Phoronix IRC Log: 2008-10-08
Milyardo: Good Morning #phoronix
Milyardo: Hey michaellarabel, did any else confirm the issue with performace between different version of x.org/xserver yet? I had to goto class yesterday and haven't been able to work on them since
Milyardo: and since I have class today as well, I won't be able to work on them today either
michaellarabel: Not sure, haven't looked.
Milyardo: thats ok, I'll need time to do benchmarks anyways to make sure its not just my imagination
michaellarabel: Hi Deanjo
michaellarabel: Hi Ivanovic
Ivanovic: hi michaellarabel
michaellarabel: When you have the time, can you check out PTS git master? I committed hard drive detection support this morning and just want to make sure it's working for people.
Ivanovic: what do i have to test the drive detection?
michaellarabel: just run phoronix-test-suite system-info
michaellarabel: Disk: 210GB Maxtor 6B200M0 + 990GB SAMSUNG HD103UJ + 250GB SAMSUNG SP2514N Great :)
Ivanovic: jupp, looks good
michaellarabel: Except for the 1TB drive as 990
Ivanovic: and the 200gb drive for 210gb
michaellarabel: and the 200 as 210
maligor: does it show real size or marketing size?
michaellarabel: I am going after the marketing size
Ivanovic: and some of the other data is wrong: "OpenGL: 2.1 Mesa 7.2 (fglrx)"
Ivanovic: mesa 7.2 is correct
Ivanovic: but fglrx is not atm
michaellarabel: Do you have the fglrx module loaded?
Ivanovic: yeah, it is loaded
Ivanovic: seems to get auto loaded just when it is available on my system
Ivanovic: and do not ask me why it is not able to detect the chipset though it is able to detect the motherboard
Ivanovic: might be something permissionrelated with the progs used for that part of the detection
michaellarabel: If you git pull, hopefully the size detection is correct now
Ivanovic: Disk: 220GB Maxtor 6B200M0 + 1000GB SAMSUNG HD103UJ + 260GB SAMSUNG SP2514N
michaellarabel: Well, 1TB is correct :)
Ivanovic: yeah, the two others are not...
Ivanovic: afk, got to hunt for lunch/dunner
michaellarabel: Ivanovic when you get back I made another commit
Milyardo: michaellarabel: I just pulled the lastest from git and ran =====================================
Milyardo: michaellarabel: I just ran phoronix-test-suite system-info from lastest git and it returned no information about HDDs
Milyardo: Also you can visit http://zliog.ath.cx/phpsysinfo for purposes of comparision
Milyardo: My Drives are both IDE but in phpsysinfo they're listed as SCSI, does PTS make any assumptions about interfaces?
michaellarabel: can you send me your dmesg output?
michaellarabel: or pastebin it
maligor: Milyardo, SATA?
maligor: all the hdd interfaces are moving to the scsi subsystem
maligor: pata stuff is experimental tho
Milyardo: Yeah but my drives surely IDE
Erektium: I get the disk correctly but. motherboard and chipset are wrong.. http://pastebin.com/m718c2821
Milyardo: Hrrmmm now that I look at my output my motherboard and chipset is wrong too
Milyardo: I can't have a AMD K8 if I have an Nvidia 8200 o_o
Deanjo: What processor do you have?
michaellarabel: Milyardo: Try git again, I think yours should be fixed
Milyardo: Disk: 40GB ST340014A
michaellarabel: Great, just had to add an entry to read for ATA-6.
Deanjo: Disk: 2x 500GB ST3500320AS + 2 x 500GB Maxtor 6L250S0
Deanjo: Doesn't detect all my drives and capacity on the 250S0's is wrong
redeeman: ah lol
redeeman: so apparently the new nvidia driver
Deanjo: The 250S0's are running on dmraid
redeeman: does not make xrender usable speeds at all
Milyardo: redeeman: You were have performace issues as well?
redeeman: not me personally no, this is a friends box
redeeman: im just saying
redeeman: nvidia says they fixed it
michaellarabel: Deanjo: I have yet to touch anything to do with calculating RAID. Right now all of this disk information (since hdparm, fdisk, dmidecode, etc require root) is through dmesg
redeeman: they haven't fixed shit
Deanjo: Well it works fine here and nobody is bitching in their forums redeeman . PEBKAC
redeeman: yes so tell me, how exactly might one use it wrong?
Deanjo: More then likely, (as is usual in cases like that) the person installed without uninstalling the previous first
redeeman: i garantuee you this is not the case
Milyardo: WHat version of xorg/xserver were they using?
redeeman: this guy is using 7.2
Milyardo: I observed a discrepancy in performance with 7/4/1.5 but I don't think it was as bad as before(though I have yet to benchmark this).
