Re: [Xenomai-help] linux 2.6.11
On 21/04/12 23:03, Gilles Chanteperdrix wrote: On 04/21/2012 10:12 PM, xenophile wrote: Hi, I am trying to see why a trivial driver reading an ADC on an EP9312 ARM board has slowed by a factor of three. Since I installed it under 2.6.29 a couple fo years ago. It's looking like this is due to a kernel update I did in the mean time to 2.6.32 If the kernel is patched with Xenomai, did you check that you have the same setting for FCSE ? Other than that, you can have an idea of how the time is spent in both cases using the I-pipe tracer. I'm interested in exploring RT on this hardware and part of the change of kernel was to move to more recent kernel which has better support for this board and to find a kernel that has been used RT with this type of hardware so that I don't have to start from scratch. Yesterday I tested 2.6.33 and it was equally slow. I need to find out if this is due to kernel version or gcc. (4.3.4) I need to resolve the cause of this factor of 2 or 3 slow down before I get involved in xenomia patching so as to start from a sound base. Would there be any value in trying to rebuild a 2.6.11 based system for realtime work ? xenomai comes with a latency test that you can run on any two configuration to compare their latencies. If you go on the adeos project download area: http://download.gna.org/adeos/patches/ You will find kernel patches for versions such as 2.6.13 or 2.6.14, they are supposed to work with xenomai latest stable release. You can even find patches for 2.4 kernels for i386 and ppc. Chances are however that these patches have bugs which have been fixed since then (after all, 2.6.11 is 7 years old). Thanks, I was hoping to find some comments on which kernels are more responsive. I'm sure there's good and bad since RT is not one of the main priorities for kernel development, this may not be a case of the most recent being the best. From a simple user experience circa 2.6.11 was hugely more responsive than the more recent offerings. Maybe 2.6.13 or 2.6.14 are close enough to be in the same league. I would have thought this sort of thing would have been thoroughly investigated by RT people and would be documented. regards. ___ Xenomai-help mailing list Xenomai-help@gna.org https://mail.gna.org/listinfo/xenomai-help
Re: [Xenomai-help] linux 2.6.11
On 04/21/2012 10:12 PM, xenophile wrote: Hi, I am trying to see why a trivial driver reading an ADC on an EP9312 ARM board has slowed by a factor of three. Since I installed it under 2.6.29 a couple fo years ago. It's looking like this is due to a kernel update I did in the mean time to 2.6.32 I have just built a new toolchain and rebuilt the kernel to 2.6.33 and the result is about the same. I recall a number of years ago running a 2.6.11 kernel and the system was incredibily responsive. I could be building kde-libs, downloading at full speed browsing and listening to mp3 decoded music all without the slightest lack of responsiveness in the browser not a glitch in the mp3 playback. It was truly impressive. Now I find if I am rebuilding the toolchain , the browser or any other window can take over a second to respond to a redraw ( 2.6.32 on this machine). What premption mode does the kernel use? CONFIG_PREEMPT should be enabled. Now I've seen comments on audio forums that there was nothing better than linux 2.6.11 for real time audio work. I'm wondering whether there is not a general degradation in the linux kernel as time goes on. It certainly seems to be getting forever bigger and slower. Yes, there is definitively a degradation which will hurt especially low-end systems. Just compare the size of the kernel. It's getting bigger and bigger. Also the timer now runs at 250 Hz and there are many more features, interfaces, etc.. You may regain some degradation by optimizing the kernel configuration. On low-end systems, if it's an option, choose a 2.4 kernel. Would there be any value in trying to rebuild a 2.6.11 based system for realtime work ? Well, then I would prefer 2.4.x but for both Xenomai support (for your board) is not available, I guess. But normally it's not an option as people do want more recent kernel features and drivers. But all this does not explain a factor of 3 with your driver. I doubt that it's due to a move from 2.6.29 - 2.6.32. Try using ftrace to find out what's going on (if it's already available). Wolfgang. ___ Xenomai-help mailing list Xenomai-help@gna.org https://mail.gna.org/listinfo/xenomai-help
