AMD E2-1800 APU - could I be running it better?

Message

Havin_it · Post by **Havin_it** » Wed Aug 28, 2013 6:06 pm

Hi,

I've had my new netbook (not sure I can really call it that, might be more sub-notebook territory with a 1366x768 screen) for a while now since my trusty Sammy NC10 died on me, and I'm not too impressed with performance I have to say. It's a HP Pavilion dm1-4400as with AMD E2-1800 APU which from what little I could find out about it, sounded OK for my usage (no games; bit of video; browsing and NetBeans mainly). I went with ~amd64 install (standard x64 ABI, not x32) since I haven't had the opportunity with a "desktop" machine before, and I'm using the radeon driver.

My findings:

* Video plays OK but can still be a bit sluggish on HD stuff
* Audio is temperamental and often goes into loops (certain files seem to trigger this but KDE system sounds are the most common culprit, eg login music)
* UI (KDE) can get very laggy at times, like 5-10 seconds between input action and visual feedback - requires a reboot
* Sometimes won't go to sleep
* Sometimes when display turns off for power-saving, it won't come back (VT switch no help)
* Battery life is *ass*, maybe 4hrs average

All in all, it just seems so much poorer than what I expected from the machine, and a poor successor to the NC10. Windows 8 performance wasn't amazing either, so maybe the machine is just a lemon, but I really am wondering if there might be better system choices I could have made. Catalyst? 32-bit OS? x32?

I'm certainly about ready to try Catalyst if the radeon overhaul in kernel-3.11 doesn't make much difference, but I'd welcome any other suggestions I could try to improve any of the above issues. I've already applied every tweak suggested by powertop, none of which have made much difference (except maybe disabling NMI watchdog, which may have actually caused some instability) so I'm fresh out of ideas.

Thanks in advance...

Ion Silverbolt · Post by **Ion Silverbolt** » Thu Aug 29, 2013 5:48 am

I had Gentoo on a HP DM1z which uses an e-450. (Just slightly worse than an e-1800). I had no issues with it playing HD flash and 720p videos. It was even capable at playing Minecraft at low settings. No sleep issues either.

That being said, the e-series APU's need to die. Mainly because getting a system with an A6, or even an A8 APU is barely any more money, and the performance difference is A LOT compared to the e-series. I upgraded to a Toshiba s405 for $330 which has an A6 APU. Worlds faster than the e-450.

Anyway, to your problem:

1. Have you tried a different DE? I used Xfce with my e-450.
2. What kind of cflags are you running? Optimization can make a noticeable difference with these slow APU's.Especially when it comes to mplayer and the like.
3. Catalyst is definitely better for your e-1800. I too hope the 3.11 kernel improvements make a big difference with radeon though. I use radeon now with the A6. It's fast enough for me, but Catalyst was definitely better when I was using the e-450.

Havin_it · Post by **Havin_it** » Sat Aug 31, 2013 8:00 am

Hi Ion

1. A slimmer DM would of course be an option (I use FluxBox on some older machines and am comfortable enough with that) but it's more than a little sad to think I'd have to sacrifice KDE as it brings some nice stuff to the table, some of which (kioslaves!) I'd really miss.

2. Deep breath now:

Code: Select all

CFLAGS="-O2 -pipe -fomit-frame-pointer"
CFLAGS="${CFLAGS} -march=amdfam10 -mcx16 -msahf -mpopcnt --param l1-cache-size=32 --param l1-cache-line-size=64 --param 
l2-cache-size=512 -mtune=amdfam10"

The 2nd row are all the flags gcc infers from "-march-native", listed explicitly (because I use distcc).

My kernel config uses Athlon/Hammer/K8 for subarchitecture type, but I was never sure about this: should I actually be using Generic 64-bit instead? (Both work but I didn't notice much difference)

3. One additional question about this lot: I'm using radeon/r600 which seems right based on the x.org "decoder ring" although my model isn't explicitly mentioned (the GPU core is Wrestler / Radeon HD 7340). I did wonder if I should be using radeonsi instead, because I'm a little unsure of my facts on that one.

