) | Megatron-Bridge (25.11) | Pre-training | Slurm | [Link](./training/a3ultra/llama3_70b/megatron-bridge-slurm/nemo2511/)
36+
**Llama-3-8B** | [A3 Ultra (NVIDIA H200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-ultra-vms) | Megatron-Bridge (25.11) | Pre-training | Slurm | [Link](./training/a3ultra/llama3_8b/megatron-bridge-slurm/nemo2511/)
37+
**Llama-3.1-405B** | [A3 Ultra (NVIDIA H200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-ultra-vms) | MaxText | Pre-training | GKE | [Link](./training/a3ultra/llama3-1-405b/maxtext-pretraining-gke/README.md)
38+
**Llama-3.1-405B** | [A3 Ultra (NVIDIA H200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-ultra-vms) | NeMo (24.12) | Pre-training | GKE | [Link](./training/a3ultra/llama31_405b/nemo-gke/nemo2412/recipe/)
39+
**Mixtral-8-7B** | [A3 Ultra (NVIDIA H200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-ultra-vms) | NeMo (24.07) | Pre-training | GKE | [Link](./training/a3ultra/mixtral_8x7b/nemo-gke/nemo2407/recipe/)
40+
**DeepSeek-V3** | [A3 Ultra (NVIDIA H200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-ultra-vms) | Megatron-Bridge (26.02) | Pre-training | GKE | [Link](./training/a3ultra/deepseek_v3/megatron-bridge-gke/nemo2602/)
41+
**GPT OSS 120B** | [A3 Ultra (NVIDIA H200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-ultra-vms) | NeMo (26.02) | Pre-training | GKE | [Link](./training/a3ultra/gpt_oss_120b/nemo-gke/nemo2602/)
42+
**Qwen-3-30B** | [A3 Ultra (NVIDIA H200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-ultra-vms) | NeMo (26.02) | Pre-training | GKE | [Link](./training/a3ultra/qwen3_30b_a3b/nemo-gke/nemo2602/)
43+
**Wan-2.1** | [A3 Ultra (NVIDIA H200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-ultra-vms) | Megatron-Bridge (26.02) | Pre-training | GKE | [Link](./training/a3ultra/wan/megatron-bridge-gke/nemo2602/)
44+
3845

3946
### Training benchmarks A4
4047

4148
Models | GPU Machine Type | Framework / Library | Workload Type | Orchestrator | Link to the recipe
42-
------------------ | ---------------------------------------------------------------------------------------------------- | --------- | ------------- | ------------ | ------------------
43-
**Llama-3.1-70B** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | MaxText | Pre-training | GKE | [Link](./training/a4/llama3-1-70b/maxtext-pretraining-gke/README.md)
44-
**Llama-3.1-70B** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | NeMo | Pre-training | GKE | [Link](./training/a4/llama3-1-70b/nemo-pretraining-gke)
45-
**Llama-3.1-405B** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | MaxText | Pre-training | GKE | [Link](./training/a4/llama3-1-405b/maxtext-pretraining-gke/README.md)
46-
**Llama-3.1-405B** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | NeMo | Pre-training | GKE | [Link](./training/a4/llama3-1-405b/nemo-pretraining-gke/README.md)
47-
**Mixtral-8-7B** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | NeMo | Pre-training | GKE | [Link](./training/a4/mixtral-8x7b/nemo-pretraining-gke/README.md)
48-
**PaliGemma2** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | Hugging Face Accelerate | Finetuning | GKE | [Link](./training/a4/paligemma2/README.md)
49+
------------------ | ---------------------------------------------------------------------------------------------------- | ------------------- | ------------- | ------------ | ------------------
50+
**Llama-3.1-70B** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | MaxText | Pre-training | GKE | [Link](./training/a4/llama3-1-70b/maxtext-pretraining-gke/README.md)
51+
**Llama-3.1-70B** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | NeMo (25.07) | Pre-training | GKE | [Link](./training/a4/llama3_70b/nemo-gke/nemo2507/)
52+
**Llama-3.1-70B** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | NeMo (26.02) | Pre-training | GKE | [Link](./training/a4/llama3_70b/nemo-gke/nemo2602/)
53+
**Llama-3.1-70B** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | Megatron-Bridge (25.09) | Pre-training | Slurm | [Link](./training/a4/llama3_70b/megatron-bridge-slurm/nemo2509/)
54+
**Llama-3.1-405B** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | MaxText | Pre-training | GKE | [Link](./training/a4/llama3-1-405b/maxtext-pretraining-gke/README.md)
55+
**Llama-3.1-405B** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | NeMo (25.07) | Pre-training | GKE | [Link](./training/a4/llama31_405b/nemo-gke/nemo2507/)
56+
**Llama-3.1-405B** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | NeMo (26.02) | Pre-training | GKE | [Link](./training/a4/llama31_405b/nemo-gke/nemo2602/)
57+
**Llama-3.1-405B** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | Megatron-Bridge (25.09) | Pre-training | Slurm | [Link](./training/a4/llama31_405b/megatron-bridge-slurm/nemo2509/)
58+
**Mixtral-8-7B** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | NeMo (25.07) | Pre-training | GKE | [Link](./training/a4/mixtral_8x7b/nemo-gke/nemo2507/recipe/)
59+
**PaliGemma2** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | Hugging Face Accelerate | Finetuning | GKE | [Link](./training/a4/paligemma2/README.md)
60+
**DeepSeek-V3** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | Megatron-Bridge (25.11) | Pre-training | GKE | [Link](./training/a4/deepseek_v3/megatron-bridge-gke/nemo2511/)
61+
**DeepSeek-V3** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | Megatron-Bridge (26.02) | Pre-training | GKE | [Link](./training/a4/deepseek_v3/megatron-bridge-gke/nemo2602/)
62+
**GPT OSS 120B** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | Megatron-Bridge (26.02) | Pre-training | GKE | [Link](./training/a4/gpt_oss_120b/megatron-bridge-gke/nemo2602/)
63+
**Llama-3-8B** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | Megatron-Bridge (26.02) | Pre-training | GKE | [Link](./training/a4/llama3-8b/megatron-bridge-gke/nemo2602/)
64+
**Qwen-3-235B** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | Megatron-Bridge (25.11) | Pre-training | GKE | [Link](./training/a4/qwen3_235b_a22b/megatron-bridge-gke/nemo2511/)
65+
**Qwen-3-235B** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | Megatron-Bridge (26.02) | Pre-training | GKE | [Link](./training/a4/qwen3_235b_a22b/megatron-bridge-gke/nemo2602/)
66+
**Qwen-3-235B** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | Megatron-Bridge (25.11) | Pre-training | Slurm | [Link](./training/a4/qwen3_235b_a22b/megatron-bridge-slurm/nemo2511/)
67+
**Qwen-3-30B** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | NeMo (26.02) | Pre-training | GKE | [Link](./training/a4/qwen3_30b_a3b/nemo-gke/nemo2602/)
68+
**Wan-2.1-14B** | [A4 (NVIDIA B200)](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a4-vms) | NeMo (25.11) | Pre-training | GKE | [Link](./training/a4/wan_14b/nemo-gke/nemo2511/)
69+
4970

5071
### Training benchmarks A4X
5172

0 commit comments

Comments
 (0)