Viktor Szakáts wrote:
Test #1: Comparing -gc2 vs. -gc3 for speedtst in ST and MT modes.
Test #2: Comparing -ko vs. non -ko modes in -gc2 vs. -gc3 speedtsts in
ST mode.
Command lines run from Far Manager command prompt, in
C:\harbour-msvc-gc3\bin directory:
> test#1-hbbench.bat > test#1-results.txt 2>&1
> test#2-hbbench.bat > test#2-results.txt 2>&1
Link:
http://www.syenar.hu/harbour/winbench2.zip
Please challenge the measuring methodology (and I'll try
to further enhance it), also try to run it on your own system
(even with other Harbour builds / compilers), and post the
results here. In the end we can even include such batch
files under /tests directory.
Hi,
it seems, the only significant execution speed difference in gc3 mode is
in T030 test. It's related to macro compiler. It is natural, that macro
compiler speed is decreased, if -ko makes additional optimisation in
macro compiler is done (but this is a subject to be checked!!!, I guess
.prg compilation switch should not change macro compiler behaviour).
Though, it is a little strange execution test T031 is not affected by
-ko. Macro-compiled expressions are very similar.
N_LOOPS: 1000000 df ko
[ T000: empty loop overhead ]............................0.03 .0.03
============================================================= =====
[ T001: x := L_C ].......................................0.05 .0.05
...
[ T030: x := &( "f1(" + str(i) + ")" ) ].................3.78 .6.14
[ T031: bc := &( "{|x|f1(x)}" ), eval( bc, i ) ].........4.25 .4.31
...
============================================================= =====
[ total application time: ].............................29.86 32.13
[ total real time: ]....................................29.92 32.19
Regards,
Mindaugas
_______________________________________________
Harbour mailing list
[email protected]
http://lists.harbour-project.org/mailman/listinfo/harbour