Post by **NeddySeagoon** » Sat Aug 31, 2013 8:31 am

Havin_it,

Please post your emerge --info and your /proc/cpuinfo
I run an E450 as a media player, mostly because its totally silent. It dates from before the time when SSDs were a reasonable price so it net boots.

A small number of hardware specific USE flags make quite a difference to graphics performance.

Athlon/Hammer/K8 is correct for the kernel. Its good that it appears to make little difference, since the kernel arbitrates between processes and provides nothing (almost) to the furtherance of those processes. You want the kernel code to run as little as possible.

Havin_it · Post by **Havin_it** » Sat Aug 31, 2013 8:53 am

emerge --info:

Code: Select all

Portage 2.1.13.1 (default/linux/amd64/13.0/desktop/kde, gcc-4.6.3, glibc-2.17, 3.10.0-gentoo x86_64)
=================================================================
System uname: Linux-3.10.0-gentoo-x86_64-AMD_E2-1800_APU_with_Radeon-tm-_HD_Graphics-with-gentoo-2.2
KiB Mem:     3627048 total,    502876 free
KiB Swap:    9047036 total,   9047036 free
Timestamp of tree: Sat, 10 Aug 2013 19:15:01 +0000
ld GNU ld (GNU Binutils) 2.23.1
distcc 3.1 x86_64-pc-linux-gnu [enabled]
ccache version 3.1.9 [enabled]
app-shells/bash:          4.2_p45
dev-java/java-config:     2.2.0
dev-lang/python:          2.7.5-r1, 3.2.5-r1, 3.3.2-r1
dev-util/ccache:          3.1.9
dev-util/cmake:           2.8.11.1
dev-util/pkgconfig:       0.28
sys-apps/baselayout:      2.2
sys-apps/openrc:          0.11.8
sys-apps/sandbox:         2.6-r1
sys-devel/autoconf:       2.13, 2.69
sys-devel/automake:       1.10.3, 1.11.6, 1.12.6, 1.13.4, 1.14
sys-devel/binutils:       2.23.1
sys-devel/gcc:            4.6.3
sys-devel/gcc-config:     1.8
sys-devel/libtool:        2.4.2
sys-devel/make:           3.82-r4
sys-kernel/linux-headers: 3.9 (virtual/os-headers)
sys-libs/glibc:           2.17
Repositories: gentoo
ACCEPT_KEYWORDS="amd64 ~amd64"
ACCEPT_LICENSE="*"
CBUILD="x86_64-pc-linux-gnu"
CFLAGS="-O2 -pipe -fomit-frame-pointer -march=amdfam10 -mcx16 -msahf -mpopcnt --param l1-cache-size=32 --param l1-cache-line-size=64 --param l2-cache-size=512 -mtune=amdfam10"
CHOST="x86_64-pc-linux-gnu"
CONFIG_PROTECT="/etc /usr/share/config /usr/share/gnupg/qualified.txt /usr/share/themes/oxygen-gtk/gtk-2.0"
CONFIG_PROTECT_MASK="/etc/ca-certificates.conf /etc/env.d /etc/fonts/fonts.conf /etc/gconf /etc/gentoo-release /etc/php/apache2-php5.5/ext-active/ /etc/php/cgi-php5.5/ext-active/ /etc/php/cli-php5.5/ext-active/ /etc/revdep-rebuild /etc/sandbox.d /etc/terminfo"
CXXFLAGS="-O2 -pipe -fomit-frame-pointer -march=amdfam10 -mcx16 -msahf -mpopcnt --param l1-cache-size=32 --param l1-cache-line-size=64 --param l2-cache-size=512 -mtune=amdfam10"
DISTDIR="/usr/portage/distfiles"
EMERGE_DEFAULT_OPTS="--nospinner --quiet-build=n"
FCFLAGS="-O2 -pipe"
FEATURES="assume-digests binpkg-logs ccache distcc distlocks ebuild-locks fail-clean fixlafiles merge-sync news preserve-libs protect-owned sandbox sfperms strict unknown-features-warn unmerge-logs unmerge-orphans userfetch userpriv usersandbox usersync"
FFLAGS="-O2 -pipe"
GENTOO_MIRRORS="http://www.mirrorservice.org/sites/www.ibiblio.org/gentoo/"
LANG="en_GB.UTF-8"
LDFLAGS="-Wl,-O1 -Wl,--as-needed"
MAKEOPTS="-j20 -l2.6"
PKGDIR="/usr/portage/packages"
PORTAGE_CONFIGROOT="/"
