This is just for my app only. All build times are with --production set true. RSPack tests were done via Meteor 3.4 beta 11.
the reduction of client bundle size is a game changer.
Hey @vikr00001!
Great to see your changes, especially the bundle size improvements.
Regarding HMR and builds, how did you gather the data? Did you use meteor profile on the app in both Meteor 3.3 and 3.4-beta.12? It’s odd I don’t see gains in HMR client. Did you revise the numbers?
I recently adapted a real-world app showing:
Client bundle size
Client bundle size dropped from 8.56 MB to 2.51 MB (~70.68% dropped), with a few optimizations still pending.
These metrics are easy to read in your Galaxy dashboard or in the browser.
Build times
For build times, see the attached meteor profile example for that app:
| Scenario | Original (ms) | Improved (ms) | % Reduction | Times Faster |
|---|---|---|---|---|
| Cold start | 26 486 | 8 304 | 68.64% | 3.19× |
| Cache start | 15 562 | 5 970 | 61.64% | 2.61× |
| Rebuild client #1 | 1 764 | 431 | 75.57% | 4.09× |
| Rebuild client #2 | 1 704 | 431 | 74.71% | 3.95× |
| Rebuild server #1 | 7 561 | 2 390 | 68.38% | 3.16× |
| Rebuild server #2 | 6 900 | 1 846 | 73.24% | 3.74× |
Meteor 3.4-beta.12 for this app shows around 70% faster build and rebuild times, including both client and server HMR/rebuild. I applied other optimizations I will talk about in the future for some phases, but gains should show just by adopting Rspack in all phases.
I’ve seen this pattern in many apps I’ve migrated, so I’m curious how you measured the client rebuild data. It should show gains as now Rspack dev server and HMR strategy takes place.
I had no idea those stats were available on Galaxy!
For client change HMR I just used a stopwatch and waited for changes on the client.
For server change updates while the client is live, I used a stopwatch and waited for the console log saying “App is now running on http://localhost:3000/”,
I got the numbers for the client bundle from bundle-visualizer.
Amazing improvements!


