CPU schedulers compared (bfs vs cfs)

Abstract
Con Kolivas’ Brain Fuck Scheduler (bfs) was designed to provide superior desktop interactivity and responsiveness to machines running it.[1] However, it was not implicitly designed to provide superior performance. The purpose of this study was to evaluate the Completely Fair Scheduler (cfs) in the vanilla Linux kernel and the bfs in the corresponding kernel patched with the ck1 patchset. Seven (7) different machines were used to see if differences exist and, to what degree they scale using performance based metrics. Again, these end-points were never factors in the primary design goals of the bfs. Results were encouraging.
Kernels patched with the ck1 patch set including the bfs outperformed the vanilla kernel using the cfs at nearly all the performance-based benchmarks tested. Further study with a larger test set could be conducted, but based on the small test set of 7 PCs evaluated, these increases in process queuing, efficiency/speed are, on the whole, independent of CPU type (mono, dual, quad, hyperthreaded, etc.), CPU architecture (32-bit and 64-bit), 64 bit) and of CPU multiplicity (mono or dual socket).
Moreover, several "modern" CPUs (Intel C2D and Ci7) that represent common workstations and laptops, consistently outperformed the cfs in the vanilla kernel at all benchmarks. Efficiency and speed gains were small to moderate.
Link to complete study
http://repo-ck.com/bench/cpu_schedulers_compared.pdf
[1] http://ck.kolivas.org/patches/bfs/bfs-faq.txt
Comments are welcomed.
Last edited by graysky (2012-10-20 20:07:46)

Thaodan wrote:Will it replace cfs?
AFAIK, Con Kolivas (creator of BFS) has no intentions of bringing it to the mainline: http://ck.kolivas.org/patches/bfs/bfs-faq.txt
Are you looking at getting this into mainline?
LOL.
No really, are you?
LOL.
Really really, are you?
No. They would be crazy to use this scheduler anyway since it won't scale to
their 4096 cpu machines. The only way is to rewrite it to work that way, or
to have more than one scheduler in the kernel. I don't want to do the former,
and mainline doesn't want to do the latter. Besides, apparently I'm a bad
maintainer, which makes sense since for some reason I seem to want to have
a career, a life, raise a family with kids and have hobbies, all of which
have nothing to do with linux.

Similar Messages

What is the difference between bfs and bfq ??

I read a lot of good things about BFS for desktops and I wanted to give it a try, so I installed a linux-ck kernel for core2 from repo-ck (what I'm using).
I checked what I/O schedulers I had with the command "dmesg | grep scheduler" and found out I could choose between bfs and bfq...I have no idea what is the difference..and having trouble finding a clear answer ..

If you want to use BFQ:
https://wiki.archlinux.org/index.php/li … _Scheduler
I have no idea if this would be a good idea with an SSD though. As the above posts already indicated, BFS and BFQ share two thirds of their acronyms, that's it. They are entirely different things - and different "types" of things at that. You do not chose between them. You can chose between BFS and CFS for your process scheduler - and using linux-ck means you chose BFS. You can also chose between BFQ and other I/O schedulers like CFQ, NOOP, and others.
Last edited by Trilby (2013-02-10 20:27:37)

Call Park service causing high CPU usage

Hi,
I've got a Lync 2013 pool with 4 servers running. Everything is working fine except I notice that when I enable the call park service, the CPU usage goes from hovering at around 4% to jumping all over the place (anything from 11% to 30%). I've
not seen this behaviour on any other Lync platform I have deployed.
All machines are virtuals, running 2 Sockets with 8 cores each. Each server has 25GB of RAM. Any perfmon report I run only confirm that the Call Park Service is using a lot of CPU resource (compared with other services)
Any help greatly appreciated.
Thanks
Mike

Please check if the CPU usage drops down when you stop the call park service.
Please check the Call Park Configurations are correct by articles at
http://technet.microsoft.com/en-us/library/gg399014.aspx
Lisa Zheng
TechNet Community Support

