2.4.2-ac3 == 400% speed improvement with Java

From: Mark Swanson (swansma@yahoo.com)
Date: Sat Feb 24 2001 - 09:33:54 EST


any 2.4.1-ac5 kernel and above until 2.4.2-ac3 had a bug (Ingo guessed
it might be the yield() bug) that made Java (and some other things
*really* slow). This is fixed now. When I profiled slow kernels the
mtrr_file_add was taking up all of the time - even though I had
disabled mtrr in my kernel (it wasn't in /proc).

(Alan Cox: I'm not sure if the yield() bug was the problem, or if that
patch was in the TUX patch (based on -ac3) or the -ac3 patch.)

Here was the old test: (With 2.4.2-ac3 things are even faster than
2.2.14 at 0:00.44elapsed!!)

> time java
(done multiple times so everything is in cache)

On 2.2.14 (400MHz Celeron/192MB RAM) this takes about 0.66ms.
0.37 user 0.07 system 0:00.62elapsed 70% CPU (oavgtext+0avgdata
0inputs+0outputs (3957major+1707minor)pagefulats 0swaps

On linux 2.4.2-ac1-TUX (same machine)

0.33 user 2.17 system 0:03.49elapsed 71%CPU (ovgtext+0avgdata
0inputs+0outputs (3987major+1784minor)pagefaults 0swaps

Do You Yahoo!?
Get email at your own domain with Yahoo! Mail.
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

This archive was generated by hypermail 2b29 : Wed Feb 28 2001 - 21:00:07 EST