**** BEGIN LOGGING AT Tue Jun 28 23:59:57 2005
**** BEGIN LOGGING AT Wed Jun 29 01:28:03 2005
Jun 29 05:35:34 <NAiL> Another slug joining the army :-)
Jun 29 05:35:53 <NAiL> "All in all, it's just another slug in the wall"
Jun 29 05:59:21 <VoodooZ_Work> hehehe
Jun 29 05:59:43 <VoodooZ_Work> NAiL: Funny you said that as I'm currently listening to Floyd!
Jun 29 06:01:17 <NAiL> Great minds think alike. Or something :-P
Jun 29 06:03:27 <VoodooZ_Work> yep
Jun 29 08:12:28 <jacques> anyone know the implications of this:
Jun 29 08:12:36 <jacques> [PATCH] ARM: 2759/1: Fix IXP4xx debug code (again)
Jun 29 08:12:41 <jacques> Patch from Deepak Saxena
Jun 29 08:12:46 <jacques> Accidently swapped the order of movne and orrne. Bad.
**** BEGIN LOGGING AT Wed Jun 29 11:10:53 2005
Jun 29 11:14:49 <NAiL> Anyone using mysql on the slug?
Jun 29 11:18:49 <sebasvv> i had experience under unslung but was slow
Jun 29 11:20:11 <NAiL> yeah, I know ;)
Jun 29 11:20:29 <NAiL> just wondering if the ipkg was still screwy. 
Jun 29 11:21:17 <NAiL> And it is. The paths are nutty, there's no init-script and it's compiled with support for unneeded stuff 
Jun 29 14:44:33 <VoodooZ> judging from the DS the only way to change those bits is to use 10K pull-downs on the appropriate pins.
Jun 29 14:45:12 <NAiL> And those pins are not exactly easy to reach?
Jun 29 14:45:33 <dyoung-zzzz> VoodooZ, did you mean to put that in a differnt channel?
Jun 29 14:46:34 <VoodooZ> oops. Yes, wrong channel.
Jun 29 14:46:36 <VoodooZ> sorry
Jun 29 20:55:34 <NAiL> Is it a good idea to require the root disk to be plugged into the "Disk 1" port? It'd make a lot of stuff easier :-)
Jun 29 21:08:50 <beewoolie-afk> NAiL: I thought that the determination of the root disk is made by a file in the root of the jffs2 partition.
Jun 29 21:11:56 <NAiL> yeah, but sda/sdb currently depends on which order you plug in the disks, or what port they are attached to at boot time
Jun 29 21:12:12 <NAiL> so .sda1root doesn't really make much sense
Jun 29 21:12:30 <NAiL> .usb1-1root makes more sense to me
Jun 29 21:12:53 <NAiL> So the first partition of the disk/flashstick in the first port would be bootable
Jun 29 21:13:45 <NAiL> Of course, you could specify usb2-1 as well, and leave the root disk in the second port.
Jun 29 21:15:49 <NAiL> It makes more sense to require the root disk in a specific port, than to hope that the disks are detected in the correct order
Jun 29 21:15:56 <beewoolie-afk> I don't claim to be proficient in this area.  IMHO, the whole scheme is kinda wacko.
Jun 29 21:16:08 <beewoolie-afk> I'd rather see us use a UUID for the volume.
Jun 29 21:16:37 <NAiL> Afaik, not all fses use UUID
Jun 29 21:16:37 <beewoolie-afk> UUID -> volume name
Jun 29 21:16:52 <beewoolie-afk> volume name->usage
Jun 29 21:17:00 <beewoolie-afk> Well, we're using ext2 which has it.
Jun 29 21:17:08 <beewoolie-afk> What volume type doesn't have UUID?
Jun 29 21:17:13 <beewoolie-afk> FAT has something.
Jun 29 21:17:19 <beewoolie-afk> Reiser?
Jun 29 21:17:20 <NAiL> fat has label only, afaik
Jun 29 21:17:26 <beewoolie-afk> Label is OK, too.
Jun 29 21:17:34 <beewoolie-afk> Actually, FAT does have a serial number.
Jun 29 21:17:47 <beewoolie-afk> We don't need to care about the form, just that we can link it.
Jun 29 21:17:56 <beewoolie-afk> besides, using fat as a boot filesystem is a really bad idea.