(Can't Be Solved) Flash uses too much CPU

Hello all,
In the process of setting up my first Arch system, I currently experiment with OpenBox, and have tried using Flash (flashplugin) under Chromium, both latest version from the repos.
I tried watching a 720P video on youtube (Big Buck Bunny), but it was a complete failure.
After some searching, I disabled hardware decoding using
/etc/adobe/mms.cfg
Next, I completely disabled the Laptop Mode Tools daemon to make sure no powersaving was in play (and rebooted of course).
I also set the CPU governor to "performance" (though using cpupower, the CPU does scale to maximum under "ondemand").
All of the above did not help - CPU usage still reach ~80% when playing the 720P video unmaximized (and no dropped frames), and I get lots of dropped frames when going full screen.
Is it really what I can expect from flashplugin under Arch or have I missed something (hopefully)?
My CPU is Core 2 Duo P8600 which is more than capable of handling such videos (for reference only, software decoding of the same video under Windows 7 hover around 30% CPU usage (also Chromium and Flash, different Chromium version)).
Thanks, Adam.
Last edited by adam777 (2012-06-04 12:55:54)

OK, back from Ubuntu.
Did a fresh Ubuntu 12.04 LTS installation, and just installed Chromium and Flash from the software center.
Chromium version is 18.0.1025.151 (Arch has 19.0.1084.52).
Flash version is 11.2 r202 (Arch has 11.2.202.235, which I suppose is the same).
Also tried built-in Firefox and using proprietary drivers (still pure software decoding and rendering).
Under all scenarios, behavior was the same as in Arch - ridiculously high CPU usage comparing to Windows.
I guess Flash under Linux really is THAT bad, perhaps this new "chrome-only" flash will behave better (and will most definitely kept up-to-date in the AUR or so).
(I've already tried the currently available pepper flash from the AUR, with no better results).
Can't see much more that can be done in regards to direct flash rendering, really annoying.
Might try some sort of flash-video-automatic-replacement-with-vlc-or-something-like-that.
Adam.
PS.
My 20 minutes with Ubuntu were enough - Arch is definitely the way to go, even for a pretty green Linux user such as myself (have some background, though).
Last edited by adam777 (2012-06-04 12:57:06)

Sculpture CPU hungry?

I'm thinking of upgrading to the studio version just to get Sculpture. But from what I hear Sculpture is CPU hungry.
So how does Logic perform if there are multiple Structure tracks in a project? Does it crash or get really slow?

Hi,
On my machine (see sig) I can play max four (polyphonic) simultaneously, without any other plugin active. logic does not slow down, though both my processors are 100% busy. Logic has never crashed because it was overloaded. It simply stops playback when it overloads, and that simple principle works just fine.
The graphics in Logic do become unresponsive on my machine, when both processors are that busy. However, the 'freeze tracks' function can come to the rescue; it reduces the tracks load on the CPU to near zero.
But anyway, 4 instances on my G4 would equate to about 12 to 16 instances on your iMac (depending on speed). To have that many Sculptures in one piece sounds odd, a bit like a rockband with ten guitarists. But, to answer your question, yes, it is CPU hungry, compared to most other Logic Instruments. Although the EVOC can be a guzzler too. But your iMac will manage, as long as you manage your iMac.
regards, Erik.

Chromium + Flash + Pandora = very high CPU usage

I know Pandora is far from the most efficient software out there, but in Chromium it takes roughly 1.6 GHz of CPU cycles, compared to 260 MHz in Firefox. Same version of Flash (64 bit preview). I was hoping to switch from Firefox to Chromium as it seems to be faster overall, but draining my laptop's battery like this is a show-stopper.
chromium 10.0.648.205-1
xorg-server 1.10.1-1
xf86-video-ati 6.14.1-1
Firefox 4.0 (64-bit) from upstream
Adobe Flash 10.3 d162 (64-bit preview)
With Chromium:
Cn Avg residency P-states (frequencies)
C0 (cpu running) (45.4%) 2.00 Ghz 100.0%
polling 0.4ms ( 0.1%) 1.60 Ghz 0.0%
C1 mwait 0.2ms (26.9%) 1200 Mhz 0.0%
C4 mwait 0.2ms (27.6%)
Top causes for wakeups:
65.4% (1557.0) [radeon] <interrupt>
12.5% (297.2) [kernel scheduler] Load balancing tick
12.4% (295.4) kworker/0:0
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
3609 john 20 0 455m 131m 20m S 50.2 3.3 1:40.29 chromium
3151 root 20 0 115m 32m 15m R 30.6 0.8 1:06.56 X
With Firefox:
Cn Avg residency P-states (frequencies)
C0 (cpu running) (19.6%) 2.00 Ghz 0.0%
polling 2.3ms ( 0.0%) 1.60 Ghz 0.0%
C1 mwait 0.7ms ( 5.9%) 1200 Mhz 100.0%
C4 mwait 1.0ms (74.4%)
Top causes for wakeups:
29.8% (192.6) [kernel scheduler] Load balancing tick
26.3% (170.1) kworker/0:0
15.6% (100.8) firefox-bin
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
3860 john 20 0 709m 220m 37m S 22.0 5.6 0:29.94 firefox-bin
3151 root 20 0 110m 25m 8472 S 1.7 0.6 1:55.23 X
Last edited by jlindgren (2011-04-18 23:25:06)