Re: [Xenomai-help] linux 2.6.11
On 04/22/2012 08:23 AM, xenophile wrote: I'm interested in exploring RT on this hardware and part of the change of kernel was to move to more recent kernel which has better support for this board and to find a kernel that has been used RT with this type of hardware so that I don't have to start from scratch. Yesterday I tested 2.6.33 and it was equally slow. I need to find out if this is due to kernel version or gcc. (4.3.4) I need to resolve the cause of this factor of 2 or 3 slow down before I get involved in xenomia patching so as to start from a sound base. Well, given the way xenomai works, we do not depend much on the linux performances. So, if you want to compare something, it should be the xenomai latency differences between the two kernel versions. Using linux performance to get an idea of xenomai performance does not really make any sense. To give you just an example, at some point around 2.6.30, the linux kernel introduced threaded interrupts, which greatly increase interrupt latencies. Xenomai does not use threaded interrupts, so it is essentially unaffected by the change. Of course there are some indirect effect such as the size of the kernel, which has an impact on I-cache and so on the latency. Thanks, I was hoping to find some comments on which kernels are more responsive. I'm sure there's good and bad since RT is not one of the main priorities for kernel development, this may not be a case of the most recent being the best. From a simple user experience circa 2.6.11 was hugely more responsive than the more recent offerings. Maybe 2.6.13 or 2.6.14 are close enough to be in the same league. I would have thought this sort of thing would have been thoroughly investigated by RT people and would be documented. We run xenomai, so, what we investigate are xenomai performances. On a system running xenomai, the performance-sensitive job is done by xenomai, not by linux anyway. And we check between two releases that we do not get differences in performances. So, I am pretty sure that on at91rm9200 for instance (the platform I have which should be the closest to an EP9312), the xenomai user-space latency from 2.6.29 to 3.2.1 is around 220us. Besides, in terms of performance, the relation performance/version probably also depends on the platform which is running the kernel. So for instance, you may have found that 2.6.11 was good on x86, but maybe it sucked on ARM. Apart from that, as I already said, we provide you with: - the latest version of xenomai which should supports all past kernel versions - the tools to do the performance measurement. So, you have everything you need to do the comparison. -- Gilles. ___ Xenomai-help mailing list Xenomai-help@gna.org https://mail.gna.org/listinfo/xenomai-help
Re: [Xenomai-help] linux 2.6.11
On 04/22/2012 11:10 AM, Wolfgang Grandegger wrote: But all this does not explain a factor of 3 with your driver. I doubt that it's due to a move from 2.6.29 - 2.6.32. Try using ftrace to find out what's going on (if it's already available). Threaded interrupts were introduced around this time. And threaded interrupts DO have a significant impact on latencies. -- Gilles. ___ Xenomai-help mailing list Xenomai-help@gna.org https://mail.gna.org/listinfo/xenomai-help
Re: [Xenomai-help] linux 2.6.11
On 04/22/2012 01:24 PM, Gilles Chanteperdrix wrote: On 04/22/2012 11:10 AM, Wolfgang Grandegger wrote: But all this does not explain a factor of 3 with your driver. I doubt that it's due to a move from 2.6.29 - 2.6.32. Try using ftrace to find out what's going on (if it's already available). Threaded interrupts were introduced around this time. And threaded interrupts DO have a significant impact on latencies. But I agree: tracing with ftrace or the I-pipe tracer if using xenomai to see where the time is spent is the way to go. -- Gilles. ___ Xenomai-help mailing list Xenomai-help@gna.org https://mail.gna.org/listinfo/xenomai-help
Re: [Xenomai-help] linux 2.6.11
