High CPU usage on RDS 2019

[Wizeman26]

Hi,

I have an issue with Firefox 91.8.0 ESR (64 bits) on RDS (Windows 2019 Std) ; Firefox uses a lot of CPU, each time a user start Firefox, CPU take off.
It's a fresh install (server is up to date) and Firefox profiles are new (no synchronization, no plugins ...).

I've already get this issue with RDS 2016 Std, not solve and went to Chrome ...

I set a GPO to disable hardware acceleration but I haven't seen any improvements.

I didn't get this issue on Windows 2008 R2 with older Firefox versions.

Have you got an idea to improve Firefox performances ?

Thanks.

[Wizeman26]

Hi,

I've tried today to install FF ESR 32 Bits on a fresh 2019 server and I've the same issue.
On my comuter if I start Edge (100.0.1185.44 64 bits), Chrome (100.0.4896.127 64 bits) and Firefox (99.0.1 64 bits), I can see that FF consumes twice more CPU and memory than Chrome and Edge.

Does it a normal behaviour of FF or it's possible to imporve CPU and memory consumption ?
Hi,

I've tried today to install FF ESR 32 Bits on a fresh 2019 server and I've the same issue.
On my comuter (Win 11 21H2) if I start Edge (100.0.1185.44 64 bits), Chrome (100.0.4896.127 64 bits) and Firefox (99.0.1 64 bits), I can see that FF consumes twice more CPU and memory than Chrome and Edge.

Does it a normal behaviour of FF or it's possible to imporve CPU and memory consumption ?

Thanks,

[xanthon]

I doubt that it's 'normal' but I see reports and have the issue myself with RAM. I have run out of ideas for my issue. There is an article which may help. https://support.mozilla.org/en-US/kb/fi ... -resources

[therube]

I can see that FF consumes twice more CPU and memory than Chrome and Edge.

Care to explain that further?

[Wizeman26]

For example, with a fresh Windows 10 installation with Edge Version 101.0.1210.47 (Official Build) (64-bit), Chrome Version 101.0.4951.67 (Official Build) (64-bit) and Firefox 91.9.0esr (64-bit)
I set up https://www.google.fr as startup page and I start the 3 browers, after 1 minutes here is my task manager :

https://i.ibb.co/dDPpQkL/Image2.png

Firefox CPU usage = 42 %
Chrome CPU usage = 22,7%
Edge CPU usage = 8%

I understand that these values are updated regularly and sometimes Edge or Chrome consumes more CPU and RAM but generally I can see that FF consumes more CPU and RAM than Chrome and Edge.
Observed on Windows 10, Windows 11, Server 2016, Server 2019

[Wizeman26]

After 30 minutes (with no interactions), CPU usage is ~0% but RAM usage is :
- 178 Mo for FF
- 163 Mo for Edge
- 90,9 Mo for Chrome

[therube]

RDS (Windows 2019 Std)

I don't know "RDS"...

Is FF itself installed on the "server", & "users" are running FF from the server (rather then locally on their desktops), across a network, kind of thing?

I can see that FF consumes twice more CPU and memory than Chrome and Edge.

Take that with a grain of salt.

Windows 10 ... I set up https://www.google.fr as startup page and I start the 3 browers, after 1 minutes here is my task manager

Firefox CPU usage = 42 %
Chrome CPU usage = 22,7%
Edge CPU usage = 8%

Take that with a grain of salt.
That aside, what on ... green earth, how can FF, Chrome, or even Edge, be using that kind of CPU usage - after 1 minute - having loaded a relatively "simple" web page?

Something... has to be going on. Antivirus perhaps?

If instead of google.fr, I you have the browsers open to a blank page (about:blank), do you still see similar CPU usages?

After 30 minutes (with no interactions), CPU usage is ~0% but RAM usage is :

Take that (the RAM usage) with a grain of salt.
It gets confusing, it confuses me, but... What is private bytes, virtual bytes, working set?

[therube]

a grain of salt?

Defining how much CPU or RAM a program is using might not be straight forward.
So a blanket statement that Chrome uses less "CPU" or "RAM" then FF may not necessarily be a meaningful statement.

Open Chrome, open FF, set up in similar manners, load the same web page, you would expect each to be using approximately the same amounts of CPU & RAM.

Now certainly there will be instances where one performs better then the other, but overall, you would expect them to be close in behavior.

16 GB should be more then fine. I can't really fathom a browser "running" if it has eaten 16 GB of RAM. I'd expect any browser to get squirrelly after having eating 4 GB of RAM.


(I'm speaking of Windows. Linux... could be different?)