The almost exact 100 wakeups per second makes me wonder if Pandora tries to redraw constantly but gets throttled to 100 FPS in Firefox.
Edit: Okay, I guess I should have done a little more research before posting here; seems to be a known problem with Pandora and Chromium (on Windows and Mac also (!)):
http://code.google.com/p/chromium/issue … l?id=38794
http://code.google.com/p/chromium/issue … l?id=77691
http://superuser.com/questions/149230/w … -cpu-usage
http://hintsforums.macworld.com/showthread.php?p=617863
Last edited by jlindgren (2011-04-19 01:42:57)

Kernel 2.6.32 will boost x264 encode speed BIG TIME for multicore CPUs

It's due to a tweaked scheduler in the 2.6.32 codebase. I measured a 53 % speed boost comparing 2.6.31-5 to 2.6.32-rc6 on my quad. See my post over at the handbrake forum here for details and see this page too.
The test encode was a 60 fps 720p (1280x720) encode using the "High Profile" on handbrake snapshot 2907 (which you can get in the AUR here):
Arch x86_64 (kernel 2.6.31.5): avg 15.99 fps
Arch x86_64 (kernel 2.6.32-rc6): avg 24.45 fps
Linux (kernel 2.6.31-bfs): avg 24.49 fps
Win XP 64-bit edition: avg 20.41 fps
Linux is 53 % faster w/ the new kernel and 20 % faster than Windows. Can't wait for 2.6.32 to go final!
Last edited by graysky (2009-12-06 10:57:58)

bangkok_manouel wrote:
MP2E wrote:Epic! I always encode things in x264(though through FFMPEG, though that shouldn't make a difference since it uses the x264 library, should it?) but I'm a user of BFS. Since you are referring to a scheduler change, will this affect me?
you'll need to go back to the mainline CPU sched (CFS)
But this bug was noticed in the first place because BFS kicked CFS's ass with x264, so sticking with BFS is fine
@tomd123: It's a general improvement in the scheduler, which should translate to other workloads as well. x264 got a huge boost, not everything will get such a huge increase in speed.

It will break my heart, but must I abandon Firefox?

UPDATE 9th Feb Issue now SOLVED in the other thread. Disabling McAfee SiteAdvisor resolves the issue in this instance. [/questions/984055?page=2#answer-532036] Note presence of ''McAfee MSS+ NPAPI Plugin''
Please see and continue with thread [/questions/984055] ~J99 Modified February 1, 2014 9:39:33 PM BST by John99
Almost two weeks ago, after years of admirably flawless performance, my Firefox 26.0 was struck by recurrent episodes of complete unresponsiveness, and frequent crashes.
Efforts to remedy these problems, involving disabling various add-ons and JS, failed.
Now I have some important new information. When operating normally, Firefox never uses more than about 3% or, rarely 5% of my CPU, even when I have many tabs open to websites, even when several of those websites are playing videos, etc. (I have had the Task Manager open constantly this week in order to be able to close Firefox when it's so unresponsive I can't use X to close it, so I was in a position to observe CPU usage.) But when Firefox is in the early stages of an episode of unresponsiveness (for example, when I can't download files or play videos) the CPU usage is greater than 25%. What's even more remarkable is that if I then close all tabs, and have only one “New Tab” open (so that I'm theoretically not using Firefox at all), the CPU usage remains over 25%!!
Over the past few days I've a) confirmed and reconfirmed this perfect correlation between Firefox's unresponsiveness and a dramatic, unprecedented increase in CPU usage to in excess of 25%, occurring completely out of the blue, in no way triggered by my doing anything, and b) verified that when Firefox is responding normally, NOTHING I do causes the usage to go above about 5%.
But the question is: what is causing this abnormal CPU usage that is clearly associated with Firefox's unresponsiveness?
For the record, I have run repeated scans using Windows Defender and MalwareBytes and nothing malicious has been detected, and I have had no other problems with my computer besides the ones with Firefox. Chrome continues to operate perfectly normally.