Jun 29 21:18:16 <NAiL> How do you find the UUID of a volume?
Jun 29 21:19:44 <beewoolie-afk> I think there is a utility for ext2...dumpe2fs.
Jun 29 21:19:58 <beewoolie-afk> dumpe2fs /dev/hda1 | grep UUID
Jun 29 21:21:08 <beewoolie-afk> There is an additional problem which is that the user may decide to build a new root filesystem and have a reasonable expectation that that filesystem will boot.
Jun 29 21:21:24 <beewoolie-afk> The selection process should be somewhat intelligent.
Jun 29 21:21:36 <NAiL> Yeah, and you have to iterate through *all* available partitions...
Jun 29 21:21:47 <beewoolie-afk> for example, there should be some sort of marker on the disk making it 'reasonable' for a root filesystem
Jun 29 21:22:00 <beewoolie-afk> If the last one isn't present, we choose the next reasonable one.
Jun 29 21:22:15 <NAiL> Why not make that marker "being plugged into port 1"?
Jun 29 21:22:16 <beewoolie-afk> If they insert two new ones, then we need to make a reasonable guess.
Jun 29 21:22:23 <beewoolie-afk> That works, too.
Jun 29 21:22:56 <beewoolie-afk> I've been wanting the MAC like named filesystem feature for some time.
Jun 29 21:23:14 <beewoolie-afk> It can operate along side the traditional device name scheme.
Jun 29 21:23:59 <beewoolie-afk> Seems like the boot filesystem selection could be quite flexible: port 1 or the UUID of the known filesystem.
Jun 29 21:24:09 <NAiL> yeah
Jun 29 21:24:36 <beewoolie-afk> Are you going to create devices for the port1 and port2 hard drives?
Jun 29 21:24:45 <NAiL> yeah
Jun 29 21:24:52 <beewoolie-afk> /dev/hd-port1-part1
Jun 29 21:24:56 <NAiL> Just make udev create them automatically as they appear
Jun 29 21:24:58 <beewoolie-afk> Or whatever.
Jun 29 21:25:15 <beewoolie-afk> Right, but you were saying that udev does sda first, no matter which port.
Jun 29 21:25:27 <NAiL> /dev/usbdisk[a-z][0-9]
Jun 29 21:25:43 <beewoolie-afk> Perhaps it should be /dev/uda1 for USB disk, port 1, partition 1.
Jun 29 21:26:00 <NAiL> yeah, uda sounds nice ;)
Jun 29 21:26:09 <beewoolie-afk> Hope I don't get flamed.
Jun 29 21:26:12 <NAiL> uda[a-z][0-9]
Jun 29 21:26:18 <beewoolie-afk> Maybe we should use a sub directory.
Jun 29 21:26:28 <beewoolie-afk> Why udaa0? instead of uda0.
Jun 29 21:26:36 <NAiL> my bad
Jun 29 21:26:39 <NAiL> ud[a-z][0-9]
Jun 29 21:26:45 <beewoolie-afk> Actually, the partitions tend to start at 1 these days.
Jun 29 21:26:51 <NAiL> hrrm.. yeah
Jun 29 21:26:54 <NAiL> ud[a-z][1-9]
Jun 29 21:26:57 <NAiL> :)
Jun 29 21:27:08 <beewoolie-afk> /dev/usbdisk/p1p1
Jun 29 21:27:37 <beewoolie-afk> I think it's really irritating how the devices start at 0, but the partitions at 1.  Makes grub a vicious PITA.
Jun 29 21:27:39 <NAiL> Dunno if a directory is needed. /dev isn't exactly crowded when running udev.
Jun 29 21:27:52 <beewoolie-afk> Understood.  It's just to isolate the namespace.
Jun 29 21:28:00 <beewoolie-afk> udev is a really cool thing.
Jun 29 21:28:37 <NAiL> Down from 567 files to 84.
Jun 29 21:30:02 <beewoolie-afk> There's more than that.  It can be very complicated to get the contents of /dev right.  udev means no more node creation.  It also means that there should need to be nodes in /dev/ when building filesystems for embedded systems.
Jun 29 21:45:21 <NAiL> Gotta look more at this tomorrow. Going to bed now. It's 6.45 am here ;-)
Jun 29 21:46:37 <beewoolie-afk> NAiL: Wait!
