Llama 2 7B on the AMD Radeon 880M

Best measured: 16.2 tokens/s generation and 213 tokens/s prompt processing (Q4_0, ROCm, FA on).

Every run

Quant	Backend	Generate tok/s	Prompt tok/s	Test	Build / by / notes	Source
Q4_0	ROCm FA on	16.2	213	tg128	c55d53a @Hedede	llama.cpp ROCm scoreboard (#15021)
Q4_0	ROCm FA off	13.0	163	tg128	c55d53a @Hedede	llama.cpp ROCm scoreboard (#15021)

tg128 = generating 128 tokens; pp512 = reading a 512-token prompt (llama-bench). LocalScore averages several prompt and output lengths, so its numbers run lower than tg128.

Check prices

Amazon Newegg

As an Amazon Associate I earn from qualifying purchases.

No card yet?

Try it on a rented GPU (RunPod)

Affiliate link: if you sign up, RigBench may earn a referral credit. It never affects rankings or numbers.

All results for the AMD Radeon 880M · Llama 2 7B on other hardware