Llama on NVIDIA Hopper

Can I run Llama 3.2 1B on an H100 SXM 80GB?

Yes. Llama 3.2 1B at Q4_K_M uses 1.8 GB of the 74.4 GB available on a H100 SXM 80GB, and runs at about 2192 tokens per second. You can push the context to 128K.

Runs comfortably

1.8 GB of 74.4 GB · 2%
074 GB
Weights 0.7 GB
KV cache 0.3 GB
Runtime overhead 0.8 GB

Llama 3.2 1B at Q4_K_M leaves 72.6 GB spare on a H100 SXM 80GB. There is room to raise the context length or move up a quantisation level.

Generation2192tok/s
Prompt processing167492tok/s
Max context128Ktokens
KV per 1K tokens0GB

Every quantisation of Llama 3.2 1B on a H100 SXM 80GB

Highlighted row is the highest quality that still fits at 8K context.

QuantisationbpwWeights TotalVerdictMax ctxtok/s
FP16 / BF16 16.00 2.3 GB 3.3 GB Runs comfortably 128K 882
INT8 / W8A8 8.50 1.2 GB 2.2 GB Runs comfortably 128K 1546
Q8_0 (GGUF) 8.50 1.2 GB 2.2 GB Runs comfortably 128K 1546
FP8 (E4M3) 8.00 1.2 GB 2.2 GB Runs comfortably 128K 1559
AWQ 4-bit 4.25 1.0 GB 2.0 GB Runs comfortably 128K 1773
GPTQ 4-bit 4.25 1.0 GB 2.0 GB Runs comfortably 128K 1773
MXFP4 4.25 1.0 GB 2.0 GB Runs comfortably 128K 1773
Q6_K 6.56 0.9 GB 2.0 GB Runs comfortably 128K 1809
Q5_K_M 5.67 0.8 GB 1.9 GB Runs comfortably 128K 2007
Q5_K_S 5.52 0.8 GB 1.8 GB Runs comfortably 128K 2045
Q4_K_M 4.85 0.7 GB 1.8 GB Runs comfortably 128K 2192
Q4_K_S 4.58 0.7 GB 1.7 GB Runs comfortably 128K 2257
Q4_0 4.55 0.7 GB 1.7 GB Runs comfortably 128K 2264
IQ4_XS 4.25 0.7 GB 1.7 GB Runs comfortably 128K 2342
Q3_K_M 3.91 0.6 GB 1.6 GB Runs comfortably 128K 2436
IQ3_M 3.70 0.6 GB 1.6 GB Runs comfortably 128K 2498
IQ3_XXS 3.06 0.5 GB 1.6 GB Runs comfortably 128K 2707
Q2_K 2.63 0.5 GB 1.5 GB Runs comfortably 128K 2869
IQ2_XXS 2.06 0.4 GB 1.4 GB Runs comfortably 128K 3117
IQ1_M 1.75 0.4 GB 1.4 GB Runs comfortably 128K 3270

Also worth checking

Llama 3.2 1B on every GPU
Full hardware tableEverything that runs on a H100 SXM 80GB
Full model tableWhy the KV cache is the part that bites
Guide