http://bugs.winehq.org/show_bug.cgi?id=23954
Summary: Threads not allocated to multiple processors. Product: Wine Version: 1.3.0 Platform: x86 URL: http://systester.sourceforge.net/ OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: -unknown AssignedTo: wine-bugs@winehq.org ReportedBy: kphillisjr@gmail.com
when running systester with multiple threads, the program does not properly run in parallel... Threads are created, and run, but they are not parallel.
Example Single-Thread run: (512k digits) systester-cli.exe -gausslg 512K -turns 1 -threads 1 Result: 19 512K N/A 0h 0m 5.318s
Example multi-threaded run: (512k digits) systester-cli.exe -gausslg 512K -turns 1 -threads 4 Result: 19 512K OK 0h 0m 15.864s Expected result: around 7s.
Example Single-Thread run: (1M digits) systester-cli.exe -gausslg 1M -turns 1 -threads 1 result: 20 1M N/A 0h 0m 13.438s
Example multi-threaded run: (1M digits, 4 threads) systester-cli.exe -gausslg 1M -turns 1 -threads 4 result: 20 1M OK 0h 0m 53.486s Expected Result: about 15s.
example single-thread run: (1M digits, 1 thread, benchmark mode) systester-cli.exe -gausslg 1M -turns 1 -threads 1 -bench result: 20 1M N/A 0h 0m 13.418s
Example multi-threaded run: (1M digits, 4 threads, benchmark mode) systester-cli.exe -gausslg 1M -turns 1 -threads 4 -bench result: 20 1M N/A 0h 0m 50.798s expected result: around 15s.
Test system: Athlon II x4 630 Ubuntu 10.04 (amd64)
http://bugs.winehq.org/show_bug.cgi?id=23954
Stephen E. bhs2007@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |bhs2007@gmail.com
--- Comment #1 from Stephen E. bhs2007@gmail.com 2010-08-10 15:54:23 --- Confirmed using wine-1.3.0 on Ubuntu 10.04 amd64 with Intel Core i5 650 (3.20GHz). Got similar times to those provided by the OP.
http://bugs.winehq.org/show_bug.cgi?id=23954
Austin English austinenglish@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |download, source
http://bugs.winehq.org/show_bug.cgi?id=23954
--- Comment #2 from Ken Phillis Jr kphillisjr@gmail.com 2010-08-18 14:28:39 --- Tested against wine-1.3.0-396-g8333559 and the problem still exists. I'm thinking the system used to make sure all the threads sync up and start is possibly not working.
http://bugs.winehq.org/show_bug.cgi?id=23954
--- Comment #3 from butraxz@gmail.com 2013-06-09 11:39:42 CDT --- This has not been updated for over 900 days.
Is this still an issue in 1.6-rc1 or higher or is this abandoned ?
http://bugs.winehq.org/show_bug.cgi?id=23954
--- Comment #4 from Ken Phillis Jr kphillisjr@gmail.com 2013-06-09 12:24:27 CDT --- (In reply to comment #3)
This has not been updated for over 900 days.
Is this still an issue in 1.6-rc1 or higher or is this abandoned ?
There is still issues with recent versions of the plugin where the threaded run clearly takes longer than the single thread copy. However, overall the runtime has actually improved using wine v1.5.29. I'll give version 1.6-rc1 a try later.
Results: ================================ systester version 1.5.1( win32 edition) Command Line: wine systester-cli.exe -gausslg 512K -turns 1 -threads 1 Result: 19 512K N/A 0h 0m 5.850s
Command Line: wine systester-cli.exe -gausslg 512K -turns 1 -threads 4 Result: 19 512K OK 0h 0m 8.741s
Command Line: wine systester-cli.exe -gausslg 1M -turns 1 -threads 1 Result: 20 1M N/A 0h 0m 13.225s
Command Line: wine systester-cli.exe -gausslg 1M -turns 1 -threads 4 Result: 20 1M OK 0h 0m 20.330s
Command Line: wine systester-cli.exe -gausslg 1M -turns 1 -threads 1 -bench Result: 20 1M N/A 0h 0m 15.673s
Command Line: wine systester-cli.exe -gausslg 1M -turns 1 -threads 4 -bench Result: 20 1M N/A 0h 0m 20.278s
http://bugs.winehq.org/show_bug.cgi?id=23954
--- Comment #5 from Ken Phillis Jr kphillisjr@gmail.com 2013-06-09 12:25:36 CDT --- Created attachment 44717 --> http://bugs.winehq.org/attachment.cgi?id=44717 Results of running systester on a windows 7 pc. ( 64-bit )
http://bugs.winehq.org/show_bug.cgi?id=23954
--- Comment #6 from Ken Phillis Jr kphillisjr@gmail.com 2013-06-09 12:26:25 CDT --- Created attachment 44718 --> http://bugs.winehq.org/attachment.cgi?id=44718 Results from Wine v1.5.29 ( on linux mint 13 )
http://bugs.winehq.org/show_bug.cgi?id=23954
--- Comment #7 from Ken Phillis Jr kphillisjr@gmail.com 2013-06-09 23:16:24 CDT --- Created attachment 44726 --> http://bugs.winehq.org/attachment.cgi?id=44726 Linux results from the athlon II 630 pc.
Results
http://bugs.winehq.org/show_bug.cgi?id=23954
--- Comment #8 from Ken Phillis Jr kphillisjr@gmail.com 2013-06-09 23:29:47 CDT --- Created attachment 44727 --> http://bugs.winehq.org/attachment.cgi?id=44727 Windows 7 results for the a6-3400m
Based on some further checking, I think this bug can be marked as resolved. The threading behavior is now a lot better, and I am able to get results nearly identical to windows.
NOTE: The results for single threaded runs on windows at least for AMD chips should match the multi-threaded passes. The only time this does not hold true is when the "Power Options" are not set to "High Performance"
https://bugs.winehq.org/show_bug.cgi?id=23954
super_man@post.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |00cpxxx@gmail.com, | |super_man@post.com
https://bugs.winehq.org/show_bug.cgi?id=23954
Bruno Jesus 00cpxxx@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |RESOLVED Resolution|--- |FIXED
--- Comment #9 from Bruno Jesus 00cpxxx@gmail.com --- Resolving fixed as per previous comment.
https://bugs.winehq.org/show_bug.cgi?id=23954
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #10 from Alexandre Julliard julliard@winehq.org --- Closing bugs fixed in 1.9.12.
-
wine-bugs＠winehq.org