On 04/22/2012 01:34 PM, Gilles Chanteperdrix wrote: On 04/22/2012 01:24 PM, Gilles Chanteperdrix wrote: On 04/22/2012 11:10 AM, Wolfgang Grandegger wrote: But all this does not explain a factor of 3 with your driver. I doubt that it's due to a move from 2.6.29 - 2.6.32. Try using ftrace to find out what's going on (if it's already available). Threaded interrupts were introduced around this time. And threaded interrupts DO have a significant impact on latencies. Yes, but thread interrupts are normally not enabled/used by default. I'm not speaking about CONFIG_PREEMPT_RT. Well, do we speak about latencies? He just said reading an ADC on an EP9312 ARM board has slowed by a factor of three. I'm not even sure if it's related to Xenomai at all. xenophile ($#^?) , could you please be more precise. But I agree: tracing with ftrace or the I-pipe tracer if using xenomai to see where the time is spent is the way to go. Yep. Wolfgang. ___ Xenomai-help mailing list Xenomai-help@gna.org https://mail.gna.org/listinfo/xenomai-help
Re: [Xenomai-help] linux 2.6.11
On 04/22/2012 03:53 PM, Wolfgang Grandegger wrote: On 04/22/2012 01:34 PM, Gilles Chanteperdrix wrote: On 04/22/2012 01:24 PM, Gilles Chanteperdrix wrote: On 04/22/2012 11:10 AM, Wolfgang Grandegger wrote: But all this does not explain a factor of 3 with your driver. I doubt that it's due to a move from 2.6.29 - 2.6.32. Try using ftrace to find out what's going on (if it's already available). Threaded interrupts were introduced around this time. And threaded interrupts DO have a significant impact on latencies. Yes, but thread interrupts are normally not enabled/used by default. I'm not speaking about CONFIG_PREEMPT_RT. Well, do we speak about latencies? He just said reading an ADC on an EP9312 ARM board has slowed by a factor of three. I'm not even sure if it's related to Xenomai at all. xenophile ($#^?) , could you please be more precise. Yes, without answers to these questions, xenophile's post looks a lot like a troll attempt. -- Gilles. ___ Xenomai-help mailing list Xenomai-help@gna.org https://mail.gna.org/listinfo/xenomai-help
[Xenomai-help] linux 2.6.11
Hi, I am trying to see why a trivial driver reading an ADC on an EP9312 ARM board has slowed by a factor of three. Since I installed it under 2.6.29 a couple fo years ago. It's looking like this is due to a kernel update I did in the mean time to 2.6.32 I have just built a new toolchain and rebuilt the kernel to 2.6.33 and the result is about the same. I recall a number of years ago running a 2.6.11 kernel and the system was incredibily responsive. I could be building kde-libs, downloading at full speed browsing and listening to mp3 decoded music all without the slightest lack of responsiveness in the browser not a glitch in the mp3 playback. It was truly impressive. Now I find if I am rebuilding the toolchain , the browser or any other window can take over a second to respond to a redraw ( 2.6.32 on this machine). Now I've seen comments on audio forums that there was nothing better than linux 2.6.11 for real time audio work. I'm wondering whether there is not a general degradation in the linux kernel as time goes on. It certainly seems to be getting forever bigger and slower. Would there be any value in trying to rebuild a 2.6.11 based system for realtime work ? TIA, ___ Xenomai-help mailing list Xenomai-help@gna.org https://mail.gna.org/listinfo/xenomai-help
Re: [Xenomai-help] linux 2.6.11
On 04/21/2012 10:12 PM, xenophile wrote: Hi, I am trying to see why a trivial driver reading an ADC on an EP9312 ARM board has slowed by a factor of three. Since I installed it under 2.6.29 a couple fo years ago. It's looking like this is due to a kernel update I did in the mean time to 2.6.32 If the kernel is patched with Xenomai, did you check that you have the same setting for FCSE ? Other than that, you can have an idea of how the time is spent in both cases using the I-pipe tracer. Would there be any value in trying to rebuild a 2.6.11 based system for realtime work ? xenomai comes with a latency test that you can run on any two configuration to compare their latencies. If you go on the adeos project download area: http://download.gna.org/adeos/patches/ You will find kernel patches for versions such as 2.6.13 or 2.6.14, they are supposed to work with xenomai latest stable release. You can even find patches for 2.4 kernels for i386 and ppc. Chances are however that these patches have bugs which have been fixed since then (after all, 2.6.11 is 7 years old). -- Gilles. ___ Xenomai-help mailing list Xenomai-help@gna.org https://mail.gna.org/listinfo/xenomai-help