TEST #45 RESULTS
03/28/2025 - 6:58 PM
ACCELERATOR
MODEL
60.6
tokens/s
generation
656
ms
time to first token
2011
tokens/s
prompt
570
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD EPYC 7702 64-Core Processor (znver2)
RAM
503.5GB
OS
Linux
Kernel Release
6.8.0-40-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #40~22.04.3-Ubuntu SMP PREEMPT_DYNAMIC Tue Jul 30 17:30:19 UTC 2
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
2342
tokens/s
63.3
tokens/s
453
ms
pp4096+tg256
1866
tokens/s
55.5
tokens/s
2.21
sec
pp2048+tg256
2150
tokens/s
60.1
tokens/s
969
ms
pp2048+tg768
2146
tokens/s
59.3
tokens/s
971
ms
pp1024+tg1024
2329
tokens/s
60.7
tokens/s
456
ms
pp1280+tg3072
2253
tokens/s
57.6
tokens/s
585
ms
pp384+tg1152
2515
tokens/s
62.1
tokens/s
169
ms
pp64+tg1024
1982
tokens/s
63.6
tokens/s
47
ms
pp16+tg1536
518
tokens/s
63.1
tokens/s
46
ms