DEV Community

#benchmarking

Posts

đź‘‹ Sign in for the ability to sort posts by relevant, latest, or top.
Warm vs Cold: Why a Single Trial Misleads Performance Claims

Warm vs Cold: Why a Single Trial Misleads Performance Claims

1
Comments
4 min read
Why a Single ./a.out Time Misleads Your Performance Claims

Why a Single ./a.out Time Misleads Your Performance Claims

3
Comments
2 min read
Escalating to the better model made 34 answers worse

Escalating to the better model made 34 answers worse

3
Comments 6
12 min read
The Benchmark That's Half Traps — and Why That's Brilliant

The Benchmark That's Half Traps — and Why That's Brilliant

8
Comments
9 min read
A 4 GB Laptop GPU vs a 6-Core CPU on Gemma 4, Re-Measured in ABBA Order: 4.1x

A 4 GB Laptop GPU vs a 6-Core CPU on Gemma 4, Re-Measured in ABBA Order: 4.1x

9
Comments 2
11 min read
Counting bugs is the hard part of comparing AI review tools

Counting bugs is the hard part of comparing AI review tools

Comments
3 min read
When to use which: Dragonfly vs Redis vs Valkey

When to use which: Dragonfly vs Redis vs Valkey

1
Comments
4 min read
Our AI agents' "verified success" claims: 10 out of 10 failed independent recompute — including ours

Our AI agents' "verified success" claims: 10 out of 10 failed independent recompute — including ours

1
Comments
3 min read
Operational simplicity: the cross-key tax, quantified

Operational simplicity: the cross-key tax, quantified

Comments 1
3 min read
Memory efficiency: bytes per key

Memory efficiency: bytes per key

Comments 1
2 min read
Apple M2 Compute Performance: Benchmarking AMX, GPU, and Neural Engine

Apple M2 Compute Performance: Benchmarking AMX, GPU, and Neural Engine

Comments
6 min read
Latency under load: the same story as throughput, from the other side

Latency under load: the same story as throughput, from the other side

Comments 3
3 min read
Qwen 3.8 4-bit Benchmark RTX 4090: 1-bit is a Trap

Qwen 3.8 4-bit Benchmark RTX 4090: 1-bit is a Trap

Comments
7 min read
The sleep loop is the tell: agents that pay per action optimize to do nothing

The sleep loop is the tell: agents that pay per action optimize to do nothing

Comments 1
2 min read
Why a normal client gets 4.6M ops/s out of a Redis cluster that can do 40M

Why a normal client gets 4.6M ops/s out of a Redis cluster that can do 40M

Comments 3
5 min read
đź‘‹ Sign in for the ability to sort posts by relevant, latest, or top.