PORTAGE_RSYNC_OPTS="--recursive --links --safe-links --perms --times --omit-dir-times --compress --force --whole-file --delete --stats --human-readable --timeout=180 --exclude=/distfiles --exclude=/local --exclude=/packages"
PORTAGE_TMPDIR="/tmp"
PORTDIR="/usr/local/portage"
PORTDIR_OVERLAY=""
SYNC="rsync://hazel/gentoo-portage"
USE="X a52 aac acl acpi alsa amd64 bash-completion berkdb bluetooth branding bzip2 cairo cdda cddb cdr cli consolekit cracklib crypt css cups cxx dbus declarative dga djvu dri dts dvd dvdr emboss encode exif fam fbcon ffmpeg firefox flac fortran ftp gd gdbm geoip gif gimp gmp gnutls gphoto2 gpm graphviz gtk handbook hddtemp iconv imagemagick imap inotify java javascript jit jpeg kde kipi lame latex lcms ldap libkms libnotify lm_sensors lzma lzo mad maildir matroska mbox mmap mmx mmxext mng modules mp3 mp4 mpeg mplayer mtp mudflap multilib musicbrainz mysql ncurses nls nptl ocamlopt offensive ofx ogg opengl openmp pam pango pcre pda pdf perl phonon php plasma png policykit posix postscript ppds qt3support qt4 quicktime radius raw readline rss samba scanner sdl session sharedmem smp snmp sockets socks5 sound spell sqlite sqlite3 sse sse2 ssl ssse3 startup-notification subversion svg syslog tcpd theora threads tidy tiff tk tokenizer truetype udev udisks unicode upnp upnp-av upower usb v4l vcd vhosts vnc vorbis wifi wmf wxwidgets x264 xa xcb xcomposite xft xine xinerama xml xpm xscreensaver xv xvid xvmc zlib" ABI_X86="64" ALSA_CARDS="ali5451 als4000 atiixp atiixp-modem bt87x ca0106 cmipci emu10k1x ens1370 ens1371 es1938 es1968 fm801 hda-intel intel8x0 intel8x0m maestro3 trident usb-audio via82xx via82xx-modem ymfpci" APACHE2_MODULES="authn_core authz_core socache_shmcb unixd actions alias auth_basic authn_alias authn_anon authn_dbm authn_default authn_file authz_dbm authz_default authz_groupfile authz_host authz_owner authz_user autoindex cache cgi cgid dav dav_fs dav_lock deflate dir disk_cache env expires ext_filter file_cache filter headers include info log_config logio mem_cache mime mime_magic negotiation rewrite setenvif speling status unique_id userdir usertrack vhost_alias" CALLIGRA_FEATURES="kexi words flow plan sheets stage tables krita karbon braindump author" CAMERAS="ptp2" COLLECTD_PLUGINS="df interface irq load memory rrdtool swap syslog" ELIBC="glibc" GPSD_PROTOCOLS="ashtech aivdm earthmate evermore fv18 garmin garmintxt gpsclock itrax mtk3301 nmea ntrip navcom oceanserver oldstyle oncore rtcm104v2 rtcm104v3 sirf superstar2 timing tsip tripmate tnt ubx" GRUB_PLATFORMS="efi-64 multiboot" INPUT_DEVICES="keyboard mouse evdev synaptics" KERNEL="linux" LCD_DEVICES="bayrad cfontz cfontz633 glk hd44780 lb216 lcdm001 mtxorb ncurses text" LIBREOFFICE_EXTENSIONS="presenter-console presenter-minimizer" LINGUAS="en_GB" OFFICE_IMPLEMENTATION="libreoffice" PHP_TARGETS="php5-4" PYTHON_SINGLE_TARGET="python2_7" PYTHON_TARGETS="python2_7 python3_2" RUBY_TARGETS="ruby19 ruby18" SANE_BACKENDS="hp net" USERLAND="GNU" VIDEO_CARDS="radeon" XTABLES_ADDONS="quota2 psd pknock lscan length2 ipv4options ipset ipp2p iface geoip fuzzy condition tee tarpit sysrq steal rawnat logmark ipmark dhcpmac delude chaos account"
Unset:  CPPFLAGS, CTARGET, INSTALL_MASK, LC_ALL, PORTAGE_BUNZIP2_COMMAND, PORTAGE_COMPRESS, PORTAGE_COMPRESS_FLAGS, PORTAGE_RSYNC_EXTRA_OPTS, USE_PYTHON