Jun 29 21:46:43 <beewoolie-afk> NAiL: you still there?
Jun 29 21:47:42 <NAiL> yeah
Jun 29 21:47:53 <beewoolie-afk> What is the reported bogomips for the slug?
Jun 29 21:48:01 <NAiL> 133.something
Jun 29 21:48:02 <NAiL> 2 sec
Jun 29 21:48:08 <beewoolie-afk> Guess what?
Jun 29 21:48:12 <NAiL> Calibrating delay loop... 133.12 BogoMIPS (lpj=665600)
Jun 29 21:48:18 <NAiL> nooooo
Jun 29 21:48:19 <beewoolie-afk> # cat /proc/cpuinfo
Jun 29 21:48:19 <beewoolie-afk> Processor       : XScale-IXP425/IXC1100 rev 1 (v5b)
Jun 29 21:48:19 <beewoolie-afk> BogoMIPS        : 263.78
Jun 29 21:48:19 <beewoolie-afk> Features        : swp half thumb fastmult edsp 
Jun 29 21:48:19 <beewoolie-afk> Hardware        : Intel IXDP425 Development Platform
Jun 29 21:48:19 <beewoolie-afk> Revision        : 0000
Jun 29 21:48:22 <beewoolie-afk> Serial          : 0000000000000000
Jun 29 21:48:26 <NAiL> wooooooo
Jun 29 21:48:37 <NAiL> You moved R64 ?
Jun 29 21:48:44 <beewoolie-afk> I haven't plugged in a USB device, yet.
Jun 29 21:48:50 <beewoolie-afk> It's easier than that.
Jun 29 21:48:54 <beewoolie-afk> I just removed R64.
Jun 29 21:49:06 <NAiL> Didn't add anything to R81?
Jun 29 21:49:14 <beewoolie-afk> I'm not sure I know how to test if it is really faster.
Jun 29 21:49:15 * NAiL pulls out his soldering iron
Jun 29 21:49:27 <beewoolie-afk> Nope.  all I did was remove R64.
Jun 29 21:49:50 <beewoolie-afk> Setting the clock bits to 0b000 seems to be enough.
Jun 29 21:50:01 <beewoolie-afk> dyoung is gonna go wild.
Jun 29 21:50:33 <NAiL> Getting a new slug in a few days. I'm gonna try the same with mine :)
Jun 29 21:51:28 <NAiL> That is, my current slug just became a sacrificial lamb :-P
Jun 29 21:51:40 <beewoolie-afk> Well, I sacrificed one and it lived.
Jun 29 21:51:46 <NAiL> Why does my soldering iron heat up so *damn* slowly?
Jun 29 21:51:57 <beewoolie-afk> Be careful, tho.  Do you have a wick?
Jun 29 21:52:28 <NAiL> No. If I miss I'll have to wait until 9 before I can get some
Jun 29 21:52:51 <beewoolie-afk> Miss?  I suppose you could use a vacuum, but you'd lose the part in the process.
Jun 29 21:53:21 <NAiL> Once I get it off, there's no way I can get it back without getting someone else to do it for me
Jun 29 21:53:26 <beewoolie-afk> I used copper braid to remove solder from both sides and then gentle reheated each side until the resistor moved.
Jun 29 21:53:45 <beewoolie-afk> It isn't really so hard to put the parts on.
Jun 29 21:53:58 <beewoolie-afk> I've been planning to buy some dental picks to make it easier.
Jun 29 21:54:09 <NAiL> With a soldering iron the size of a finger, it's hard :-P
Jun 29 21:54:55 <beewoolie-afk> Oooo.  I have a special soldering iron.  It's ESD safe, has a 1mm tip, and is temperature controlled.
Jun 29 21:55:31 <NAiL> part removed
Jun 29 21:55:41 <NAiL> It looks way smaller when it's on my desk
Jun 29 21:56:00 <NAiL> It looks like a speck of dust ;)
Jun 29 21:56:04 <beewoolie-afk> Nice work.
Jun 29 21:56:22 <beewoolie-afk> I noticed that.  That resistor is about 1.5mm by 2mm.
