1) disable CONFIG_MODVERSIONS, after receiving the following at compilation:
--8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<--
../versions.mk:24: warning: overriding commands for target `/usr/src/linux/include/linux/modversions.h'
/usr/src/linux/.hdepend:324: warning: ignoring old commands for target `/usr/src/linux/include/linux/modversions.h'
ksyms.c:160: warning: missing braces around initializer for `symbol_table[1]'
ksyms.c:161: warning: initialization from incompatible pointer type
ksyms.c:167: initializer element for `symbol_table.name' is not constant
ksyms.c:173: initializer element for `symbol_table[3].addr' is not constant
ksyms.c:191: warning: initialization from incompatible pointer type
< many similar lines available on request >
ksyms.c:539: initializer element for `symbol_table.name' is not constant
make[1]: *** [ksyms.o] Error 1
make: *** [linuxsubdirs] Error 2
--8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<--
This happens when unmacroing (?) macro X, which depends on the above
variable. After disabling the variable, ksyms compiled OK.
2) disable the smbfs stuff: it complains about some structure element
not existing, or stuff like that. Since last night I was tooo tired to
try & understand what was going on, I just disabled smbfs. Linus says
that it requires some special attention, anyway.
So, after one release with file write/copy going fast again, I am back
to worse results. I go on not receiving any echo on this. I have
reason to believe that my machine did not undergo changes of any
particular size to justify these differences, so I post results as
they are. If anybody tells me that there is ABSOLUTELY NO REASON in
the kernel patches for those results to change that way, I might try
and give a longer look at my setup...
You can make your own comparisons at:
http://neon.ingenia.com/cgi-bin/bm
--8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<--
From:
* Linux pimpinel.fluido.org 1.3.37 #59 Sun Oct 29 10:32:57 MET 1995 i486
* Sun Oct 29 10:54:35 MET 1995
To:
* Linux pimpinel.fluido.org 1.3.38 #67 Wed Nov 8 21:54:33 MET 1995 i486
* Wed Nov 8 22:16:48 MET 1995
*******
Results
*******
Process Creation Test || 353.1 -> 366.2 +3.71%
System Call Overhead Test || 29616.0 -> 30524.3 +3.07%
Execl Throughput Test || 65.6 -> 67.1 +2.29%
Shell scripts (1 concurrent) || 92.6 -> 93.0 +0.43%
C Compiler Test || 47.0 -> 47.1 +0.21%
Shell scripts (2 concurrent) || 49.0 -> 49.1 +0.20%
Arithmetic Test (type = short) || 7503.4 -> 7506.1 +0.04%
Arithmetic Test (type = float) || 5065.8 -> 5067.4 +0.03%
Arithmetic Test (type = arithoh) || 127145.3 -> 127180.7 +0.03%
Arithmetic Test (type = register) || 8369.5 -> 8371.7 +0.03%
Arithmetic Test (type = int) || 8368.9 -> 8371.0 +0.03%
Arithmetic Test (type = double) || 5066.4 -> 5067.4 +0.02%
Arithmetic Test (type = long) || 8368.8 -> 8370.2 +0.02%
Shell scripts (4 concurrent) || 25.0 -> 25.0 +0.00%
Shell scripts (8 concurrent) || 12.0 -> 12.0 +0.00%
Recursion Test--Tower of Hanoi || 727.1 -> 726.5 -0.08%
Pipe-based Context Switching Test || 8810.1 -> 8775.5 -0.39%
Dhrystone 2 using register variables || 49710.4 -> 49398.4 -0.63%
Dc: sqrt(2) to 99 decimal places || 11668.1 -> 11372.7 -2.53%
File Read (30 seconds) || 4453.0 -> 4316.0 -3.08%
File Read (10 seconds) || 4455.0 -> 4287.0 -3.77%
Pipe Throughput Test || 18012.3 -> 17039.2 -5.40%
Dhrystone 2 without register variables || 49939.0 -> 46699.5 -6.49%
File Copy (10 seconds) || 2675.0 -> 986.0 -63.14%
File Copy (30 seconds) || 2613.0 -> 955.0 -63.45%
File Write (10 seconds) || 7380.0 -> 1400.0 -81.03%
File Write (30 seconds) || 8560.0 -> 1200.0 -85.98%
--8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<--
From:
* Linux pimpinel.fluido.org 1.2.10 #59 Tue Jun 13 09:46:17 MET DST 1995 i486
* Tue Jun 27 10:18:13 MET DST 1995
To:
* Linux pimpinel.fluido.org 1.3.38 #67 Wed Nov 8 21:54:33 MET 1995 i486
* Wed Nov 8 22:16:48 MET 1995
*******
Results
*******
Process Creation Test || 119.9 -> 366.2 +205.42%
Pipe-based Context Switching Test || 3446.9 -> 8775.5 +154.59%
Dc: sqrt(2) to 99 decimal places || 5906.2 -> 11372.7 +92.56%
Execl Throughput Test || 51.6 -> 67.1 +30.04%
Shell scripts (8 concurrent) || 10.0 -> 12.0 +20.00%
Shell scripts (2 concurrent) || 41.0 -> 49.1 +19.76%
Shell scripts (4 concurrent) || 21.0 -> 25.0 +19.05%
Shell scripts (1 concurrent) || 78.3 -> 93.0 +18.77%
C Compiler Test || 42.2 -> 47.1 +11.61%
System Call Overhead Test || 29238.2 -> 30524.3 +4.40%
Recursion Test--Tower of Hanoi || 709.2 -> 726.5 +2.44%
Dhrystone 2 without register variables || 46284.2 -> 46699.5 +0.90%
Arithmetic Test (type = float) || 5055.6 -> 5067.4 +0.23%
Arithmetic Test (type = register) || 8353.2 -> 8371.7 +0.22%
Arithmetic Test (type = int) || 8353.4 -> 8371.0 +0.21%
Arithmetic Test (type = arithoh) || 126914.4 -> 127180.7 +0.21%
Arithmetic Test (type = double) || 5056.8 -> 5067.4 +0.21%
Arithmetic Test (type = long) || 8353.5 -> 8370.2 +0.20%
Arithmetic Test (type = short) || 7496.0 -> 7506.1 +0.13%
Dhrystone 2 using register variables || 49802.4 -> 49398.4 -0.81%
Pipe Throughput Test || 18512.0 -> 17039.2 -7.96%
File Write (10 seconds) || 8100.0 -> 1400.0 -82.72%
File Write (30 seconds) || 8911.0 -> 1200.0 -86.53%
--8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<--
Enjoy!
Carlo
-- * Se la Strada e la sua Virtu' non fossero state messe da parte, * K * Carlo E. Prelz - fluido@marktest.pt che bisogno ci sarebbe * di parlare tanto di amore e di rettitudine? (Chuang-Tzu)