Page: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
<Setok> might be soon if I can't work out how the hell to get this project done
<Setok> tiredness is the best
<ijchain> <gwlester> Next you are going to say "I'm feeling much better"
<Setok> you can just BS endlessly when tired
<Setok> damn, I don't want Apple to go Intel
<ijchain> <rmax> torsten: that's how I went last year when I was listening in from home.
<rkeene> And then there's the random crap I work on for the fun of it :-P
<Setok> Intel is so gray and boring
<gav-bomb> hi guys
<gav-bomb> is openACS dope?
<ijchain> <gwlester> hello
<ijchain> <stever> stevel: exactly which compilers support the whack flag?
<Setok> with PPC there was always this little hope that some day it would kick everyone else's arses
<rkeene> Intel is grey and boring ?
<Setok> and every now and then it did
<Setok> rkeene: yeah
<ijchain> <rmax> When the conference was over I had to recover from a jet lag without really having travelled to another time zone.
<rkeene> Do you mean the chips or the i80x86 instruction set ?
<Setok> and I'll probably kill myself if an Apple ad has that Intel chime in it
<Setok> rkeene: everything really. It's so mainstream
<Setok> and the instruction set is better not to mention
<Setok> but normally we don't see that
<rkeene> Mac OS X on i80x86 is notably faster than Mac OS X on PPC.
<ijchain> <stever> fed gone
<ijchain> <stever> feed
<ijchain> <torsten> no stream.
<Setok> rkeene: since when? I find it hard to believe seeing as the G5 smokes the P4 at any real benchmarks
<rkeene> Even Mac OS X on i80x86 emulating PPC is faster than Mac OS X on PPC at times.
<ijchain> <torsten> is this the sign to go sleep now?
<patthoyts> silence rains :/
<ijchain> <aku> stream is down
<ijchain> <kbk_> wilma, fortunately is done raining (or rather, sleeting) here.
<ijchain> <stever> probable anothe web authentication timeout
<ijchain> <stever> I hate those systems
<rkeene> Setok, http://www.thinksecret.com/news/0506intelxbench.html
<Setok> I mean, I don't know what they're doing with x86 OS X, but the chips themselves atm beat everything Intel has
<Setok> according to SPECbench and stuff
<ijchain> <stever> gwlester: timeout again?
<Setok> the problem are the laptops really. G4 just isn't cutting it
<ijchain> <torsten> network seems down from hotel. all folks there leaving chat ---
<ijchain> <willdye3> Whoah. SnackAmp just freaked out. Lost of progress bars, then when I hit the stop button, it gave me a Tcl error.
<ijchain> <stever> willdye3: now you can fix the bug :)
<ijchain> <willdye3> can't read "state(http)": no such element in array while executing "lindex $state(http) 1" (procedure "playurl::cleanup" line 6) invoked from within "playurl::cleanup sock680 ::http::60" ("command" script)
<cleverly> hotel had a network hiccup
<ijchain> <willdye3> stever You think I should submit that?
<Setok> rkeene: hm, that page is saying rosetta/x86 is clearly slower than a G5
<cleverly> "what platforms does it run on?" (gwlester)
<ijchain> <stever> Id expect the author should catch network dropouts more graceful
<rkeene> Setok, Except on the few benchmarks where it beat the Dual-G5.
<Setok> even non-rosetta, apparently
<cleverly> "can it handle menus?" (bryan o.)
<ijchain> <aku> stream is back
<ijchain> <kbk_> I suspect that the bug is in the http package rather than snackamp. The http package is in major need of work
<cleverly> "what did you do on the windows to be able to automate the menus?" (bryan o.)
<Setok> rkeene: the graphcis ones? They're not CPU-intensive but gfx card intensive
<ijchain> <gwlester> stever, yeap lost connection and just got it back
<rkeene> Setok, The memory ones.
<ijchain> <willdye3> Is it back up now? SnackAmp still can't connect.
<patthoyts> tclers.tk has yet to restart.
<cleverly> "is there a url for it clif?" (gwlester)
<Setok> rkeene: only one, stream, by a bit. system memory test was much faster on a G5
<Setok> like much much faster
<rkeene> Setok, Though the page doesn't mention any non-PPC binaries, the benchmarks are all PPC binaries (running through Rosetta on x86)
<ijchain> <rmax> we got kicked off the internet again here :(
<ijchain> <aku> we noticed :P
<ijchain> <willdye3> torsten Later
<cleverly> http://www.dedasys.com/~clif is the URL to look for Clif's TkTest
<rkeene> Setok, But you have to remember the intel system is doing translating all this so it's having to do much more work. The fact that it ever wins is good.
