FYI - potential blob reading improvements, 10x in some cases

3 views

Skip to first unread message

Dan Murphy

unread,

May 29, 2026, 3:49:47 PM (14 days ago) May 29

to storage-dev

Hey all,

I prototyped a CL to see if keeping blob data in the renderer for about 5 seconds and performing an initial mojo bind to it could increase read times. It is very unsophisticated and lacks more things like cross-thread support, but it resulted in pretty big speedups for file-backed blob reading and for larger blob sizes:

https://chromium-review.git.corp.google.com/c/chromium/src/+/7808030

I think one of those categories was a 10x improvement.

Anyways - given that I thought it might be worth looking into by more official storage folks. or not, if blob reading isn't a priority right now. but I wanted to give an FYI in case others were interested!