Really poor performance under Ubuntu 12.10

[AbeFM]

So, I'm back to where I was in the beginning of the thread: My computer is really, really slow - I have no idea why, but it's sudden, unusable, and related only to a change in Handbrake and a change in the OS, and I'm wondering if:
1) There's anything I can do to fix it
2) If it's a problem with HB and one I can help to solve by providing information or testing.

All I know is it's unusable the way it is now.

[lupus9537]

I have a very similar issue. I recently built a server with following specs:
- Intel i5 CPU with 2 cores, 4 threads
- 4 Gb of RAM
- SSD system drive
- RAID 5 data disk configuration with 4x 7200 RPM 2 tB drives
The system itself runs very smooth. streaming, downloading, parity checks, etc all works very well!

However, if I use the handbrake-cli to transcode a movie, it works on average at 4 fps. It takes 6 hours to transcode an MKV episode of 40 minutes to an MP4 for iPad.
I use following command line:
handbrake-cli -i file.mkv -o file.mp4 --preset "iPad"

If I process the files over a network share and transcode the very same file from my windows PC (which has slightly lower specs), the process runs at avg 17 fps... It only takes about 50 minutes...

[AbeFM]

Huh, well, for what it's worth, my desktop is much faster (top of the line i5) on windows, perhaps I could learn from you and do it all over the network till some progress is made on this issue.

What exact OS are you running? Which handbrake? Had you run that set up before and had it be fast (sounds like no) - either way, it's obviously way, way too slow.

[JohnAStebbins]

There's nothing more I can do until I get enough information to be able to reproduce the problem. I've tried every combination I have available to me. Ubuntu 12.10, Fedora 17, 64bit, 32bit, laptop sandybridge i7, desktop sandybridge i7. The only thing I can think of is some really bad driver interaction. Can each of you give me a rundown of other hardware. I need to find some common thread here. Graphics drivers would be of particular interest. Both of my machines are nvidia graphics. One uses nvidia proprietary driver, and the other uses oss nouvelle driver.

[AbeFM]

That's what I was hoping for, John. You seem really helpful, and all I can think to do is send you info.

My machine is using the on-chip graphics, along with on-chip sound (a choice I made thinking they would get better driver support... An incorrect one) - I've had sound issues since I built the machine, stuff that goes away when I reinitialize the driver. Graphics always seem to be solid, although I don't use the 3D on there much and I've heard support for it is bad.

Is there an easy way to generate solid information on the driver that would help you the most? I'll dig around on the server and see what I can find. Thanks again!

(As an aside, I remember this happening once before, quite some time ago, getting encodes WAY slower than I thought they should be - I played with settings and somehow they ended up faster again, very fuzzy on the details but it easily could have been around 11.10)

[AbeFM]

Ubuntu12.10 32-bit, graphics experience "standard"
4GB memory, i3-2100T 2.5 GHz "4" core (I suspect this is two with hyperthreading) 30 gB SSD + 4x2GB RAID5
Graphics: Sandybridge Desktop x86/MMX/SSE2


I know 12.10 throws crash reports all the time, perhaps some of them are graphics related - hard to say, they seem to work, and for all the slow, HB is still putting out nice high quality movies.

[AbeFM]

Doing another test run since I saw a new version of HB push out last night. Using Swamp Thing again I'm still getting 42 fps on average, so no change (not that one was expected, but one can hope)....


Would more complete logs help?

Good Lord! I did the same normal-preset compression on my desktop over the network, it took 8 minutes. 11 times as fast!

[alex789]

I am having the same issue ever since I upgraded to Quantal, but it might be the 3.5 kernel?
Before the upgrade the machine was working great, but now it's probably going at 20% the speed it used to encode.
I haven't noticed this happen with other programs, just with HandBrake.
Here is some info to help JohnAStebbins, let me know if you want any other info:
Running HandBrakeCLI on a ubuntu server.

$ HandBrakeCLI -u
[09:35:51] hb_init: checking for updates
[09:35:51] Using http://handbrake.fr/appcast_unstable.xml
[09:35:51] latest: 0.9.8 i686, build 2012071700
[09:35:52] hb_init: starting libhb thread
HandBrake svn5044 (2012110601) - Linux x86_64 - http://handbrake.fr
Your version of HandBrake is up to date.

$ uname -a
Linux hostname 3.5.0-18-generic #29-Ubuntu SMP Fri Oct 19 10:26:51 UTC 2012 x86_64 x86_64 x86_64 GNU/Linux

Hardware:
Asus P8Z77-V PRO LGA 1155 Intel Z77 Motherboard http://www.asus.com/Motherboards/Intel_ ... 8Z77V_PRO/
Intel® Core™ i5-3450 Processor http://ark.intel.com/products/65511
Using the onchip graphics (Intel® HD Graphics 2500)
16 GB RAM DDR3 2133
RAID 5 with 3x Seagate 750GB HDDs

[AbeFM]

Motherboard: ZOTAC H67ITX-C-E H67 LGA1155 R

[alex789]

Any more updates?