/proc/cpuinfo:

Code: Select all

processor       : 0
vendor_id       : AuthenticAMD
cpu family      : 20
model           : 2
model name      : AMD E2-1800 APU with Radeon(tm) HD Graphics
stepping        : 0
microcode       : 0x500010d
cpu MHz         : 850.000
cache size      : 512 KB
physical id     : 0
siblings        : 2
core id         : 0
cpu cores       : 2
apicid          : 0
initial apicid  : 0
fpu             : yes
fpu_exception   : yes
cpuid level     : 6
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc extd_apicid aperfmperf pni monitor ssse3 cx16 popcnt lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch ibs skinit wdt arat hw_pstate npt lbrv svm_lock nrip_save pausefilter
bogomips        : 3393.76
TLB size        : 1024 4K pages
clflush size    : 64
cache_alignment : 64
address sizes   : 36 bits physical, 48 bits virtual
power management: ts ttp tm stc 100mhzsteps hwpstate

processor       : 1
vendor_id       : AuthenticAMD
cpu family      : 20
model           : 2
model name      : AMD E2-1800 APU with Radeon(tm) HD Graphics
stepping        : 0
microcode       : 0x500010d
cpu MHz         : 850.000
cache size      : 512 KB
physical id     : 0
siblings        : 2
core id         : 1
cpu cores       : 2
apicid          : 1
initial apicid  : 1
fpu             : yes
fpu_exception   : yes
cpuid level     : 6
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc extd_apicid aperfmperf pni monitor ssse3 cx16 popcnt lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch ibs skinit wdt arat hw_pstate npt lbrv svm_lock nrip_save pausefilter
bogomips        : 3393.76
TLB size        : 1024 4K pages
clflush size    : 64
cache_alignment : 64
address sizes   : 36 bits physical, 48 bits virtual
power management: ts ttp tm stc 100mhzsteps hwpstate

dE_logics · Post by **dE_logics** » Sat Aug 31, 2013 9:10 am

It's an AMD product. It's only Windows 8 compatible.

AMD has to die.

Hint:You should use vaapi hardware acceleration (x11-libs/libva). Only xbmc supports it, but you need the propitiatory drivers for that (which's a requirement for their closed source hardware).

Post by **NeddySeagoon** » Sat Aug 31, 2013 9:15 am

Havin_it,

That looks good. What frequency governer do you use?

Your /proc/cpuinfo shows

Code: Select all

cpu MHz         : 850.000

which may well be correct at the time of your snapshot but thats well throttled back.
I trust its not that low all the time ?

Havin_it · Post by **Havin_it** » Sat Aug 31, 2013 3:32 pm

@Neddy: I'm using ondemand governor. I used to have a plasmoid on the old box that showed the current clock speed, but I've not added it yet - might do that now actually - so not sure how it stacks up as far as what speeds are usually chosen.

