Forum moved here!

Home / Sumatra takes huge amount of RAM

britant

Hi,

I always thought that Sumatra PDF is a lightweight PDF viewer but recently I checked Task Manager and I found out that Sumatra takes around 250 MB of RAM – is it normal?

The file is not particularly big (32,3MB | 73 pages), In my opinion that’s very weird because programs like Adobe Acrobat Readers takes far less RAM than this.

I use Windows 10 Home and Sumatra 3.2 (stable version)

You can download that PDF file here:

GitHubRulesOK

It is not unusual for memory to escalate from an initial low value on opening first page as a combined 10-20MB up to 500 MB as more pages are decompressed since a compressed file itself could be several GB in decompressed storage value.

Each version may behave differently and 3.2 tends to use more memory for display compared to current release for that same file.
Memory consumtion is not a good measure of other functions such as optimising buffered text or images
image

britant

Ok, thank’s for answering my question. I wasn’t aware of that fact.

GitHubRulesOK

Out of interest I ran decompression on that file using MuPDF (which is the engine SumatraPDF uses) it ran at about 300MB to not stress memory but the output was 1.53 GB for all pages but that naturally does not equate to combined pages at screen resolution just the memory needed to hold all objects in one gulp.

It would be about right whichever way its decompressed to 96ppi that each image page in 24bit colour would be about 20MB thus 75 pages roughly = 1.5 GB one reason not to compress them in the first case just keep resolution low enough to read clearly and quickly