-
Bug
-
Resolution: Cannot Reproduce
-
P4
-
None
-
1.3.0
-
x86
-
windows_nt, windows_2000
Name: boT120536 Date: 02/19/2001
C:\jdk1.3\bin>jdb -version
jdb version 99/06/11
C:\jdk1.3\bin>java -version
java version "1.3.0"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.3.0-C)
Java HotSpot(TM) Client VM (build 1.3.0-C, mixed mode)
---------------- and ----------------
C:\jdk1.2.2\bin>jdb -version
jdb version 98/06/23
C:\jdk1.2.2\bin>java -version
java version "1.2.2"
Classic VM (build JDK-1.2.2-001, native threads, symcjit)
I am working on a rather large Java application. When I use the jdb in 1.2.2,
and run the program, I see the first bit of output in 20 seconds. When I use
the jdb in 1.3, and run the same program (no changes), I see the first bit of
output in 130 seconds (2 min. 10 seconds). I see this as a major bug.
(Review ID: 117261)
======================================================================
tim.bell@Eng 2001-02-20
From: <xxx@zzz>
To: "'###@###.###'" <###@###.###>
Subject: Performance of JDI with JDK 1.3
Date: Tue, 20 Feb 2001 13:42:17 +0100
Hello,
I have a question concerning JDI performance:
I have attached a ZIP file which contains
a debugee and a JDI client.
The debugee is very short and uses the JAXP
XML parser. It parses an XML file multiple times
and traces the execution times on stderr.
The JDI client can be attached to the debugee
(dt_shmem or dt_socket).
My measured times are listed in result.txt.
In general it shows:
debugee time (in sec) time (in sec)
JDK1.2.2 JDK1.3
-------------------------------------------------------
no JDWP at all 11,2 3,1
JDWP switched on 53,7 15,9
(no JDI client
connected)
JDI client connected 54,1 702,1
(no events requested)
JDI client connected 217,4 1320 (approx.)
few method entry events
requested
From my point of view it seems the 1.3 VM performance is
decreased enormously if any JDI client is attached.
I have noticed this behaviour during normal debugging as well.
For debugging sessions I need to switch back to JDK 1.2.2 because
of the performance decrease...
We are using KAWA resp. JBuilder4 in our department.
- relates to
-
JDK-4506433 Tools vendors: debugging is too slow
-
- Closed
-
-
JDK-4675577 RFE: need more efficient logic for BREAKPOINT/SINGLE_STEP in one set
-
- Resolved
-