Also, Ion above mentioned Flash videos playing OK: not my experience. Maybe it's something else in Firefox dragging things down, but I doubt it. Flash performance is awful and tends to turn the browser into treacle :( I heard previously that amd64 Flash was pretty bad so assumed this was normal.

@dE_logics: Could you elaborate on that somewhat polemical statement? I thought AMD were quite good on Linux generally (though I realise the ATI baggage may differ here). I will look into the VAAPI thing though, thanks for suggesting that (I never know what these things are when choosing USE flags etc...)

Ion Silverbolt · Post by **Ion Silverbolt** » Sat Aug 31, 2013 5:18 pm

My previous experiences with smooth flash playback was with the proprietary driver. Recent driver updates with radeon have made huge improvements though.

1. Do you have the vdpau use flag enabled for adobe-flash?

2. With kernel 3.10 and mesa 9.2, you should have full vdpau support using the radeon driver. So you might consider using mesa 9.2 if you don't mind using unstable and want to stick with the radeon driver.

As for AMD being "incompatible", that's not really a case on the CPU side. The graphics side can be tricky at times. Usually it involves having the right proprietary driver with the right kernel to get working graphics. It's more of a pain than an incompatibility. The best thing going right now is the improved radeon support. Hopefully using catalyst will be a thing of the past some day.

Ant P. · Post by **Ant P.** » Sun Sep 01, 2013 3:27 pm

Havin_it wrote:@dE_logics: Could you elaborate on that somewhat polemical statement? I thought AMD were quite good on Linux generally (though I realise the ATI baggage may differ here). I will look into the VAAPI thing though, thanks for suggesting that (I never know what these things are when choosing USE flags etc...)

That user is either severely misinformed or a liar. VDPAU works fine on Radeon GPUs.

Havin_it · Post by **Havin_it** » Mon Sep 02, 2013 4:53 pm

Ant P. wrote:
Havin_it wrote:@dE_logics: Could you elaborate on that somewhat polemical statement? I thought AMD were quite good on Linux generally (though I realise the ATI baggage may differ here). I will look into the VAAPI thing though, thanks for suggesting that (I never know what these things are when choosing USE flags etc...)
That user is either severely misinformed or a liar. VDPAU works fine on Radeon GPUs.

It certainly seems to - I made the required changes and mplayer does seem a little more responsive, with the machine sweating a bit less to I think (no empirical study undertaken so could be placebo effect...).

I notice that when I play an mp4 video, mplayer says:

Code: Select all

Forced video codec: ffmpeg12vdpau
Forced video codec: ffwmv3vdpau
Forced video codec: ffvc1vdpau
Forced video codec: ffh264vdpau
Opening video decoder: [ffmpeg] FFmpeg's libavcodec codec family
libavcodec version 54.92.100 (external)
Selected video codec: [ffh264vdpau] vfm: ffmpeg (FFmpeg H.264 (VDPAU))
==========================================================================
==========================================================================
Opening audio decoder: [ffmpeg] FFmpeg/libavcodec audio decoders
AUDIO: 22050 Hz, 2 ch, floatle, 64.0 kbit/4.53% (ratio: 7999->176400)
Selected audio codec: [ffaac] afm: ffmpeg (FFmpeg AAC (MPEG-2/MPEG-4 Audio))
==========================================================================
[AO_ALSA] Playback open error: Device or resource busy
[AO SDL] Samplerate: 22050Hz Channels: Stereo Format floatle
[AO SDL] Unsupported audio format: 0x1d.
[AO SDL] Unable to open audio: No available audio device
DVB card number must be between 1 and 4
AO: [null] 22050Hz 2ch floatle (4 bytes per sample)
Starting playback...
[VD_FFMPEG] Trying pixfmt=0.
[VD_FFMPEG] XVMC-accelerated MPEG-2.
Movie-Aspect is undefined - no prescaling applied.
VO: [vdpau] 640x480 => 640x480 H.264 VDPAU acceleration 
[VD_FFMPEG] XVMC-accelerated MPEG-2.
[VD_FFMPEG] Trying pixfmt=0.
[VD_FFMPEG] XVMC-accelerated MPEG-2.

