You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: docs/performance.md
+34-34Lines changed: 34 additions & 34 deletions
Original file line number
Diff line number
Diff line change
@@ -11,15 +11,15 @@ The performance tests were conducted in a controlled environment with determinis
11
11
12
12
### GCP
13
13
14
-
The indicative figures below were measured using the [CT hammer tool](/internal/hammer/) as of [commit `2872ea2`](https://github.com/transparency-dev/tesseract/commit/2872ea2387b2d3077eb832112277eb19a7a907bd).
14
+
The indicative figures below were measured using the [CT hammer tool](/internal/hammer/) as of [commit `fe7687c`](https://github.com/transparency-dev/tesseract/commit/fe7687c9ed35d11f42a211ee35544ff6c5610ee6).
15
15
16
16
The table below shows the measured performance over 12 hours in each instance type:
The write QPS is around 150. The bottleneck is Cloud Spanner CPU. Utilization is always around 100%, most of which is used for deduplication. The VM CPU usage is around 50%.
62
+
The write QPS is around 250. The bottleneck comes from the VM CPU utilization which is always around 100%. The Cloud Spanner CPU utilization is around 20%.
#### e2-medium VM Instance + Cloud Spanner 200 PUs
95
+
#### e2-standard-2 VM Instance + Cloud Spanner 100 PUs
96
96
97
-
- e2-medium (2 vCPUs, 4 GB Memory)
97
+
- e2-standard-2 (2 vCPUs, 8 GB Memory)
98
98
99
-
The write QPS is around 350. The bottleneck comes from the VM CPU usage which is always around 90%. The Cloud Spanner CPU utilization is around 50%.
99
+
The write QPS is around 600. The bottleneck comes from the VM CPU utilization which is always around 100%. The Cloud Spanner CPU utilization is around 50%.
0 commit comments