[AbeFM]

John wants more information, but I'm not sure how to get it to him.

[Deleted User 11865]

John is also a busy person, so it may take him a while to respond.

[AbeFM]

Totally. But I don't think that's it - I think he's as lost as we are. He said he's had trouble reproducing it, and obviously working through remote people - well meaning though they be - can be a pain, and very time consuming especially given the time difference. I'd love to help, and would be willing to do anything he needs to get that working (compressing and showing movies is about all I use that machine for) but I've no idea where to go next. I think he's onto something with the poor drivers, for some reason they are terribly supported under ubuntu.... But from there, I don't know what to do.

Personally, I just want to get my presets imported to my windows box so I can clear out some of my backlog.

[s55]

He can't re-create it. Until he can, there is nothing to fix.

I can't create this problem either for what it's worth.

He's literally been in another country for months at end in the last year, so is very busy and not available for periods of time.

[AbeFM]

I've honestly been impressed at how helpful he's been. A lot of projects suffer under hands a lot less capable.


But, let's be useful here. If everyone lists their system specs, maybe we can find a pattern. I'm tempted to say the issue is with the on chip graphics (and their support) in the i- series of intel processors. A quick list of who has normal performance and who has useless, along with their system specs, may give us a clue as to which rocks to look under.

I would happily run some "system specs output to text file" command if there was one that was deemed helpful.

[JohnAStebbins]

Another thing that would be helpful to clarify. I think you've said (or implied) that you are encoding source material that is on a local hdd and the output is local hdd. But I want to be certain. Is there a network drive being used anywhere in the process?

[AbeFM]

Everything is local (on a RAID). Handbrake is installed on a different drive than the source/destination (local SSD).

As another poster mentioned - when I DO run HB (on windows) over the network, it's actually faster. I could do a drive speed test or something, but since it's not limiting remotely, I don't think it is locally.

Also, if that were the issue, wouldn't the CPU load by much lower, while waiting for the drives to "catch up"? Again, if there's some test I could run on the hardware, I would gladly do it.

[alex789]

Another thing that would be helpful to clarify. I think you've said (or implied) that you are encoding source material that is on a local hdd and the output is local hdd. But I want to be certain. Is there a network drive being used anywhere in the process?

Hi John, I am using a raid5 setup with 3 hdds and when encoding, the source and destination are both on the raid5 array.
It's basically the same setup which worked great under Ubuntu 12.04 and as soon as I upgraded to 12.10 the performance dropped.
Here's what I'm using as the graphics driver:

Code: Select all

$ modinfo i915
filename:       /lib/modules/3.5.0-18-generic/kernel/drivers/gpu/drm/i915/i915.ko
license:        GPL and additional rights
description:    Intel Graphics
author:         Tungsten Graphics, Inc.
license:        GPL and additional rights

[AbeFM]

Alex, thanks for posting the command.
Mine came out nowhere near as pretty:

Code: Select all

abe@HBO:~$ modinfo i915
filename:       /lib/modules/3.5.0-18-generic/kernel/drivers/gpu/drm/i915/i915.ko
license:        GPL and additional rights
description:    Intel Graphics
author:         Tungsten Graphics, Inc.
license:        GPL and additional rights
srcversion:     3FB984B3F228714518F1DB9
alias:          pci:v00008086d00000155sv*sd*bc03sc*i*
alias:          pci:v00008086d00000157sv*sd*bc03sc*i*
alias:          pci:v00008086d00000F30sv*sd*bc03sc*i*
alias:          pci:v00008086d00000D36sv*sd*bc03sc*i*
alias:          pci:v00008086d00000D26sv*sd*bc03sc*i*
alias:          pci:v00008086d00000D16sv*sd*bc03sc*i*
alias:          pci:v00008086d00000D3Asv*sd*bc03sc*i*
alias:          pci:v00008086d00000D2Asv*sd*bc03sc*i*
alias:          pci:v00008086d00000D1Asv*sd*bc03sc*i*
alias:          pci:v00008086d00000D32sv*sd*bc03sc*i*
alias:          pci:v00008086d00000D22sv*sd*bc03sc*i*
alias:          pci:v00008086d00000D12sv*sd*bc03sc*i*
alias:          pci:v00008086d00000A26sv*sd*bc03sc*i*
alias:          pci:v00008086d00000A16sv*sd*bc03sc*i*
alias:          pci:v00008086d00000A06sv*sd*bc03sc*i*
alias:          pci:v00008086d00000A2Asv*sd*bc03sc*i*
alias:          pci:v00008086d00000A1Asv*sd*bc03sc*i*
alias:          pci:v00008086d00000A0Asv*sd*bc03sc*i*
alias:          pci:v00008086d00000A22sv*sd*bc03sc*i*
alias:          pci:v00008086d00000A12sv*sd*bc03sc*i*
alias:          pci:v00008086d00000A02sv*sd*bc03sc*i*
alias:          pci:v00008086d00000C26sv*sd*bc03sc*i*
alias:          pci:v00008086d00000C16sv*sd*bc03sc*i*
alias:          pci:v00008086d00000C06sv*sd*bc03sc*i*
alias:          pci:v00008086d00000C2Asv*sd*bc03sc*i*
alias:          pci:v00008086d00000C1Asv*sd*bc03sc*i*
alias:          pci:v00008086d00000C0Asv*sd*bc03sc*i*
alias:          pci:v00008086d00000C22sv*sd*bc03sc*i*
alias:          pci:v00008086d00000C12sv*sd*bc03sc*i*
alias:          pci:v00008086d00000C02sv*sd*bc03sc*i*
alias:          pci:v00008086d00000426sv*sd*bc03sc*i*
alias:          pci:v00008086d00000416sv*sd*bc03sc*i*
alias:          pci:v00008086d00000406sv*sd*bc03sc*i*
alias:          pci:v00008086d0000042Asv*sd*bc03sc*i*
alias:          pci:v00008086d0000041Asv*sd*bc03sc*i*
alias:          pci:v00008086d0000040Asv*sd*bc03sc*i*
alias:          pci:v00008086d00000422sv*sd*bc03sc*i*
alias:          pci:v00008086d00000412sv*sd*bc03sc*i*
alias:          pci:v00008086d00000402sv*sd*bc03sc*i*
alias:          pci:v00008086d0000016Asv*sd*bc03sc*i*
alias:          pci:v00008086d0000015Asv*sd*bc03sc*i*
alias:          pci:v00008086d00000162sv*sd*bc03sc*i*
alias:          pci:v00008086d00000152sv*sd*bc03sc*i*
alias:          pci:v00008086d00000166sv*sd*bc03sc*i*
alias:          pci:v00008086d00000156sv*sd*bc03sc*i*
alias:          pci:v00008086d0000010Asv*sd*bc03sc*i*
alias:          pci:v00008086d00000126sv*sd*bc03sc*i*
alias:          pci:v00008086d00000116sv*sd*bc03sc*i*
alias:          pci:v00008086d00000106sv*sd*bc03sc*i*
alias:          pci:v00008086d00000122sv*sd*bc03sc*i*
alias:          pci:v00008086d00000112sv*sd*bc03sc*i*
alias:          pci:v00008086d00000102sv*sd*bc03sc*i*
alias:          pci:v00008086d00000046sv*sd*bc03sc*i*
alias:          pci:v00008086d00000042sv*sd*bc03sc*i*
alias:          pci:v00008086d0000A011sv*sd*bc03sc*i*
alias:          pci:v00008086d0000A001sv*sd*bc03sc*i*
alias:          pci:v00008086d00002E92sv*sd*bc03sc*i*
alias:          pci:v00008086d00002E42sv*sd*bc03sc*i*
alias:          pci:v00008086d00002E32sv*sd*bc03sc*i*
alias:          pci:v00008086d00002E22sv*sd*bc03sc*i*
alias:          pci:v00008086d00002E12sv*sd*bc03sc*i*
alias:          pci:v00008086d00002E02sv*sd*bc03sc*i*
alias:          pci:v00008086d00002A42sv*sd*bc03sc*i*
alias:          pci:v00008086d00002A12sv*sd*bc03sc*i*
alias:          pci:v00008086d00002A02sv*sd*bc03sc*i*
alias:          pci:v00008086d000029D2sv*sd*bc03sc*i*
alias:          pci:v00008086d000029C2sv*sd*bc03sc*i*
alias:          pci:v00008086d000029B2sv*sd*bc03sc*i*
alias:          pci:v00008086d000029A2sv*sd*bc03sc*i*
alias:          pci:v00008086d00002992sv*sd*bc03sc*i*
alias:          pci:v00008086d00002982sv*sd*bc03sc*i*
alias:          pci:v00008086d00002972sv*sd*bc03sc*i*
alias:          pci:v00008086d000027AEsv*sd*bc03sc*i*
alias:          pci:v00008086d000027A2sv*sd*bc03sc*i*
alias:          pci:v00008086d00002772sv*sd*bc03sc*i*
alias:          pci:v00008086d00002592sv*sd*bc03sc*i*
alias:          pci:v00008086d0000258Asv*sd*bc03sc*i*
alias:          pci:v00008086d00002582sv*sd*bc03sc*i*
alias:          pci:v00008086d00002572sv*sd*bc03sc*i*
alias:          pci:v00008086d0000358Esv*sd*bc03sc*i*
alias:          pci:v00008086d00003582sv*sd*bc03sc*i*
alias:          pci:v00008086d00002562sv*sd*bc03sc*i*
alias:          pci:v00008086d00003577sv*sd*bc03sc*i*
depends:        drm,drm_kms_helper,video,i2c-algo-bit
intree:         Y
vermagic:       3.5.0-18-generic SMP mod_unload modversions 686 
parm:           invert_brightness:Invert backlight brightness (-1 force normal, 0 machine defaults, 1 force inversion), please report PCI device ID, subsystem vendor and subsystem device ID to dri-devel@lists.freedesktop.org, if your machine needs it. It will then be included in an upcoming module version. (int)
parm:           modeset:Use kernel modesetting [KMS] (0=DRM_I915_KMS from .config, 1=on, -1=force vga console preference [default]) (int)
parm:           fbpercrtc:int
parm:           panel_ignore_lid:Override lid status (0=autodetect [default], 1=lid open, -1=lid closed) (int)
parm:           powersave:Enable powersavings, fbc, downclocking, etc. (default: true) (int)
parm:           semaphores:Use semaphores for inter-ring sync (default: -1 (use per-chip defaults)) (int)
parm:           i915_enable_rc6:Enable power-saving render C-state 6. Different stages can be selected via bitmask values (0 = disable; 1 = enable rc6; 2 = enable deep rc6; 4 = enable deepest rc6). For example, 3 would enable rc6 and deep rc6, and 7 would enable everything. default: -1 (use per-chip default) (int)
parm:           i915_enable_fbc:Enable frame buffer compression for power savings (default: -1 (use per-chip default)) (int)
parm:           lvds_downclock:Use panel (LVDS/eDP) downclocking for power savings (default: false) (int)
parm:           lvds_channel_mode:Specify LVDS channel mode (0=probe BIOS [default], 1=single-channel, 2=dual-channel) (int)
parm:           lvds_use_ssc:Use Spread Spectrum Clock with panels [LVDS/eDP] (default: auto from VBT) (int)
parm:           vbt_sdvo_panel_type:Override/Ignore selection of SDVO panel mode in the VBT (-2=ignore, -1=auto [default], index in VBT BIOS table) (int)
parm:           reset:Attempt GPU resets (default: true) (bool)
parm:           enable_hangcheck:Periodically check GPU activity for detecting hangs. WARNING: Disabling this can cause system wide hangs. (default: true) (bool)
parm:           i915_enable_ppgtt:Enable PPGTT (default: true) (int)
abe@HBO:~$