I suggest you continue this conversation in the thread you have already opened.
The fact that you are using the CPU heavily compared with normal may help explain the Firefox unresponsiveness.
The question is what is it doing.
*Is it a direct Firefox process taking up this CPU use.
** If so why. is malware running and making use of Firefox ? Is this only on close down ? is it only on a particular site ?
*Is it the plugincontgainer ?
*Is it another process that may affect or slowdown Firefox ?
** can you identify the process.
** is security software involved, scanning or updating
Presumably sysinternals software - now available from microsoft - will help you answer such questions.
Ensure you give the information in the other thread so it may be considered along with the details you have already posted.
*''' The thread is ''After years of happily using Firefox, sudden, recurrent calamity'' [/questions/984055]'''
Back to basics
You really need to troubleshoot by using Firefox
*in a new clean profile
* with all plugins disabled
*in Firefox's safemode
*even try it with only this forum open as the only tab
I would hope in such a configuration you do not get
* 25% CPU usage.
* Crashes
We should be able to work from that to figure out what else is implicated in your problems.
Please answer in the other thread [/questions/984055] .

A lot of questions

Background.
A while ago I asked for some advice about external drives. I currently have 4 WD 4TB attached to a sonnet eSATA card and 3 WD 2 TB attached to a sonnet FW card and was having some issues with access times when saving to them. I decided to order some OWC mercury enclosures, unfortunately I had to order them from the US and I think they're sending them on a raft, since they haven't yet arrived.
After yet another failure of one of the WD 4TB drives, I've now ordered 2 x WD 2TB green drives as a temporary solution which I intend putting in my remaining 2 internal bays.
I've had a few issues with my Mac Pro which I've largely ignored over the 3 years I've had it, and I'm tempted to throw in the towel and go for a new mac at the same time as I'm re-organising my external storage. However I thought I'd list the issues I have and see if anyone has any comments about them in addition to some questions I have relating to configuring the possible new mac.
_Current issues._
The first issue I had was from new, it had FW issues from day one and FW completely died after about a week. Apple did say they would take it in but rather than be without it I purchased a sonnet FW card.
About a year and a half ago the display failed to light up after a shut down and restart, several attempts later it did eventually restart properly. This has happened regularly since but a couple of shut downs and restarts always work.
When this happens I can hear the mac starting normally but I just have no display, on occasions the attempts to restart it will ramp up the fans to maximum and I have to shut down and pull the power cord to reset.
My graphics card is a Radeon X1900, but without really having enough knowledge to make such a statement I wonder if it isn't PCI related rather than just the graphics card.
For the avoidance of doubt this only happens with some shut downs and never happens with restarts.
I also have some doubts about my RAM, the info from profiler and activity monitor don't seem to match up.
_A new mac or not and what configuration._
First question is how long has it been since the mac pro line was updated.
Not sure what CPU's I have they're the first Quad cores (3 Ghz), I've usually bought my macs by going for the top model in the belief they'll serve my needs for longer and not that I particularly need the horsepower they give at the point of purchase. If I do this today it's a hefty price, so does someone have some sort of guide as to how my current CPU's compare with todays line up. My current CPU's are still more than acceptable for what I do.
My intention has been (when the raft arrives) to move all my data internally and max out the macs HDD bays, to backup everything via time machine to one eSATA enclosure, back up data with 3rd party software to a second enclosure and have a 3rd enclosure for temporary use, a spare so to speak. My intention was to use green drives in the enclosures because speed isn't an issue and they're used only occasionally. Speed isn't really an issue internally either, would greens be sufficient or are the blacks not only faster but more robust.
_Last question_ (for now)
I have a lot of failed drives lying about, I appreciate it's speculative but what proportion of those drives are likely to have failed because of their cheap enclosures and how many because of the drives themselves, in short what's the chance I can use some of them in my new mercury enclosures (once the raft arrives)

