Phoronix IRC Log: 2009-01-08
Soul_keeper: http://www.phoronix-test-suite.com/screenshots/pts-malvik-3.png what was run to generate those graphs in phoronix-test-suite ?
Soul_keeper: phoronix-test-suite MONITOR=all benchmark opengl-demos doesn't seem to do it for me
ndim: michaellarabel: http://fedorapeople.org/gitweb?p=ndim/public_git/upstream/phoronix-test-suite.git;a=summary AKA git://fedorapeople.org/~ndim/upstream/phoronix-test-suite.git
Soul_keeper: MONITOR=all phoronix-test-suite benchmark GLMark ok i figured it out
Soul_keeper: found some mention on the forums
ndim: ./glmark: line 6: 17296 Segmentation fault ./glmark $@ > $THIS_RUN_TIME.result
ndim: Hmm. The latest commit I get via "git clone http://www.phorogit.com/repo/phoronix-test-suite.git" is from 2009-01-04, while http://www.phorogit.com/index.php?p=phoronix-test-suite.git is showing commits from 2009-01-07?
ndim: Missing git-update-server-info post-update
Ivanovic: michaellarabel: cf what ndim wrote
michaellarabel: ndim: Should be fixed now. Sometimes the git server needs to be manually updated
Ivanovic: michaellarabel: s/sometimes/always
ndim: michaellarabel: And there are two more patches in my git repo. :)
michaellarabel: ndim: Yep, thanks. Committed the fixes to master.
ndim: goes look whether the update now worked.
ndim: Looks good :)
michaellarabel: Great :)
michaellarabel: Milyardo: How are the documentation patches coming?
ndim: "documentation" as in "contains the command line to run for just testing all the basic OpenGL stuff without user input"?
michaellarabel: ndim: Those capabilities should already be there
michaellarabel: say.... phoronix-test-suite batch-benchmark universe-x and there will be no user input after you run batch-setup once
ndim: The first part of the benchmark is a network download benchmark over several gigabytes, I guess.
michaellarabel: For the first time, yes, to procure all of the needed tests.
ndim: Is there any documentation which gives me any idea how much space I should have? 1GB? 10GB? 100GB?
michaellarabel: It says on a test-by-test basis and if it runs out of room or will, it will alert you. Though I think if you install *every* test right now it's like 16GB installed
ndim: takes a deep breath
michaellarabel: ndim: Well there's about 90 different tests...
Xemanth: any java guru here atm
Xemanth: http://java.sun.com/j2se/1.5.0/docs/api/java/lang/System.html#currentTimeMillis%28%29 i want to return current time in milliseconds and put it to label's value
Milyardo: Haven't worked on them today
Milyardo: I have work today
Milyardo: I think tomarrow I might head to school and do some Sysadmin work on the ACM servers and get alot of documentation done then
Milyardo: but tomarrow might also be my day to watch the my daughter
Milyardo: Where I still may have some time to do some documentation
Milyardo: depends on how fussy she is
EvilGuru: Xemanth: #java
EvilGuru: But you there is a wide collection of date/time classes in Java
EvilGuru: nanaTime is a timer
Milyardo: EvilGuru: He was inquiring about the currentTimeMillis() method, and how to convert it to a string
EvilGuru: Oh, Safari jumped to nanTime
EvilGuru: Converting it to a string isn't too difficult
EvilGuru: Also, I am looking through the review for the ATI 4870X2 and am wondering why there are no nVidia/lower end cards in the mx
EvilGuru: makes performance comparisons very difficult if only a single chip (RV770) is being benchmarked
Milyardo: EvilGuru: IIRC at the time nvidia had no comparable cards officaily supported by thier driver at the time
Milyardo: That the the cards didn't work with the driver
EvilGuru: Milyardo: There is always an 8800GTX
EvilGuru: Or any other single GPU card from nVidia, if it is directly comparable should depend purely on performance
Milyardo: 8800GTX comaprable to a 4870X2 IMO
Milyardo: *isn't comparable
EvilGuru: Under Windows I wouldn't think so, under Linux, lets see some FPS
EvilGuru: Maybe a GTX280, both are in the upper-high-end price bracket
EvilGuru: (On Windows they are often compared as such.)
EvilGuru: It is great to see performance reviews under Linux as opposed to Windows, but no one looking to buy a card will take them seriously unless a wide spectrum of cards are compared
bkero: There are many more factors which make that more difficult on linux, such as different versions of libraries, driver maturity, etc.
EvilGuru: bkero: Agreed, driver maturity affects windows benchmarks as well
EvilGuru: so long as the same system is used most libraries should be the same
bkero: Except for different versions of mesa, X, drm, etc.
EvilGuru: bkero: Last time I checked the nvidia drivers don't even use mesa/drm stack (could be wrong)
EvilGuru: Either way, I think one would have a very hard time arguing that taking a distro X (Ubuntu/Fedora) installing the ATI drivers with a set of ATI cards, running some benchmarks then swapping it for the nVidia drivers and an nVidia card and running some benchmarks is unscientific
EvilGuru: As it will reflect real world performance, which is what most people care about, potential performance in the future is all well and good, but what you can get today is what matters
Milyardo: Nvidia stable yum