We can see there that the vdpau-enabled codec is being used, but what's this about XVMC towards the end? I have that flag enabled too, but should I?

Also, from what I've been reading, VAAPI is something I could also use on top of VDPAU, but is there any point in doing that? Is it just a question of which apps use which API? For example I see that ffmpeg has both flags, but mplayer has only vdpau and vlc (which I also have installed purely for a Phonon backend, I don't use it on its own) only has vaapi. Is there any point in enabling vaapi just for phonon purposes (which I don't think involves video)? Could that have any effect on the audio issues I'm having (this can be seen in the mplayer output above; currently stuck in a half-second loop of the login music)?

I can't imagine how folk on binary distros cope with this alphabet soup!

dE_logics · Post by **dE_logics** » Tue Sep 03, 2013 1:33 am

" [VD_FFMPEG] XVMC-accelerated MPEG-2. "

Humm...

crshbndct · Post by **crshbndct** » Tue Sep 03, 2013 1:46 am

Try installing Mesa 9.2 and enabling the vdpau USE flag, and if you are compiling firefox, enable the gstreamer flag for it.

This will give you hardware acceleration for flash and other media. (Select vdpau in the player of your choice, I find smplayer is best.)

Check to see what your GPU speed is running at, and for the DPM stuff to be enabled, you need to pass radeon.dpm=1 to the kernel at boot. There is a good chance that your GPU is running at whatever its lowest speed is, so the DPM stuff should make a difference, but you can test it by manually turning the power level to high and seeing if that speeds things up, if it does then 3.11 is your fix. (It was released like 2 hours ago, so it should be available very shortly) I would suggest playing with the setting manually anyway, (as a way to test the difference in performance) before enabling it permanently in your kernel.

If you need better 3d performance, (which it doesn't sound like you do from your description, but its worth a try) set R600_DEBUG=sb in your environment, it should speed things up a little.

You need to set some other things in your environment to get vdpau working perfectly:

VDPAU_DRIVER=r600
LIBVA_DRIVER_NAME=vdpau

dE_logics · Post by **dE_logics** » Tue Sep 03, 2013 2:04 am

For flash you may also try Gnash.

Havin_it · Post by **Havin_it** » Tue Sep 03, 2013 10:26 am

crshbndct wrote:Try installing Mesa 9.2 and enabling the vdpau USE flag, and if you are compiling firefox, enable the gstreamer flag for it.

This will give you hardware acceleration for flash and other media. (Select vdpau in the player of your choice, I find smplayer is best.)

Check to see what your GPU speed is running at, and for the DPM stuff to be enabled, you need to pass radeon.dpm=1 to the kernel at boot. There is a good chance that your GPU is running at whatever its lowest speed is, so the DPM stuff should make a difference, but you can test it by manually turning the power level to high and seeing if that speeds things up, if it does then 3.11 is your fix. (It was released like 2 hours ago, so it should be available very shortly) I would suggest playing with the setting manually anyway, (as a way to test the difference in performance) before enabling it permanently in your kernel.

If you need better 3d performance, (which it doesn't sound like you do from your description, but its worth a try) set R600_DEBUG=sb in your environment, it should speed things up a little.

You need to set some other things in your environment to get vdpau working perfectly:

VDPAU_DRIVER=r600
LIBVA_DRIVER_NAME=vdpau

Already upgraded mesa and enabled vdpau: that's what I was intending to show by the post above. I have added "export VDPAU_DRIVER=r600" in my .bashrc, but not the other line as I haven't enabled vaapi yet, which is what that line's for according to the wiki. Refer again to the questions I actually ask in my post above. Particularly, given the weird output dE_logics flagged above, should I actually be disabling xvmc flag?

