Llama on NVIDIA Ampere

Can I run Llama 3.2 1B on an RTX 3090?

Yes. Llama 3.2 1B at Q4_K_M uses 1.8 GB of the 21.8 GB available on a RTX 3090, and runs at about 650 tokens per second. You can push the context to 128K.

Runs comfortably

1.8 GB of 21.8 GB · 8%
022 GB
Weights 0.7 GB
KV cache 0.3 GB
Runtime overhead 0.8 GB

Llama 3.2 1B at Q4_K_M leaves 20.0 GB spare on a RTX 3090. There is room to raise the context length or move up a quantisation level.

Generation650tok/s
Prompt processing12024tok/s
Max context128Ktokens
KV per 1K tokens0GB

Every quantisation of Llama 3.2 1B on a RTX 3090

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 235
INT8 / W8A8 8.50 1.2 GB 2.2 GB Runs comfortably 128K 434
Q8_0 (GGUF) 8.50 1.2 GB 2.2 GB Runs comfortably 128K 434
FP8 (E4M3) 8.00 1.2 GB 2.2 GB Runs comfortably 128K 439
AWQ 4-bit 4.25 1.0 GB 2.0 GB Runs comfortably 128K 508
GPTQ 4-bit 4.25 1.0 GB 2.0 GB Runs comfortably 128K 508
MXFP4 4.25 1.0 GB 2.0 GB Runs comfortably 128K 508
Q6_K 6.56 0.9 GB 2.0 GB Runs comfortably 128K 519
Q5_K_M 5.67 0.8 GB 1.9 GB Runs comfortably 128K 586
Q5_K_S 5.52 0.8 GB 1.8 GB Runs comfortably 128K 599
Q4_K_M 4.85 0.7 GB 1.8 GB Runs comfortably 128K 650
Q4_K_S 4.58 0.7 GB 1.7 GB Runs comfortably 128K 673
Q4_0 4.55 0.7 GB 1.7 GB Runs comfortably 128K 676
IQ4_XS 4.25 0.7 GB 1.7 GB Runs comfortably 128K 703
Q3_K_M 3.91 0.6 GB 1.6 GB Runs comfortably 128K 738
IQ3_M 3.70 0.6 GB 1.6 GB Runs comfortably 128K 760
IQ3_XXS 3.06 0.5 GB 1.6 GB Runs comfortably 128K 840
Q2_K 2.63 0.5 GB 1.5 GB Runs comfortably 128K 903
IQ2_XXS 2.06 0.4 GB 1.4 GB Runs comfortably 128K 1004
IQ1_M 1.75 0.4 GB 1.4 GB Runs comfortably 128K 1068

Also worth checking

Llama 3.2 1B on every GPU
Full hardware tableEverything that runs on a RTX 3090
Full model tableWhy the KV cache is the part that bites
Guide