TEST #90 RESULTS
03/28/2025 - 9:24 PM
ACCELERATOR
MODEL
51.3
tokens/s
generation
252
ms
time to first token
5487
tokens/s
prompt
1038
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD EPYC 9354 32-Core Processor (znver4)
RAM
755.5GB
OS
Linux
Kernel Release
6.8.0-52-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #53~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Wed Jan 15 19:18:46 UTC 2
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
7577
tokens/s
54.4
tokens/s
155
ms
pp4096+tg256
4771
tokens/s
44.2
tokens/s
885
ms
pp2048+tg256
6200
tokens/s
49.9
tokens/s
351
ms
pp2048+tg768
6183
tokens/s
48.2
tokens/s
352
ms
pp1024+tg1024
7637
tokens/s
51.5
tokens/s
155
ms
pp1280+tg3072
6829
tokens/s
47.3
tokens/s
208
ms
pp384+tg1152
7398
tokens/s
54.4
tokens/s
70
ms
pp64+tg1024
2125
tokens/s
56.6
tokens/s
47
ms
pp16+tg1536
667
tokens/s
55.2
tokens/s
41
ms