Winston Churchill wrote:
What about my current MP how would it compare to the current Mac pro line up.
I have the same basic machine as yours and have used Cinebench CPU Benchmark to compare with the latest Mac Pro but obviously yours may be faster or slower depending the precise components in your machine:
*My Machine:*
8C/8T @3.00GHz, Intel Xeon CPU X5365 - Score 6.24
*Reference Machine*
8C/16T @2.92 Ghz, Intel Xeon CPU X5570 - Score 10.33 (65% Improvement)
Obviously there is more to performance than raw CPU performance but having mod'ed mine pretty well now (256 GB SSD Boot drive, 4,.5 TB Raid0 data drive, ATI Radeon HD 4870 Graphics - see blog posts below) I'm finding much of the work I am doing (video work mainly) is processor bound. I've also been thinking about upgrading. You should maybe consider the same.
Your 1900XT for example: I had one of these BTO and hated the noise and lack of performance. I jumped to the 8800GT when it was made available for the early Mac Pro's and now have a 4870 which is fact and quiet. This is hundreds of pounds of graphics upgrades that I could have used on a new machine.
The mods I have made have significantly improved my machine as you can see in the results of my blog posts below. However the best solution for you is probably to upgrade to a new machine. But I would encourage you NOT to buy a new Mac Pro now as the current model is 18 months old and Apple must be launching their new models soon.
Implementing an SSD for the boot drive may actually be a good investment now whilst waiting for the new models to come out as it will significantly enhance your existing system and you can swap it into your new machine when it arrives so you wouldn't lose anything.
Good luck anyway
Alan
*ATI 487*0
http://www.theapplechap.com/TheAppleChap/Blog/Entries/2010/4/16The_ultimate_graphics_on_the_earlyMacPro.html
*Crucial C300 SSD*
http://www.theapplechap.com/TheAppleChap/Blog/Entries/2010/4/9Speed_and_Silence_-_The_joys_ofSSD.html
*Raid0 4.5 TB Disk Array*
http://www.theapplechap.com/TheAppleChap/Blog/Entries/2010/4/2Earning_your_stripes_-_Raid0_on_a_MacPro.html

How to make use of the crypto modules in the new Sparc T2 processors?

Having just received some frontal indoctrination about the features of the new Niagara (T1/T2) processor platform, I've started to dig around its ncp cryptography support cores a bit, but I've got a few open ends that I hope someone here can tie up for me..
- Is there any way to make current OpenSSL take advantage of the ncp cores? The FAQs don't mention it anywhere except a general "Included ability to use crypto hardware options". Does this already include Sun ncp? The openssl derivate that comes with Solaris10 is antique and I really want to replace it.. but if that means losing hardware support, I'm considering ditching the T2 altogether and going for US4+, doing crypto in the "old" way on the main cpu cores.
- The Niagara crypto blueprint mentions some commands to run speed tests, which I did on a T2000 server I've got around here, and on an old 420 in parallel, and I got some results that puzzle me:
The T2000:
sign verify sign/s verify/s
dsa 512 bits 0.0001s 0.0001s 13025.0 11979.2
dsa 1024 bits 0.0001s 0.0001s 12835.7 12426.3
dsa 2048 bits 0.3030s 0.6135s 3.3 1.6
The 420:
sign verify sign/s verify/s
dsa 512 bits 0.001171s 0.001427s 853.8 700.7
dsa 1024 bits 0.003521s 0.004292s 284.0 233.0
dsa 2048 bits 0.011990s 0.014306s 83.4 69.9
The T2k looks nice for short keys, but does it really drop from 13k/sec to 1.6/sec for the 2kbit keys, or is that a display bug?
If not, does that mean that the ncp in the T1 cpu is limited to 1kbit keys, and if I intend to use longer DSA keys, I better stay doing it on normal CPU core?
I see the same huge performance drop in RSA when switching from 2048 to 4096 bit key sizes. Ok, RSA isn't really that popular anymore.. but even DSA-1024 is aged a bit and many crypto researchers recommend the 2kbit key sizes already, at least for applications that are supposed to be running for a few years.
Are there any improvements being made in the new T2 cpu series, compared to the T1 I ran the tests on?
- The T1/T2 brochure also lists functions like crc32, md5, sha1 hashing algorithms in the crypto cores - how can I make use of those? I've got a few applications that do heavy checksumming and file hashing, and I'd love to offload these operations to the crypto cores; however I couldn't yet find any tool that allows me to do that. The regular md5sum binaries coming with Solaris still do the math on the main cpu core.

I'd love to, but I don't have the money for a T2 system, and even if I did, I don't have the time to play with it right now anyways :(

Memory Usage with new Tab Apps and Panoramas

I actually don't have a problem, just a question. I love the new tab apps and panorama features in FF4, but I am wondering how the CPU usage compares with e earlier versions of FF. (I never checked and I'm not going to :)) Are there any recommendations on how tab apps and panoramas should be used? My assumption is that the "dormant" panos will use less cycles than if they were all open on the desktop, is this true?