I use firefox-bin for preference, and I'm not over the moon about having to get gstreamer involved: I've gone to a lot of effort to avoid it over the years, is it the only way? (Next thing you know I'll be saddled with bloody pulseaudio too

)

CPU is scaling, it's just in low-power most of the time. I ran

Code: Select all

watch -n0.5 grep MHz /proc/cpuinfo

for a while with just firefox running and it mostly stayed at 850 on both cores, but if I moved the mouse around or scrolled on the firefox page one or both cores would jump briefly to 1700. But this is cpufreq; the improvements in 3.11 are for the GPU power management aren't they?

I'll try the 3d tweak you mention; although I'm not really "doing" anything 3d-heavy (yet!), I do use compositing in kwin so it might help with that, right?

One further question that came to light when upgrading mesa: how about this USE=r600-llvm-compiler flag? Anyone have any experience with this?

aim nano · Post by **aim nano** » Wed Sep 04, 2013 7:04 am

i thought amd64 and ~amd64 clashed:
http://wiki.gentoo.org/wiki/ACCEPT_KEYWORDS
"One should not specify the stable keyword (amd64) when adding the testing keyword (~amd64) because ACCEPT_KEYWORDS is an incremental variable."
-just curious

also, are you using the 64bit flash? i was, then downgraded to 32 with nspluginwrapper and my flash performance improved, there was a bug in hw accel for 64 bit (not sure if it's applicable in your situation)

Havin_it · Post by **Havin_it** » Wed Sep 04, 2013 11:42 am

aim nano wrote:i thought amd64 and ~amd64 clashed:
http://wiki.gentoo.org/wiki/ACCEPT_KEYWORDS
"One should not specify the stable keyword (amd64) when adding the testing keyword (~amd64) because ACCEPT_KEYWORDS is an incremental variable."
-just curious

Hmm, can't say why I did that -- just a weird brainfart on my part I guess. The somewhat cryptic passage above does little to explain what's wrong with doing this, but I'm sure there is no reason for it (I know I'm getting unstable packages fine currently) so I'll take the stable keyword out. This hasn't changed since install-time, FWIW.

aim nano wrote:also, are you using the 64bit flash? i was, then downgraded to 32 with nspluginwrapper and my flash performance improved, there was a bug in hw accel for 64 bit (not sure if it's applicable in your situation)

I had heard something to this effect, and it's probably worth a look. So far I've made some edits to /etc/adobe/mms.cfg to force accelerated video decoding (in original config it only accelerates display) and Youtube's "data for nerds" panel shows me this has worked, however videos on some other sites have caused the plugin to crash every time.

Ant P. · Post by **Ant P.** » Wed Sep 04, 2013 7:32 pm

Havin_it wrote:I notice that when I play an mp4 video, mplayer says:

Code: Select all

Forced video codec: ffmpeg12vdpau
Forced video codec: ffwmv3vdpau
Forced video codec: ffvc1vdpau
Forced video codec: ffh264vdpau
Opening video decoder: [ffmpeg] FFmpeg's libavcodec codec family
libavcodec version 54.92.100 (external)
Selected video codec: [ffh264vdpau] vfm: ffmpeg (FFmpeg H.264 (VDPAU))
==========================================================================
==========================================================================
Opening audio decoder: [ffmpeg] FFmpeg/libavcodec audio decoders
AUDIO: 22050 Hz, 2 ch, floatle, 64.0 kbit/4.53% (ratio: 7999->176400)
Selected audio codec: [ffaac] afm: ffmpeg (FFmpeg AAC (MPEG-2/MPEG-4 Audio))
==========================================================================
[AO_ALSA] Playback open error: Device or resource busy
[AO SDL] Samplerate: 22050Hz Channels: Stereo Format floatle
[AO SDL] Unsupported audio format: 0x1d.
[AO SDL] Unable to open audio: No available audio device
DVB card number must be between 1 and 4
AO: [null] 22050Hz 2ch floatle (4 bytes per sample)
Starting playback...
[VD_FFMPEG] Trying pixfmt=0.
[VD_FFMPEG] XVMC-accelerated MPEG-2.
Movie-Aspect is undefined - no prescaling applied.
VO: [vdpau] 640x480 => 640x480 H.264 VDPAU acceleration 
[VD_FFMPEG] XVMC-accelerated MPEG-2.
[VD_FFMPEG] Trying pixfmt=0.
[VD_FFMPEG] XVMC-accelerated MPEG-2.

