Nvidia/libcairo2 memory use

[dbVU]

All,

After the last update and even though I had locked libcairo2 at the stable version,(Most likely removed Debian stable from repository list.) (http://forums.solydxk.nl/viewtopic.php?f=7&t=1916 my at desktop ram use had jumped to nearly 700MB. I checked which version of libcario2 I was running and it was reset to the testing version. Rather than downgrade again which wanted to remove Gimp and much more I changed from the Nvidia driver 304.xx to Nouveau and my ram usage dropped a whopping 550MB to 159MB at the desktop. I notice no difference in the video performance. This could be a huge deal for for a low power machine. Edit 11-21-13 still happy with Nouveau.
I used smxi to make the video driver change and it worked perfectly. This has not been true in the past as I was unable to revert to Nouveau. I'm not sure why it went so smoothly this time so be careful.

db

[drc]

For me, reverting to nouveau (or better yet, not installing the nvidia drivers) is really not an option. I really have no need for the 3D aspect of nvidia, but thru weeks of testing, I find that on my laptop with a nVidia G86 [GeForce 8400M GS] the gpu runs in excess of 10C hotter, while the cpu runs about 5C hotter (I'm assuming it's heat bleed over from the gpu, but maybe not) with the nouveau driver. To be fair, it runs perfectly, just significantly hotter.

Here's hoping this problem is resolved soon (and I don't mean every individual having to personally revert to an earlier version of libcairo2, that's not really a "fix").

[dbVU]

Interesting, My card is a NVIDIA NV44A [GeForce 6200] and is running at 43C which seems normal. I agree with your statement about not installing Nvidia at all at this time.

[drc]

OK, it's been almost a month now, time to ask the question "Is this (nvidia/libcairo2/memory-use) still a problem for some folks? Or has Schoelje, in his copious free time, with nothing else to do, found the answer?

[dbVU]

I don't want to try using the Nonfree nvidia drivers. It was too hard to get everthing sorted out the last time. If anyone else is willing, it would be nice to know if this is still a problem. I use Glances from the cl to monitor memorey use.

[drc]

If anyone else is willing, it would be nice to know if this is still a problem.

I swear there used to be somewhat of an answer (here) to my (above) question, stating that "yes it was still a problem", but it appears to have disappeared into the "nether-net".

[zerozero]

there was and it said that the problem was still present in the current libcairo2 version. the user chose to delete it.

[dbVU]

Oooh wish I would have checked here before doing a test reinstalling Nvidia driver 304.117, with smxi by the way DDM just didn't work. Anyhow I can confirm the memory bug is still present. I went from ~200mb with nouveau to over 600mb with Nvidia. As a point of interest smxi failed removing nouveau had to sgfxi to reinstall after a reboot. Seems to be a bit of a mess right now.

db

[drc]

Anyhow I can confirm the memory bug is still present. I went from ~200mb with nouveau to over 600mb with Nvidia.

Kinda like having a 64 Volkswagon that got 35 mpg, but deciding that you wanted a bit more bling, so you decided to buy a Lexis that got 25 mpg. Then, after you took it in for your annual maintenance check, you found that it now got 15 mpg, and the dealer said "oh yeah, that's an parts-manifacturer/upstream problem, we can't do anything about it".

Makes one want to stick with one's 64 Volkswagon (which one did), but one is still looking longingly at the Lexis, wonder when they're going to get things fixed?

[Chris M]

Sorry to resurrect this if this was fixed, but I recently installed an Nvidia card, and I'm getting the same type of memory jump issue. The libcairo2 versions in Testing and SID are the same at version 1.12.16-2

When I look at the libcairo2 Source Package, the date stamps/last modified dates are 09/17/2013.

What's strange is that the jump was pronounced right after installation with sgfxi, but then it seemed to settle down after subsequent reboots. I think it settled down recently after an upgrade of xserver from about 420mb to roughly 220mb. Then a kernel update landed me at the cli, so I reinstalled the Nvidia driver with sgfxi again, but it won't settle down. I do feel like something is still off, performance-wise.

Is anyone else still experiencing the same issue?

420mb at start up is not astronomical. If I hadn't experienced 200mb at start up with Nvidia working, I may not have thought too much about it.

I'm running a very lean XFCE Debian testing 64 bit. For those using the Nvidia driver, what's your start up RAM usage look like?

Thanks

[ScottQuier]

I don't think I've ever seen start up ram usage below about 440mb

[zerozero]

honestly i lost track of this topic (so thanks Chris for bringing it back )