Jun 29 21:56:29 <NAiL> root@maimedslug:~# dmesg|grep MIPS
Jun 29 21:56:29 <NAiL> Calibrating delay loop... 266.24 BogoMIPS (lpj=1331200)
Jun 29 21:56:40 <NAiL> It booted a lot faster, tbh
Jun 29 21:56:48 <NAiL> ~hail beewoolie-afk 
Jun 29 21:56:53 <NAiL> duh
Jun 29 21:57:55 <ByronT-Away> overclocked slug?!
Jun 29 21:58:03 <ByronT-Away> am I reading that correctly?
Jun 29 21:58:17 <NAiL> yes, you are
Jun 29 21:58:20 <NAiL> and yes, it WORKS
Jun 29 21:58:34 <NAiL> It still hasn't heated up though
Jun 29 21:58:41 <NAiL> I'm pressing it a bit right now
Jun 29 21:59:11 <beewoolie-afk> go go go go go 
Jun 29 21:59:20 <NAiL> I'm gonna stay up a while longer ;)
Jun 29 21:59:28 <NAiL> Definitely a *lot* faster
Jun 29 21:59:37 <NAiL> Man, the slug just became a *really* neat toy
Jun 29 21:59:39 <beewoolie-afk> OK.  This is pretty darn cool.
Jun 29 21:59:45 <beewoolie-afk> heck yeah.
Jun 29 21:59:54 <beewoolie-afk> Better go buy a bunch before linksys gets wise.
Jun 29 22:00:03 <NAiL> Samba is faster :D
Jun 29 22:00:04 <ByronT-Away> <- trying to wake up core developers....
Jun 29 22:00:10 <beewoolie-afk> ByronT-Away: yeah.  Do.
Jun 29 22:00:18 <beewoolie-afk> All we've done is remove a single resistor, R64.
Jun 29 22:00:26 <NAiL> haha
Jun 29 22:00:35 <NAiL> A side effect is that the leds blink faster :P
Jun 29 22:00:41 <NAiL> But that might be my
Jun 29 22:00:42 <NAiL> me
Jun 29 22:00:53 <NAiL> Nah, just my imagination
Jun 29 22:01:22 <NAiL> But this is nothing less than brilliant
Jun 29 22:01:36 <NAiL> Still not warm to the touch
Jun 29 22:01:42 <ByronT-Away> need to do some benchmarking right now
Jun 29 22:02:11 <beewoolie-afk> How can I prove it is faster?
Jun 29 22:02:29 <beewoolie-afk> The bogomips are definitely doubled.
Jun 29 22:02:32 <NAiL> Well.. someone needs to do benchmarks with a "normal" slug :)
Jun 29 22:02:33 <beewoolie-afk> I have two slugs.
Jun 29 22:02:49 <beewoolie-afk> One is modified and one is not.
Jun 29 22:02:52 <NAiL> Perl's "make test" took 17 mins, afaik
Jun 29 22:02:58 <beewoolie-afk> However, one is running a newer kernel, too.
Jun 29 22:03:10 * ByronT-Away goes and fetches FrankenSlug.... you's getting a new modification...
Jun 29 22:03:18 <beewoolie-afk> I'm really curious if the network performance is improved.
Jun 29 22:03:24 <ByronT-Away> Well, a castration of sorts...
Jun 29 22:03:39 <NAiL> beewoolie-afk: It *feels* like it's going faster. Samba is using a lot less cpu anyhow
Jun 29 22:03:48 <beewoolie-afk> That's awesome.
Jun 29 22:03:55 <NAiL> I'm timing perl's "make test" right now
Jun 29 22:03:57 <beewoolie-afk> What about transfer times?
Jun 29 22:04:01 <beewoolie-afk> OK.
Jun 29 22:04:02 <beewoolie-afk> Excellent.
Jun 29 22:04:09 <ByronT-Away> empirical evidence!
Jun 29 22:04:38 <NAiL> If this works, I've gotta buy a lot more slugs :-D
Jun 29 22:05:30 <beewoolie-afk> If this is really twice the clock speed, we should see close to that in performance improvement.  The cache may have something to do with it, but not much.
Jun 29 22:05:53 <NAiL> Yeah
Jun 29 22:05:55 <beewoolie-afk> How's the temperature?
