Re: [qvm86-devel] New changes

[Christian MICHON]

good for you :)

are you trying to "over-sell" gtk ? with "-serial stdio", I would expect
no need to use any graphic layer :)

yet something doesn't look right. The numbers don't look right...

If you're testing on a x86-32 machine, which I would assume is above
1Ghz, you should get indexes around 4 to 6 times higher (last column).

It looks as if:
- yes, qvm86 is enabled
- no, qvm86 is not accelerating anything

For info, on my windows 1.6GHz pentium M, I get 0.99 for
NUMERIC SORT without qvm86 acceleration, but I get 5.86
with qvm86 acceleration.

What is the arch of your 2.4.26 linux machine ?

Christian

On 6/23/05, Jim C. Brown <address@hidden> wrote:
> On Wed, Jun 22, 2005 at 11:10:52PM +0200, Christian MICHON wrote:
> > actually, there's another floppy for qemu ready made for testing :)
> > ftp://ftp.win4lin.com/pub/testing/nbench-qemu/nbench.flp
> >
> > when you switch to the serial console (CTRL-ALT-3), you can see
> > the benchmark results. With qvm86 enabled, it confirms nbench
> > crashes after the 1st sorting test.
> >
> > Christian
> >
> 
> I got your image. And test with qvm86 under a linux 2.4.26 kernel.
> 
> Mine doesn't crash until the powerdown.
> 
> $ qemu/qemu/i386-softmmu/qemu -m 32 -boot a -fda nbench.flp -use-gtk -serial 
> stdio
> QEMU acceleration layer enabled
> 
> (<unknown>:31870): Gdk-CRITICAL **: gdk_draw_image: assertion 
> `GDK_IS_DRAWABLE (drawable)' failed
> PCI: PIIX3: Enabling Passive Release on 00:01.0
> 
> BYTEmark* Native Mode Benchmark ver. 2 (10/95)
> Index-split by Andrew D. Balsa (11/97)
> Linux/Unix* port by Uwe F. Mayer (12/96,11/97)
> Power-off on exit code by Win4Lin, Inc. (4/05)
> 
> TEST                : Iterations/sec.  : Old Index   : New Index
>                     :                  : Pentium 90* : AMD K6/233*
> --------------------:------------------:-------------:------------
> NUMERIC SORT        :          126.48  :       3.24  :       1.07
> STRING SORT         :          6.1659  :       2.76  :       0.43
> BITFIELD            :      3.3823e+07  :       5.80  :       1.21
> FP EMULATION        :          13.554  :       6.50  :       1.50
> FOURIER             :          1908.4  :       2.17  :       1.22
> ASSIGNMENT          :          4.2006  :      15.98  :       4.15
> IDEA                :          404.79  :       6.19  :       1.84
> HUFFMAN             :          99.352  :       2.76  :       0.88
> NEURAL NET          :         0.83494  :       1.34  :       0.56
> LU DECOMPOSITION    :          40.639  :       2.11  :       1.52
> ==========================ORIGINAL BYTEMARK RESULTS==========================
> INTEGER INDEX       : 5.118
> FLOATING-POINT INDEX: 1.830
> Baseline (MSDOS*)   : Pentium* 90, 256 KB L2-cache, Watcom* compiler 10.0
> ==============================LINUX DATA BELOW===============================
> CPU                 :
> L2 Cache            :
> OS                  :
> C compiler          : gcc version 3.3.5 (Debian 1:3.3.5-8)
> libc                : ld-2.3.2.so
> MEMORY INDEX        : 1.289
> INTEGER INDEX       : 1.268
> FLOATING-POINT INDEX: 1.015
> Baseline (LINUX)    : AMD K6/233*, 512 KB L2-cache, gcc 2.7.2.3, libc-5.4.38
> * Trademarks are property of their respective holder.
> Power down.
> Segmentation fault
> 
> --
> Infinite complexity begets infinite beauty.
> Infinite precision begets infinite perfection.
> 


-- 
Christian