**** BEGIN LOGGING AT Sun Jun 27 02:59:56 2010 Jun 27 08:43:21 hi, anyone knows of an asm instruction (or instructions) to find out the L1/L2 (well no ARM with L3 cache yet) cache sizes on ARM? Or some way to read it from the kernel? Jun 27 08:45:01 similar to cpuid on recent x86 cpus Jun 27 15:16:16 lool: I'm about to give up, no matter what I do, wanna-build just fails to even produce the database, you did mention an alternative a while ago, a gsoc project, got any more info on that? Jun 27 15:23:54 markos_: the registers which can be used to get this information are described in ARM ARM, but they are not accessible from userspace Jun 27 15:24:37 markos_: kernel provides some of the cpu specific information via /proc/cpuinfo and /proc/self/auxv Jun 27 15:26:38 ssvb: cpuinfo doesn't carry this info though, actually I thought of auxv, but it's not straightforward and I thought I'd ask first Jun 27 15:28:06 markos_: yes, this information may be incomplete, for example auxv did not even report NEON availability before linux kernel 2.6.29 Jun 27 15:29:02 markos_: if you are really interested in having this information, pushing some patches to the kernel to get it may be useful Jun 27 15:30:06 markos_: btw, http://pastebin.com/3rvXx68s Jun 27 15:30:45 hm, I guess it depends on the platform Jun 27 15:30:53 what kernel version? Jun 27 15:31:16 2.6.21-omap1 Jun 27 15:31:47 strange, I get none of the cache info here (2.6.31.12.3-ER1-efikamx) on the imx515 Jun 27 15:31:53 looks like somebody is being lazy and is not keeping this code up to date :) Jun 27 15:32:32 markos_: pkern was working on a python-based wanna-build replacement Jun 27 15:32:38 ssvb: still on x86 and ppc, L1/L2 cache info can be found on /sys/devices/system/cpu/cpu0/cache/index# etc Jun 27 15:33:10 ssvb: would be nice to have it work the same on arm also Jun 27 15:33:18 lool: anything published? Jun 27 15:34:25 markos_: google "Debian Autobuilding Infrastructure Rewrite"; I don't know how complete the rewrite ended up being Jun 27 15:34:48 markos_: Ping pkern or HE on #debian-devel@oftc? Jun 27 15:35:30 markos_: I have only 'cpufreq' there, but not 'cache', anyway it is clearly a problem on the kernel side, preferably it should provide this information in a consistent way Jun 27 15:36:39 markos_: alternatively it would be interesting if the kernel could just provide access to cpuid related registers (or emulate their behavior) Jun 27 15:37:52 why do you need to know cache size? Jun 27 15:38:58 eigen library does block size configuration -or is able to do so- by matching matrix blocks according to the L1/L2/L3 cache sizes, it gets some extra % speed that way -at least on x86 Jun 27 15:39:29 there's cpuid on x86 which gives that info -amongst others Jun 27 15:41:31 a somewhat ugly solution would be to run a small benchmark to get this information in an experimental way Jun 27 15:41:57 that's too error prone Jun 27 15:41:59 iirc, fftw does something like this Jun 27 15:42:35 right, but in the worst case it will just impact performance, nothing else Jun 27 15:42:58 which is the point behind all this :) Jun 27 15:44:24 are you going to bug kernel developers about providing this information? Jun 27 15:44:45 if it's possible, I can use it, if not, well, eigen is only a linear algebra library, running mini benchmarks at startup isn't going to be accepted by the authors anyway Jun 27 15:45:18 ssvb: I might, time permitting Jun 27 15:45:41 tbh, I've never filed a bug report on the kernel before, but there's always a first time :) Jun 27 15:46:05 the info is already available anyway Jun 27 15:46:35 first I have to setup this this autobuilder setup Jun 27 15:46:45 s/setup$/ Jun 27 15:50:56 lool: I guess soyuz cannot be installed standalone... Jun 27 15:50:57 ? Jun 27 18:25:35 markos_: No, but we're working exactly on that Jun 27 18:25:59 markos_: Actually, according to the soyuz devs it's not that hard to run it standalone Jun 27 18:26:10 markos_: What you will miss is the Launchpad API Jun 27 18:26:16 and obviously the web UI Jun 27 18:26:54 markos_: I'm not sure where the work related to splitting soyuz out is tracked; it should be in a blueprint but I'm not sure where Jun 27 18:28:58 markos_: Ideally, ping james_w if you want to follow these discussions **** ENDING LOGGING AT Mon Jun 28 02:59:57 2010