CS3 is still 32-bit so the user address space is 3.5GB. However, the OS can still use additional memory for cache to help performance, which comes into play when working with larger files. You need to download and install the enable VMBuffer plug-in, otherwise VMBuffering is disabled.
See: http://www.crn.com/software/197700697
Adobe forums:
http://www.adobeforums.com/cgi-bin/webx?14@@.3bc45977/5
some benchmarks:
http://www.barefeats.com/octopro1.html
http://www.barefeats.com/octopro4.html

Intel HM77 Express/C210 Series Chipset using eSATA port multiplier

I have just purchased a Zotac Nano computer which uses Intel HM77 Express chipset and has an eSATA port. I was surprised to notice my 8-bay hard disk enclosure did not work on this machine. I have read online that the Intel chipset supports eSATA port multipler in hardware but that the Windows drivers do not have it supported. However, some reports are that it should work under Linux.
Does anyone have any idea about getting the on-board eSATA port to work with port multipliers?
Many thanks,
Graham

see the following doc:
Acquiring from GigE Vision Cameras with Vision Acquisition Software - Part I
http://www.ni.com/white-paper/5651/en
Compares the Universal to High performance driver,
specifically both provide the same max bandwidth of 125MB but the high performance driver takes advantage of the intel pro1000 chipset to offload some of the CPU load, again this is less important as processors become faster.
"Advantages
While the high performance driver provides better CPU performance compared to the universal driver, it will only work on Gigabit Ethernet cards with the Intel Pro 1000 chipset. The universal driver can work on any Gigabit Ethernet card recognized by the Operating System. CPU usage during acquisition will affect the CPU cycles available for image analysis. So if your application requires in-line processing, it is better to invest in a Gigabit Ethernet card with the Intel Pro 1000 chipset like the NI PCIe-8231. While your choice of driver affects the CPU usage, it does not limit the maximum bandwidth possible. Either driver will be able to achieve the maximum bandwidth of 125 MB/s."
Paul Falkenstein
Coleman Technologies Inc.
CLA, CPI, AIA-Vision
Labview 4.0- 2013, RT, Vision, FPGA

Force oracle 9i installer to use green thread

I want to install Oracle 9i on freebsd but I was unable to launch the installer because of the java runtime envoronnment, for the Oracle 8i installer We had to specify the installer to use green threads but I was unable to find where to specify it to the 9i installer.
Any help would be welcome.
Thanks in advance

<How can I force oracle 9i to use more CPU >
Never seen that before; usually we're trying to get Oracle to use LESS CPU!
Comparing 10g and 9i are difficult and probably an exercise in futility. 10g is a different database and will perform differently. If all of your initialization parameters are identical between instances, the OS is the same version with the same settings, and the hardware is identical then there's not a lot of things to look at. If any of these things are not identical they may or may not be a factor as to why the systems are performing differently.
The lower CPU on the 9i side suggests something else is slower, possibly data access. You could perform traces on both systems to see things like I/O and memory reads.

OATS - 12.1.0.1.0361 Issue with running a scenario

Hi,
I am doing a comparison of Hits/sec and cpu usage for a scenario with OATS and LR.
I have remote load generator and app server.
With OATS, i am getting almost half the hits/sec and cpu as compared to lr.
Till now what i have noticed is that the time for the transaction in OATS is higher than that of LR. But the transaction contains one single URL only. So am not sure what is causing this increase.
I have checked the server response content size and it is same in OATS and LR.
Logging and think time are disabled.
I would like some help with this issue.
Thanks

this thread is closed

Streaming with H.264 (jiggle)

Hello,
After streaming with Flash Media Encoder 3.1 and ManyCam/Splitcam with VP6 and 320x240,30fps (works good!!),
I tried to stream HQ with H.264,30FPS,640x360,700kbps...
Buter after pressing start,the both Windows (Input and Output) will jiggle very strong.(In the Justin.tv stream,too,of course)
Looking in the Encoding Log,it shows me an Output with 11-12 fps.
But where is the Problem? (PC,Internet...?)
My System:
Laptop Core 2 duo T5250@1,50GHz
3GB RAM
Windows Vista
enough Harddisk
ATI Mobility Radeon X2300
DSL 20.000/1.024
Thanks a lot =)

Ithink the problem is that the H.264 codec needs more CPU resources compared to ON2 VP6.
I tried 640x480, 25fps, 200Kbps with the 2 codec (VP6, and h264). I have the H.264 codec with a delay of 3 seconds and a delay vp6 less than 1 second.
I think these should you need more CPU resources with H264
the cpu works a lot with h.264 and this makes for more delay
http://help.adobe.com/es_ES/Flash/10.0_UsingFlash/WS9222D73A-676D-41cd-9222-A4884858BBA3.h tml
Enrique Figueroa

CPU schedulers compared (bfs vs cfs)

Similar Messages

Maybe you are looking for