Jun 29 22:06:24 <NAiL> I'd say it's somewhere around 25-29 degrees. Still not very hot.
Jun 29 22:06:40 <NAiL> It's *far* away from being dangerously hot anyway
Jun 29 22:06:53 <beewoolie-afk> Are you using the ethernet port for the compile?
Jun 29 22:06:59 <beewoolie-afk> Or is this on a local disk?
Jun 29 22:07:08 <NAiL> local disk
Jun 29 22:08:05 <beewoolie-afk> ByronT-Away: any luck contacting other developers?
Jun 29 22:08:12 <NAiL> Right now, the slug is serving "Gorillaz - Feel Good Inc." while compiling perl and running top ;)
Jun 29 22:08:17 * NAiL feels good
Jun 29 22:08:36 <ByronT-Away> except for me, they are all asleep or away
Jun 29 22:08:55 <ByronT-Away> except for dyoung, who is supposedly getting a chewing out
Jun 29 22:08:59 <NAiL> hahaha
Jun 29 22:09:01 <ByronT-Away> or butt kicking
Jun 29 22:09:07 <NAiL> Well, he'll be happy when he gets back 
Jun 29 22:09:07 <beewoolie-afk> I thought that Tiersten might be up.
Jun 29 22:09:32 <ByronT-Away> a bit early for him, I think
Jun 29 22:09:46 <ByronT-Away> plus the fact that he didn't go to sleep until 3 hours ago
Jun 29 22:09:50 <beewoolie-afk> ah.
Jun 29 22:10:05 <beewoolie-afk> As far as I can see, it's dyoung who did the hard part on this project.
Jun 29 22:10:06 <NAiL> I've started editing http://www.nslu2-linux.org/wiki/HowTo/OverClockTheSlug, so you know. Won't save it until we know fer sure though 
Jun 29 22:10:10 <ByronT-Away> I would have thought rwhitby would have checked in...
Jun 29 22:11:37 <beewoolie-afk> NAiL: still building?
Jun 29 22:12:34 <NAiL> yeah
Jun 29 22:13:00 <NAiL> Actually, I think it's rebuilding something it's not supposed to. Might have to run it twice to see the real effect.
Jun 29 22:18:52 <NAiL> It's slowly heating up 
Jun 29 22:18:56 <NAiL> But still not body tem
Jun 29 22:18:57 <NAiL> p
Jun 29 22:24:47 <jbowler-zzz> Is it inside the case?
Jun 29 22:25:07 <beewoolie-afk> doubt that
Jun 29 22:25:08 <NAiL> no, and that wouldn't make much of a diff in my case. I've got Maimedslug.
Jun 29 22:25:24 <NAiL> Ok, now it's at body temp
Jun 29 22:25:52 <NAiL> Compiling lots of stuff though. Perl seems to be building lotsa stuff before running the test stuff
Jun 29 22:26:04 <NAiL> 3x stuff on one like. I must like stuff.
Jun 29 22:26:11 <NAiL> line, even
Jun 29 22:27:06 <NAiL> beewoolie-afk: Did you time your bootup?
Jun 29 22:27:21 <beewoolie-afk> There isn't any way for me to compare boot times.
Jun 29 22:27:36 <ByronT-Away> a watch
Jun 29 22:27:46 <NAiL> I'm pretty sure the kernel part took a lot less time. The blinky-led didn't blink for as long as usual.
Jun 29 22:27:55 <beewoolie-afk> Yeah.  I have a stopwatch, but I don't know how fast it used to be.
Jun 29 22:28:15 <NAiL> beewoolie-afk: Boot, measure time. Resolder R64, boot, measure time ;-)
Jun 29 22:28:30 <beewoolie-afk> yeah, right.
Jun 29 22:28:40 <NAiL> Should've measured stuff before I desoldered... 
Jun 29 22:28:42 <beewoolie-afk> You resolder it.
Jun 29 22:28:56 <beewoolie-afk> Why?  bogomips is higher.  That's all I care about.  :-)
Jun 29 22:29:00 <NAiL> With my >1mm tip? I don't think so ;)
Jun 29 22:29:20 <NAiL> The tip is almost larger than the component
Jun 29 22:29:30 <beewoolie-afk> I get email all the time about how to get a bigger tip...
