masaori335 commented on PR #13248:
URL: https://github.com/apache/trafficserver/pull/13248#issuecomment-4655047058
Our benchmark says below.
| Scenario | no last-open-read | 10.2.0 (vanilla) | Δ |
|---|---|---|---|
| max-tps-10url (plain text) | 441,754 req/s | 564,269 req/s | −122,515
req/s (−21.7%) |
| max-tps-10url-tls | 505,910 req/s | 505,717 req/s | +193 req/s
(+0.0%) |
| max-tps-10url-h2 | 463,392 req/s | 465,356 req/s | −1,964 req/s
(−0.4%) |
| max-tps-100url (plain text) | 514,221 req/s | 578,793 req/s | −64,572
req/s (−11.2%) |
| max-tps-100url-tls | 517,851 req/s | 524,341 req/s | −6,490 req/s
(−1.2%) |
| max-tps-100url-h2 | 481,452 req/s | 487,708 req/s | −6,256 req/s
(−1.3%) |
Hardware & Config Info
- CPU: Intel(R) Xeon(R) Gold 5218 64 core
- 40 Stripes = 8 (NVME SSD Drive) x 5 (Volumes)
- Thread configs in records.yaml
```
exec_thread:
affinity: 4
limit: 63
listen: 1
```
The mini-cache (last-open-read) is working even if target URL is 100. Now,
I'm wondering we should fix the small RAM cache promotion bug and keep this
mini cache.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