<ijchain> <torsten> bye, have a nice evening in Portland!
<ijchain> <willdye3> Restarted SnackAmp (dunno if that mattered), and now it's working.
<ijchain> <rmax> 'nite torsten
<cleverly> try #3 for richard hipp
<patthoyts> euro feed ok now.
<Setok> rkeene: sure, I wouldn't expect rosetta to be faster really, but that doesn't prove my claim false :)
<Setok> I mean, I was just comparing SPEC CPU results
<Setok> the new G5 beats anything from intel that is listed there
<rkeene> Setok, Except for the interesting cases where the system that should be slower is not slower.
<Setok> and is about on par with opteron. For SPECint and fp. Didn't check the rate as I don't know them for dual-g5
<Setok> rkeene: yeah, dunno what happens there (but remember that rosetta translates once, so with certain tests you might easily get comparable performance)
<Setok> still, the overall picture is fairly clear
<Setok> not that rosetta does badly at all
<Setok> I'd have expected it to do worse. Clearly the technology is not bad
<ijchain> <kbk_> clicky : http://tkhtml.tcl.tk/
<rkeene> I can't even find Apple in the latest SPEC results.
<rkeene> What was its base/peak ?
<Setok> hm, I wonder if tkhtml is maintained
<ijchain> <miguel> /darw
<Setok> rkeene: they're not listed and I'm not sure if they are base or peak. What would IBM normally announce? Base or peak?
<ijchain> <aku> 3dcanvas quesiton - could POVRAY be translated to 3dcanvas API ?
<Setok> anyway.. results (predicted, not yet submitted I think) are:
<cleverly> "does the tkhtml widget work on all platforms?" (bryan o.)
<ijchain> <aku> The general way of describing things seems to be same
<ijchain> <aku> I.e. could 3dcanvas be used as preview for povray scenes ?
<Setok> 1,438 for SPECint
<Setok> 2076 for SPECfp
<cleverly> "is it under a bsd license?" (gwlester)
<ijchain> <GPS> aku: does it do constructive solid geometry? so you can combine shapes to form other shapes, and subtract a shape's area from another?
<ijchain> <aku> I have no idea - these are questions for drh
<rkeene> Intel beats them in atleast SPECint with 1810/1821 (base/peak)
<ijchain> <kbk_> Does the Eolas-sponsored work contain patent bombs?
<Setok> oh yes, with SPECint they would
<rkeene> 3.8 GHz Xeon, 2MB L2 Cache
<rkeene> Checking SPECfp now
<Setok> they do benefit there from the high frequency
<ijchain> <GPS> who's on now? I only just reconnected.
<ijchain> <aku> drh
<ijchain> <stever> richard hipp
<ijchain> <willdye3> URL mentioned just now: http://3dcanvas.tcl.tk/
<ijchain> <aku> richard hipp of sqlite
<cleverly> "looking at the requirements -- producing postscript output, do you know what the status of that is?" (tom a.)
<ijchain> <GPS> thanks
<Setok> I think the 3.8 Xeon beats the opteron too with int
<rkeene> Ye olde dell poweredge 750 gets a 2545 in SPECfp
<rkeene> 7250, rather
<Setok> but of course in digital media, fp is probably more important. with games, vice versa
<Setok> rkeene: what kind of machine is that? Many of the high end intel workstations were getting clearly worse figures
<Setok> is that Itanium?
<Setok> Itanium is a totally different story
<ijchain> <gwlester> bye
<Setok> Itanium rocks at fp
<ijchain> <stever> applause
<rkeene> Intel Itanium 2
<ijchain> <willdye3> Extra thanks to rmax, cleverly, and all those helping us lurkers.
<Setok> different story
<Setok> itaniums are very nice for fp
<ijchain> <willdye3> miguel The audio is being recorded, I'm told.
<Setok> although there were a couple POWER5 which were even nicer
<rkeene> Itaniums are also very nice for SMP (significantly better than its nearest competitor)
<Setok> some of the IBM powers were doing over 3000
<Setok> they're mean beasts
<ijchain> <kbk_> What bofs are tonight, and is rmax going to stream any of them?
<rkeene> The highest Intel SPECfp I see is 2013/2016
<Setok> the g5 is basically a cut-down power5, I think
<ijchain> <dkf> I don't like this network
<rkeene> Intel Pentium, that is
<rkeene> (3.73GHz Intel Pentium4)
Return to tcl
Go to some related
how to display utorrent in irc