You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@xalan.apache.org by Jacek Ambroziak <ja...@yahoo.com> on 2002/07/02 19:56:07 UTC
XSLTC and Sun's j2sdk1.4.0_01 problem
Guys,
as I've been benchmarking various XSLT solutions for
Java
I noticed that XSLTC takes forever (practically loops)
to complete the popular XSLTMark benchmark suite.
It runs OK on 1.4.0.
I am running Red Hat Linux 7.1, kernel 2.4.18.
XSLTC comes either from Sun's XML Pack Summer 02 or
Xalan 2.4D1
-- it makes no difference :-(
Is it a known problem? A VM bug? Am I missing
something?
The latest-and-greatest from Sun should work together
well...
If you know a solution, please write to me at
jra@ambrosoft.com
--Jacek
__________________________________________________
Do You Yahoo!?
Sign up for SBC Yahoo! Dial - First Month Free
http://sbc.yahoo.com
Re: XSLTC and Sun's j2sdk1.4.0_01 problem
Posted by Jacek Ambroziak <ja...@yahoo.com>.
> Hi Jacek,
>
> I couldn't reproduce this problem with 1.4.0_01 on
> Solaris 8.
I haven't run anything on Windows... in a long while;
just Linux. So the problems I reported affect
only 1.4.0_01 and 1.4.1-beta for Linux.
Others seem to have experienced problems with
this setup too. 1.4.0 is fine and quite fast.
--Jacek
__________________________________________________
Do You Yahoo!?
Sign up for SBC Yahoo! Dial - First Month Free
http://sbc.yahoo.com
Re: XSLTC and Sun's j2sdk1.4.0_01 problem
Posted by Tom Amiro <To...@Sun.COM>.
Hi Jacek,
I couldn't reproduce this problem with 1.4.0_01 on Solaris 8.
> XSLTMark 2.0 (Feb 28 2001) running on SunOS 5.8 sparc
> Java Runtime Environment is Sun Microsystems Inc. 1.4.0_01
> Current time is 14:45:56 EDT, 08 Jul 2002
>
> Results for driver XSLTCDriver
>
> prep exec
> Test Component iter ms ms KB In KB Out KB/s status
>
> alphabetize 100 1676 11525 1966 156699 6883.52 CHK OUTPUT
> attsets 100 1040 1273 121 25595 10100.94 CHK OUTPUT
> avts 100 513 1721 1966 115693 34183.43 CHK OUTPUT
> axis 100 981 470 35 6093 6520.11 CHK OUTPUT
> backwards 100 989 1867 260 26181 7081.44 CHK OUTPUT
> bottles 100 545 1119 4 113046 50514.42 CHK OUTPUT
> breadth 100 395 1103 415 20078 9290.14 CHK OUTPUT
> brutal 100 2684 1561 130 39990 12850.96 CHK OUTPUT
> chart 100 375 1244 121 40732 16420.3 CHK OUTPUT
> creation 50 935 1180 983 76042 32637.92 CHK OUTPUT
> current 100 1281 703 18 2998 2145.73 CHK OUTPUT
> dbonerow 10 648 3176 19621 89 3103.05 CHK OUTPUT
> dbtail 100 651 1247 1966 50390 20993.04 CHK OUTPUT
> decoy 100 482 5545 1966 197460 17982.59 CHK OUTPUT
> depth 100 399 619 588 37363 30656.1 CHK OUTPUT
> encrypt 10 1669 1566 196 1967 690.96 CHK OUTPUT
> functions 100 1090 7583 1966 61845 4207.55 CHK OUTPUT
> game 100 1450 605 260 4658 4065 CHK OUTPUT
> html 100 828 509 29 4863 4805.78 CHK OUTPUT
> identity 10 505 1329 1962 19626 8122.26 CHK OUTPUT
> inventory 100 1701 598 184 20205 17048.45 CHK OUTPUT
> metric 100 679 608 88 6582 5485.61 CHK OUTPUT
> number 100 458 802 18 7958 4973.71 NO REF
> oddtemplate 100 2344 861 16 1552 911.46 CHK OUTPUT
> patterns 100 801 2730 1966 197460 36525.08 CHK OUTPUT
> prettyprint 10 805 4557 196 14363 1597.57 CHK OUTPUT
> priority 100 731 296 91 5722 9820.56 CHK OUTPUT
> products 100 1096 696 196 4101 3087.47 CHK OUTPUT
> queens 0 0 0 0 0 ? SKIPPED
> reverser 100 632 1197 129 12939 5459.2 CHK OUTPUT
> stringsort 10 655 2902 1962 19333 3669.18 CHK OUTPUT
> summarize 100 1324 1065 712 2939 1714.53 CHK OUTPUT
> total 100 369 538 121 3408 3280.2 CHK OUTPUT
> tower 20 871 3462 1 27628 3990.35 CHK OUTPUT
> trend 100 2426 11756 176 80683 3439.09 CHK OUTPUT
> union 100 1619 683 16 1181 877.4 CHK OUTPUT
> xpath 100 2488 977 28 2021 1049.33 CHK OUTPUT
> xslbench1 100 752 905 36 72285 39956.84 CHK OUTPUT
> xslbench2 20 826 2173 2920 42801 10520.38 CHK OUTPUT
> xslbench3 20 1093 393 2920 503 4355.62 CHK OUTPUT
>
> Aggr. Results 100 458 802 18 7958 4973.71
> Arithmetic Mean 458 4973.71
> Geometric Mean 4973.71
> Harmonic Mean 4973.71
I'm not running dbnorm, but I doubt that has anything to do with it.
Have you tried it on any other OS besides Red Hat Linux 7.1?
Tom
Jacek Ambroziak wrote:
>
> Guys,
>
> as I've been benchmarking various XSLT solutions for
> Java
> I noticed that XSLTC takes forever (practically loops)
> to complete the popular XSLTMark benchmark suite.
> It runs OK on 1.4.0.
>
> I am running Red Hat Linux 7.1, kernel 2.4.18.
> XSLTC comes either from Sun's XML Pack Summer 02 or
> Xalan 2.4D1
> -- it makes no difference :-(
>
> Is it a known problem? A VM bug? Am I missing
> something?
>
> The latest-and-greatest from Sun should work together
> well...
>
> If you know a solution, please write to me at
>
> jra@ambrosoft.com
>
> --Jacek
>
> __________________________________________________
> Do You Yahoo!?
> Sign up for SBC Yahoo! Dial - First Month Free
> http://sbc.yahoo.com
--
Tom Amiro -- SQA Engineer
Sun XML Technology Development
voice: 781-442-0589 Fax: 781-442-1437
eMail: tom.amiro@.sun.com