Jun 29 22:29:35 <NAiL> hahaha
Jun 29 22:29:51 <NAiL> I don't need a bigger tip. A lot smaller tip though
Jun 29 22:30:04 <NAiL> (kept typing tit, which made even less sense)
Jun 29 22:30:09 <beewoolie-afk> That's not what the email says.
Jun 29 22:30:57 <beewoolie-afk> BTW, the main reason that packages don't cross compile is really bad autoconf garbage.
Jun 29 22:31:06 <beewoolie-afk> ...such as can be found in samba...
Jun 29 22:31:53 <NAiL> Temp is still in the comfortable zone
Jun 29 22:37:22 <beewoolie-afk> NAiL: do we have a new compile time?
Jun 29 22:37:58 <NAiL> It's still compiling. I'll have to rerun make test to get the results...
Jun 29 22:38:59 <NAiL> perl is the only thing I know some timings from :(
Jun 29 22:43:21 <NAiL> pausing perl build. Transferring a large file to see if the disk performance is any good
Jun 29 22:45:41 <NAiL> For some reason, my debian box won't mount the samba share from my slug :(
Jun 29 22:46:29 <NAiL> smbfs: mount_data version 1919251317 is not supported
Jun 29 22:47:10 <NAiL> hmm.. disk performance looks about the same
Jun 29 22:49:26 <beewoolie-afk> I wouldn't expect the disk to perform too differently.
Jun 29 22:49:41 <beewoolie-afk> I *would* expect that network traffic would be improved.
Jun 29 22:49:58 <beewoolie-afk> Interrupt latency should be lower.
Jun 29 22:49:58 <ByronT-Away> because CPU% should drop some
Jun 29 22:50:07 <beewoolie-afk> And that, too.
Jun 29 22:50:35 <NAiL> yeah, disk performance is about the same
Jun 29 22:50:37 <beewoolie-afk> I HATE AUTOTOOLS
Jun 29 23:10:57 <beewoolie-afk> NAiL: you still here?  I thought you'd headed for the sleep chamber.
Jun 29 23:11:21 <beewoolie-afk> I just finished cross-compiling samba.  There are some errors in the autoconf scripts, but it does build, at least.
Jun 29 23:11:28 <NAiL> How could I sleep now?
Jun 29 23:11:51 <ByronT-Away> one word for you NAiL, "Ambien"
Jun 29 23:11:56 <ByronT-Away> :)
Jun 29 23:12:19 <beewoolie-afk> Tee hee.
Jun 29 23:12:44 <NAiL> Ambien?
Jun 29 23:13:47 <ByronT-Away> prescription drug for sleeping
Jun 29 23:13:57 <beewoolie-afk> http://www.ambien.com/
Jun 29 23:14:28 <NAiL> hahaha
Jun 29 23:14:51 <NAiL> If I'm gonna sleep, I need to get off the SlugDrug first :-D
Jun 29 23:15:25 <NAiL> wee, the slug I ordered has been sent
Jun 29 23:16:10 <NAiL> Neat. I might have it later today if I'm really lucky
Jun 29 23:16:29 <NAiL> Door-to-door business service is fantastic :)
Jun 29 23:17:38 <beewoolie-afk> Nite
Jun 29 23:17:43 <NAiL> nite :)
Jun 29 23:23:16 <NAiL> real    2891967m43.831s
Jun 29 23:23:16 <NAiL> user    28m44.070s
Jun 29 23:23:16 <NAiL> sys     1m2.550s
Jun 29 23:23:30 <NAiL> Hmm.. didn't know my slug had been up that long ;)
Jun 29 23:32:32 <NAiL> Anyone that can write a short C-program that does x iterations? So we can compile on both a modded and an unmodded slug
Jun 29 23:32:36 <NAiL> rwhitby:)
Jun 29 23:32:45 <NAiL> Processor       : XScale-IXP42x Family rev 1 (v5b)
Jun 29 23:32:45 <NAiL> BogoMIPS        : 266.24
Jun 29 23:36:20 <NAiL> s/compile/run/
Jun 29 23:37:53 <ByronT-Away> NAiL, he is "notified" ;)
Jun 29 23:38:00 <NAiL> ah, good :D
Jun 29 23:39:31 <NAiL> Anyone care to run a openslug binary for me?
