Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Top-K Identification on 1M 128-dimensional vectors (99% recall)
Loading...
6.55
Matmul Latency (ms)
Generalized Two-Stage Approximate Top-K
6.4496
7.1273
7.805
8.4827
Jun 4, 2025
Matmul Latency (ms)
Stage 1 Latency (ms)
Stage 2 Latency (ms)
Total Latency (ms)
Updated 20d ago
Evaluation Results
Method
Method
Links
Matmul Latency (ms)
Stage 1 Latency (ms)
Stage 2 Latency (ms)
Total Latency (ms)
Generalized Two-Stage Approximate Top-K
K'=4, Fusion=fused
2025.06
6.55
-
3.51
10
Generalized Two-Stage Approximate Top-K
K'=4, Fusion=unfused
2025.06
7.31
10.8
3.51
22
jax.lax.top_k
implementation=exact T...
2025.06
7.32
-
587
594
Generalized Two-Stage Approximate Top-K
K'=1, Fusion=unfused
2025.06
7.32
6.58
50
64
Generalized Two-Stage Approximate Top-K
K'=1, Fusion=fused
2025.06
9.03
-
50
59
jax.lax.approx_max_k
K'=1, Fusion=fused
2025.06
9.06
-
118
127
Feedback
Search any
task
Search any
task