Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Matrix Multiplication Latency on Llama-3 8B
Loading...
152.69
Kernel-level latency (µs)
CodeGEMM
132.9772
266.0386
399.1
532.1614
Dec 19, 2025
Kernel-level latency (µs)
Updated 4d ago
Evaluation Results
Method
Method
Links
Kernel-level latency (µs)
CodeGEMM
config=m1v4g128
2025.12
152.69
LUTGEMM
quantization=q2-g128
2025.12
160.1
QuIP#
config=e8p
2025.12
162.63
CodeGEMM
config=m2v8g128
2025.12
172.18
QTIP
config=r2
2025.12
189.94
AQLM
config=2x8
2025.12
250.12
cuBLAS
precision=fp16
2025.12
332.45
AQLM
config=1x16
2025.12
645.51
Feedback
Search any
task
Search any
task