not a direct comparison but

Code: Select all

zerozero@x-test:~$ inxi -Fzx
System:    Host: x-test Kernel: 3.14-1-amd64 x86_64 (64 bit gcc: 4.8.3) 
           Desktop: N/A Distro: SolydXK 1 testing 
Machine:   System: LENOVO product: Lenovo G585 v: Lenovo G585
           Mobo: LENOVO model: Lenovo G585 v: 31900003WIN8 STD MLT
           Bios: LENOVO v: 6CCN88WW(V8.01) date: 08/22/2012
CPU:       Dual core AMD E2-1800 APU with Radeon HD Graphics (-MCP-) cache: 1024 KB
           flags: (lm nx sse sse2 sse3 sse4a ssse3 svm) bmips: 6787 
           Clock Speeds: 1: 850 MHz 2: 1360 MHz
Graphics:  Card: Advanced Micro Devices [AMD/ATI] Wrestler [Radeon HD 7340]
           bus-ID: 00:01.0
           Display Server: X.Org 1.15.1 driver: fglrx
           Resolution: 1366x768@60.0hz
           GLX Renderer: AMD Radeon HD 7340 Graphics
           GLX Version: 4.4.12874 - CPC 14.10.1006.1001 Direct Rendering: Yes
Audio:     Card-1 Advanced Micro Devices [AMD] FCH Azalia Controller 
           driver: snd_hda_intel bus-ID: 00:14.2 
           Card-2 Advanced Micro Devices [AMD/ATI] Wrestler HDMI Audio 
           driver: snd_hda_intel bus-ID: 00:01.1 
           Sound: Advanced Linux Sound Architecture v: k3.14-1-amd64
Network:   Card-1: Qualcomm Atheros AR9285 Wireless Network Adapter (PCI-Express)
           driver: ath9k bus-ID: 06:00.0
           IF: wlan0 state: up mac: <filter>
           Card-2: Realtek RTL8101E/RTL8102E PCI Express Fast Ethernet controller
           driver: r8169 v: 2.3LK-NAPI port: 2000 bus-ID: 02:00.0
           IF: eth0 state: down mac: <filter>
Drives:    HDD Total Size: 320.1GB (3.1% used)
           ID-1: /dev/sda model: WDC_WD3200BPVT size: 320.1GB
Partition: ID-1: / size: 54G used: 5.6G (11%) fs: ext4 dev: /dev/sda6 
           ID-2: swap-1 size: 4.19GB used: 0.00GB (0%) fs: swap dev: /dev/sda1 
Sensors:   System Temperatures: cpu: 42.8C mobo: N/A gpu: 43.00C 
           Fan Speeds (in rpm): cpu: N/A 
Info:      Processes: 161 Uptime: 28 min Memory: 323.1/3534.9MB 
           Init: SysVinit runlevel: 2 Gcc sys: 4.9.0 
           Client: Shell (bash 4.3.111) inxi: 2.1.28 
zerozero@x-test:~$ 

is stable at 317MB

Code: Select all

zerozero@x-test:~$ free -ms 30
             total       used       free     shared    buffers     cached
Mem:          3534        845       2689          2         33        493
-/+ buffers/cache:        317       3217
Swap:         3998          0       3998

             total       used       free     shared    buffers     cached
Mem:          3534        845       2689          2         33        493
-/+ buffers/cache:        317       3217
Swap:         3998          0       3998

             total       used       free     shared    buffers     cached
Mem:          3534        845       2689          2         33        493
-/+ buffers/cache:        317       3217
Swap:         3998          0       3998

             total       used       free     shared    buffers     cached
Mem:          3534        845       2689          2         33        493
-/+ buffers/cache:        317       3217
Swap:         3998          0       3998

             total       used       free     shared    buffers     cached
Mem:          3534        845       2689          2         33        493
-/+ buffers/cache:        317       3217
Swap:         3998          0       3998

             total       used       free     shared    buffers     cached
Mem:          3534        845       2689          2         33        493
-/+ buffers/cache:        317       3216
Swap:         3998          0       3998

             total       used       free     shared    buffers     cached
Mem:          3534        845       2689          2         33        493
-/+ buffers/cache:        318       3216
Swap:         3998          0       3998

             total       used       free     shared    buffers     cached
Mem:          3534        845       2689          2         34        493
-/+ buffers/cache:        317       3217
Swap:         3998          0       3998

             total       used       free     shared    buffers     cached
Mem:          3534        845       2689          2         34        493
-/+ buffers/cache:        317       3217
Swap:         3998          0       3998

[Chris M]

And just like that, it settled back down to 220mb at start up. My last 2 apt-get update && apt-get upgrade && apt-get dist-upgrade :

Code: Select all

Start-Date: 2014-08-18  07:41:53
Commandline: apt-get upgrade
Upgrade: libio-socket-ssl-perl:amd64 (1.992-1, 1.997-2), iucode-tool:amd64 (1.0.2-1, 1.0.3-1), python-pkg-resources:amd64 (5.4.1-1, 5.5.1-1), python3-pkg-resources:amd64 (5.4.1-1, 5.5.1-1)
End-Date: 2014-08-18  07:42:00

Start-Date: 2014-08-19  08:16:36
Commandline: apt-get upgrade
Upgrade: libfontembed1:amd64 (1.0.55-1, 1.0.57-1), isc-dhcp-common:amd64 (4.3.0+dfsg-2, 4.3.1-1), python-numpy:amd64 (1.8.1-1+b1, 1.8.2-1), console-setup:amd64 (1.102, 1.111), libfontconfig1:amd64 (2.11.0-5, 2.11.0-6), console-setup-linux:amd64 (1.102, 1.111), fontconfig:amd64 (2.11.0-5, 2.11.0-6), fontconfig-config:amd64 (2.11.0-5, 2.11.0-6), libv4lconvert0:amd64 (1.0.1-1, 1.2.1-2), cups-browsed:amd64 (1.0.55-1, 1.0.57-1), libv4l-0:amd64 (1.0.1-1, 1.2.1-2), cups-filters-core-drivers:amd64 (1.0.55-1, 1.0.57-1), isc-dhcp-client:amd64 (4.3.0+dfsg-2, 4.3.1-1), cups-filters:amd64 (1.0.55-1, 1.0.57-1), keyboard-configuration:amd64 (1.102, 1.111), libcupsfilters1:amd64 (1.0.55-1, 1.0.57-1)
End-Date: 2014-08-19  08:16:59 

I guess it's something within the 2014-08-19 upgrades. There have been a lot of cups upgrades of late. Go figure.

No question that my install is snappier. A real head scratcher. "Self-healing linux"