Jun 29 23:40:51 <NAiL> (or who has a dev env set up, and can compile?)
Jun 29 23:41:53 <ByronT-Away> it's almost 2am for this sleepy head.... and I have to be up early...
Jun 29 23:41:58 <ByronT-Away> good night, NAiL
Jun 29 23:42:02 <NAiL> nite
Jun 29 23:43:41 <NAiL> Time spent finding 9009 digits of e:
Jun 29 23:43:42 <NAiL> real    0m26.250s
Jun 29 23:43:42 <NAiL> user    0m26.140s
Jun 29 23:43:42 <NAiL> sys     0m0.070s
Jun 29 23:44:18 <NAiL> (ok, it's obfuscated c, but it should work to benchmark)
Jun 29 23:45:01 <jbowler-zzz> Put it up on pastebin, I believe I can cross-compile
Jun 29 23:45:40 <jbowler-zzz> BTW, these are pull down resistors - removing R64 sets the A23:21 lines to b110, which is undefined.  Removing 
Jun 29 23:45:45 <NAiL> http://pastebin.ca/16449
Jun 29 23:46:21 <jbowler-zzz> R83 is probably a better bet - b011 which is the setting for 266MHz.
Jun 29 23:46:35 <NAiL> R83 is harder to remove than R64 :p
Jun 29 23:47:28 <NAiL> btw, compiled without any optimizations.
Jun 29 23:48:20 <NAiL> with -O2 it's down to 19 secs
Jun 29 23:49:45 <jbowler-zzz> Command exited with non-zero status 9
Jun 29 23:49:46 <jbowler-zzz> real    0m 52.80s
Jun 29 23:49:46 <jbowler-zzz> user    0m 52.22s
Jun 29 23:49:46 <jbowler-zzz> sys     0m 0.02s
Jun 29 23:49:52 <jbowler-zzz> gcc, no options.
Jun 29 23:49:55 <NAiL> Neat-O :D
Jun 29 23:50:07 <NAiL> My slug is faster than yours :-P
Jun 29 23:50:19 <ByronT-Away> NAiL's looks exactly twice as fast...
Jun 29 23:50:32 <NAiL> yup
Jun 29 23:50:45 * NAiL does the happy dance
Jun 29 23:53:00 <jbowler-zzz> Just to double check try this with gcc -Os:
Jun 29 23:53:03 <jbowler-zzz> int main(void) { register volatile int i; int j; for (j=0;j<266000000;++j) ++i; return 0; }
Jun 29 23:53:13 <jbowler-zzz> 12.02s user on my slug
Jun 29 23:54:31 <jbowler-zzz> Stupid gcc ignores the register, never mind.
Jun 29 23:54:38 <NAiL> real    0m6.012s
Jun 29 23:54:38 <NAiL> user    0m5.990s
Jun 29 23:54:38 <NAiL> sys     0m0.020s
Jun 29 23:54:44 <NAiL> oh, heh
Jun 29 23:55:33 <jbowler-zzz> That's the right number - 6 clock cycles:
Jun 29 23:55:42 <jbowler-zzz> .L5:
Jun 29 23:55:42 <jbowler-zzz>         ldr     r3, [sp, #0]
Jun 29 23:55:42 <jbowler-zzz>         subs    r2, r2, #1
Jun 29 23:55:42 <jbowler-zzz>         add     r3, r3, #1
Jun 29 23:55:43 <jbowler-zzz>         str     r3, [sp, #0]
Jun 29 23:55:43 <jbowler-zzz>         bne     .L5
Jun 29 23:56:18 <jbowler-zzz> Five instructions but I think there isn't enough delay between the ldr and the add for the data to come down the pipe.
Jun 29 23:57:49 <NAiL> I'm satisfied with the fact that my times are half yours ;)
Jun 29 23:58:06 <NAiL> (actually, slightly less, but you might be running more stuff on yours)
Jun 29 23:59:00 <jbowler-zzz> .04s in my slugtime
Jun 29 23:59:35 <jbowler-zzz> Explained by the time to load the program I think.
Jun 29 23:59:44 <NAiL> ah, might be
Jun 29 23:59:54 <NAiL> ka6sox: Good morning. Got the good news? :)
**** ENDING LOGGING AT Wed Jun 29 23:59:56 2005