Hope it helps.

[alex789]

AbeFM I have truncated my output to just the top part which I thought was relevant.
It seems we are using the same graphics driver, so I wonder if that might be the issue?

[AbeFM]

Another thing that would be helpful to clarify. I think you've said (or implied) that you are encoding source material that is on a local hdd and the output is local hdd. But I want to be certain. Is there a network drive being used anywhere in the process?

Hey, John, any other bits you might want to know, tests I could run, etc? Obviously if there's nothing to work with, there's nothing to work with, but I'd love to get you any info I can, even on a longshot

[JohnAStebbins]

I really don't have any idea how to proceed with this. The folks on that fedora bugzilla thread (https://bugzilla.redhat.com/show_bug.cgi?id=847551) have been trying to hunt down some common cause as well and nothing is really emerging. The last proposed theory about what is common between users with the problem is turboboost. But there have been no responses to that yet.

[AbeFM]

Ah, very awesome. Still reading through it. Think I could disable enough of the graphics to make it use CPU only? I guess I could try to steal the nvidia card out of my desktop for a test run.

Definitely didn't think you weren't taking it seriously, just offering to help if I can. Far from a professional developer, but can diddle with things and report back. Will continue to read.

[AbeFM]

John,
How do I check where the temp files are? I suppose if I wait just a bit, the RAID will spin down and I can confirm it's all running off the SSD - so far, source/destination doesn't matter where they are (SSD or RAID) all four are slow.
It does seem with source and destination on the SSD, it's 10-15% faster on the first pass... But that's still not enough to get even close to where I was before. 42 fps verses 37 - but I used to get 80-120.
-------------
Edit: Difference might not even be as big as I said.
Unmounted RAID, didn't change compression, so I'm pretty sure everything was in the right spot. Went into bios, disables speedstep and turboboost and no difference.

[AbeFM]

Tried one at a time:
VMM, SpeedStep, EnergySaver, GPU Render Standby all in the bios, none helped.