Re: a stupid question on speed on 68VZ328

Philippe Ney <[email protected]>
Newsgroups gmane.comp.video.picogui.devel
Message-ID <[email protected]>
> hi Micah, 
> hi Philippe
> 
> 
> > I haven't personally run PicoGUI on anything slower than an Agenda VR3
> > in a long time. You could run gprof to find the bottleneck, but it's
> > almost certainly the video driver. What video driver are you using,
> > and at what color depth? pgserver also has a benchmarking mode for
> > quantitatively measuring video speed.
> > 
>   we use the depth is 1bpp, and the benchmarking is flowing,
>  
> load video driver successfully, name is vz328
> vidw, vidh, vidd, vidf = 160, 160, 1, 0
> 
> Running video driver benchmarks...
> 
> Time             Name                           LGOP            Size
> -----------------------------------------------------------------------
>        10000 ns  nop                            none            128x128
>        30000 ns  color pgtohwr                  none            128x128
>        20000 ns  color hwrtopg                  none            128x128
>        20000 ns  update                         none            128x128
>        50000 ns  pixel                          none            128x128
>        50000 ns  getpixel                       none            128x128
>       160000 ns  slab                           none            128x128
>       520000 ns  bar                            none            128x128
>      1190000 ns  line                           none            128x128
>     19000000 ns  rect                           none            128x128
>   1273720000 ns  gradient                       none            128x128
>     16330000 ns  blit                           none            128x128
>     16240000 ns  scrollblit up                  none            128x128
>     27950000 ns  scrollblit down                none            128x128
>     16380000 ns  scrollblit left                none            128x128
>    182920000 ns  scrollblit right               none            128x128
>     43210000 ns  multiblit (1x1 tile)           none            128x128
>     67860000 ns  multiblit (1x32 tile)          none            128x128
>     30390000 ns  multiblit (32x1 tile)          none            128x128
>     18410000 ns  multiblit (32x32 tile)         none            128x128
>   1440330000 ns  rotateblit (0 degrees)         none            128x128
>   1440320000 ns  rotateblit (90 degrees)        none            128x128
>   1440330000 ns  rotateblit (180 degrees)       none            128x128
>   1440320000 ns  rotateblit (270 degrees)       none            128x128
>     19780000 ns  ellipse                        none            128x128
>     22330000 ns  fellipse                       none            128x128
>     27540000 ns  polygon (star)                 none            128x128
> 
> > To have an idea about what will be the behavior of the programs on our
> > platform, we used the XCopilot emulator on pc. Is this also what you 
> spoke
> > about when you said "... when run in PC, it's ..."?
>    when we debug on X86 platform, we use SDL as video driver on XWindow
> 
> > Unfortunately, I don't have one that works here... but if my memories
> > are right, it wasn't possible to have such informations with the 2.0.x
> > kernel.
> > (or maybe it was something else that wasn't accessible...)
> > Anyway, if I could get some info, I'll tell you.
>    
>    Thanks. we had used 2.0.38.kernel when we use M68EZ328 , we also can 
> find in /proc/cpuinfo file.and the speed about 1.4BogoMips  
>                                                sanit 

Hi sanit,

I got some benchmarks from our last working VZ board...
mmh.. seems the CPU is clocked at 17MHz... I don't know if this show the
real clocking of it or if it's an error...
The pgserver is a bit old and then the benchmarks aren't the same as yours
I even don't remember their signification...
Maybe Micah could help about this?


-philippe

--------------8<-------------------------
# uname -a
uClinux (none) 2.0.38.6 #18, 2002-05-31 21:53:51 CEST chipslice

# cat /proc/cpuinfo
CPU:            MC68VZ328
MMU:            none
FPU:            none
Clocking:       17.1MHz
BogoMips:       2.14
Calibration:    1072000 loops

# cat /proc/meminfo
        total:    used:    free:  shared: buffers:  cached:
Mem:   7950336   647168  7303168        0   352256     4096
Swap:        0        0        0
MemTotal:      7764 kB
MemFree:       7132 kB
MemShared:        0 kB
Buffers:        344 kB
Cached:           4 kB
SwapTotal:        0 kB
SwapFree:         0 kB

# pgserver --benchmark=1
Please wait... Each test will take 4 seconds to complete.
Results are measured in frames per second. Complete includes
screen hardware updates, raw is just the software. Raw should
equal complete if the driver does not double-buffer

 Test  | Complete / Raw
------------------------
     1 |     1.50 /     2.50
     2 |     0.14 /     0.12
     3 |     4.50 /     5.00
     4 |    20.50 /    22.00
     5 |    10.50 /    10.50
     6 |     1.00 /     1.00
Alpha channel test requires PNG loader
[...]
Alpha channel test requires PNG loader
     7 |    20.00 / Alpha channel test requires PNG loader
Alpha channel test requires PNG loader
[...]
Alpha channel test requires PNG loader
   21.00
     8 |     0.02 /     0.02

Done.
kfree of non-kmalloced memory: 0007c018, next= 00000000, order=0
#


-------------------------------------------------------
This SF.NET email is sponsored by:
SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See!
http://www.vasoftware.com
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.