We can see there that the vdpau-enabled codec is being used, but what's this about XVMC towards the end? I have that flag enabled too, but should I?

It's working, those weird messages are just a cosmetic issue; the important word is "accelerated". If you run it via `time mplayer` and then `time mplayer -vo xv` (the default) you should notice a large difference in CPU time. It'll be more noticeable with 720p or larger videos.

Also, from what I've been reading, VAAPI is something I could also use on top of VDPAU, but is there any point in doing that? Is it just a question of which apps use which API? For example I see that ffmpeg has both flags, but mplayer has only vdpau and vlc (which I also have installed purely for a Phonon backend, I don't use it on its own) only has vaapi. Is there any point in enabling vaapi just for phonon purposes (which I don't think involves video)? Could that have any effect on the audio issues I'm having (this can be seen in the mplayer output above; currently stuck in a half-second loop of the login music)?

You can safely leave it enabled if you want, but in this case it's just another layer of bloat.

Havin_it · Post by **Havin_it** » Fri Sep 06, 2013 3:29 pm

OK, just upgraded to kernel-3.11 and, well, this is certainly no placebo effect

How do I know: I've had the box turned on for a good 20min, and the APU fan vent area isn't lethally hot to the touch. Battery % is definitely ticking down slower too, if I'm any judge.

Feeling a whole lot happier about this box today -- great job, Radeon team

On the other hand, wifi performance seems to have dropped severely; always something, isn't there?

That was something I missed from my original list of grievances (as it wasn't concerning the APU): the Ralink RT3290 (rt2800_pci) wifi'n'bluetooth combo chip (no driver for the bluetooth part, other than an OEM one that doesn't build against recent kernels). I knew the bluetooth was probably a non-starter and can live with that, but the wifi reception took a bad hit with this kernel - I now can't sit in my favourite chair, so this shit just got real

Ah well, swings and roundabouts...

So, can anyone articulate the relative merits of USE=r600-llvm-compiler? Because the Phoronix articles I've read are clear as mud :S

Ant P. · Post by **Ant P.** » Fri Sep 06, 2013 6:35 pm

The rt2800 issue sounds a bit worrying; I'm using one of those and they're usually fine...

I don't think I've ever successfully gotten that LLVM thing to compile. You can get similar benefits by putting R600_DEBUG="sb" in an env.d file, but I've no idea if their effects stack or not (though it'd be nice...)

Havin_it · Post by **Havin_it** » Sat Sep 07, 2013 5:27 pm

Ant P. wrote:The rt2800 issue sounds a bit worrying; I'm using one of those and they're usually fine...

Unless it's an RT3290 or RT5390, you might be OK. I found this patch that went into the kernel between 3.10 and 3.11 which might be responsible, and this proposed patch that sounded like it might address the issue. However, after applying it, I'm now also getting panics

I might see if I can revert that previous patch, or maybe there's some other change in that code between the kernels, because I had no complaints before. Anyway, guess that's a bit OT for this thread, I'll start a bug for that I guess.

Ant P. wrote:I don't think I've ever successfully gotten that LLVM thing to compile. You can get similar benefits by putting R600_DEBUG="sb" in an env.d file, but I've no idea if their effects stack or not (though it'd be nice...)

Odd thing is, R600_DEBUG="sb" is already in my environment (not put there by me, AFAICR). Maybe this or some other flag causes this to be set? I'm not too clear on how they relate to each other (never understood this 3d juju too well, as various laptops will attest).