🚀 go-pugleaf

In article <exuptr.349.714338170@exu.ericsson.se> exuptr@exu.ericsson.se (Patrick Taylor) writes:
}In article <1992Aug20.131829.5606@tc.cornell.edu> homer@theory.TC.Cornell.EDU (Homer Smith) writes:
}>     I wrote a short program in assembler to write to every pixel on
}>     On my 386/33 with Trident 8900 chip the program takes 1:10
}>minutes.
}>     On my 486/33 with identical video board the program takes 11:46
}>minutes.
}
}First off, both times are too slow.
}
}>     Could someone please give me a hint why the 486 is 11 times
}>slower than the 386. 
}> 
}>     Using another program that did a simple INC AX,1 repeatedly,
}>the 486 was 2.8 times faster than the 386.
}
}How are you drawing pixels?  What video mode? If you are calling an INT 10h 
}function, there's your bottleneck.  As for why it might be slower, check 
}these things:
}
}1>  Do you have VIDEO / BIOS shadow on either machine?
}2>  Do you have any non-standard bus timings on either machine?  (Extended
}    setup)
}3>  Compare the loaded drivers, etc... on both machines, especially packet 
}    drivers, keyboard intercepts, video intercepts, etc...

4> Is the board in a 16-bit slot in the 386 and in an 8-bit slot in the 486?

5> Is the memory cache on the 486 enabled?

6> Are you running a memory manager on the 486 but not on the 386?  Or
   different memory managers on the two machines?  When doing INT NN calls,
   the memory manager gets control, so the efficiency of the memory manager
   is very important (QEMM is much faster than Microsoft's EMM386, for
   example).
-- 
Internet: RALF+@CS.CMU.EDU   |The University would disclaim this if it knew...
FIDO: Ralf Brown 1:129/26.1  |
BIT: RALF%CS.CMU.EDU@CARNEGIE|"Success has a simple formula: do your best,
AT&Tnet: (412)268-3053 school| and people may like it."   -- Sam Ewing