Milyardo: Yesterday with 7.2 I did see an improvement (again no benchmark :( )
Ivanovic: michaellarabel: getting closer...
Ivanovic: Disk: 200GB Maxtor 6B200M0 + 990GB SAMSUNG HD103UJ + 250GB SAMSUNG SP2514N
michaellarabel: So just the 1TB drive is off now?
michaellarabel: Try now, found a typo
Deanjo: longs for last weeks hot weather
Deanjo: thinks there are gonna be a lot of pissed off DDR3 customers when they find out their performance ram can't be used on a i7
maligor: heh, i7 is ddr2?
Deanjo: No, it just can't use high voltage ram
michaellarabel: maligor: High voltage DDR3 RAM can kill the i7 CPU
Deanjo: You can fry the processor by using ram that requires voltages greater then 1.65
maligor: heh, that'll REALLY piss off some people
maligor: good to know tho
Deanjo: Especially when they paid premium dollar for that ram
maligor: and premium dollar for the cpu too
maligor: and mb
Exopaladin: RAM and CPU voltages are locked in sync?
Deanjo: heh, might be enough for people to switch back to the AM3's systems
maligor: they haven't been released yet :P
Deanjo: back to AMD that is
Exopaladin: I'm sticking with my C2Q setup for now
Deanjo: but at least their ram would work in it
maligor: oh, they're going to limit it
maligor: or add stickers
Deanjo: I imagine that limitation will be resolved in a future revision of i7
Deanjo: Never buy first gen architectures
maligor: unless it's a fundamental design problem
maligor: which would mean they'd have to spend a few more dollars on redesign
maligor: quite unlikely tho
Deanjo: Not surprising to see items like this on i7 being that it's intels first kick at the cat with a integrated memory controller
maligor: no, I don't think it is
maligor: pretty sure xscale had one
Deanjo: 1 gen opterons required Registered ram when they first came out
Deanjo: 1 gen quirks. It happens all the time.
maligor: having a separate memory controller on a embedded arm atleast sounds a bit silly
Deanjo: OK 1st kick at a integrated memory controller on a x86 cpu
Deanjo: Until this day I can't say I actually consider ARM's a "CPU" per say
Deanjo: heh, you know people that run port scans should really not be running XP without a firewall, at least a router, and have file and printer sharing turned on while doing their script kiddie crap.
Deanjo: Heh, I love printing stuff off on their printers
Deanjo: I think 100 pages of "People who try to hack should learn about securing their own system first" should suffice in this case.
Deanjo: Just about as bad as the 9 out of 13 routers I can see with my wireless that are still running default settings with the default login and pass.
maligor: why wouldn't arms be cpus?
maligor: you mean they're more like microcontrollers?
Deanjo: Ya I put them in the same class. It's a mental thing.
maligor: I suppose
maligor: just very fast microcontrollers
Deanjo: It's like I don't consider my routers as having CPU's either.
maligor: yes, they have PPU's, Packet Processing Units
Darklock: so Linux is not an OS either because it runs in those routers? :-D
maligor: some distinctive features I've seen of microcontrollers were that they had flash and ram ondie
maligor: and stuff like ADCs
Deanjo: It all depends on how you view what a CPU is I suppose. S0C's I really don't consider CPU's wither
Deanjo: Damn, this new opera seems even quicker then 9.5
Deanjo: Still doesn't pass the acid 3 test though, but still better then FF
Deanjo: I wonder if google is actually going to port chrome or do their typical wine wrapper crap
Darklock: FF nightlies are at 97 I think
Exopaladin: dunno, there's at least some code in the repository, so I'd assume they're working on it
Deanjo: odd, I've got last nights FF here and it gets 79
Darklock: right, just looked again, it was an experimental build with some unfinished patches
Darklock: err, http://ehsanakhgari.org/blog/2008-10-03/experimental-builds-latest-acid3-patches
Deanjo: meh the race is over anyways. Webkit again beat them to the punch
Darklock: yeah, no clue why you brought it up
maligor: beats them to what?
Deanjo: The Acid tests
maligor: while standards compliance is something to achieve, it's not exactly that important :P
Deanjo: It's all about the bragging rights baby!
Deanjo: Webkits ePenis is bigger then Mozilla's :P
maligor: and ie has the most penises in total
maligor: and it doesn't even pass acid -1
Darklock: exactly, where is WebKit's