🚀 go-pugleaf

In article <BvnJ4A.pL.2@cs.cmu.edu> nash+@cs.cmu.edu (Richard Nash) writes:
> >In article <dillon.0oq2@overload.Berkeley.CA.US>  
dillon@overload.Berkeley.CA.US  
> >(Matthew Dillon) writes:
> >> 
> >>     The main problem with 3.0 is in paging from disk, a faster processor
> >>     will not help at all in that area.
> >> 
> >>                                       -Matt
> >> 
> >> --
> >> 
> >>     Matthew Dillon            dillon@Overload.Berkeley.CA.US
> >>     1005 Apollo Way           uunet.uu.net!overload!dillon
> >>     Incline Village, NV. 89451        ham: KC6LVW (no mail drop)
> >>     USA                       Sandel-Avery Engineering (702)831-8000
> >
> >If only it were really that simple!
>
> But it is true! Most of the perceived 'slowness' about next machines have to
> do with disk or network access. Even if the processor was infinitely fast,
> executing in zero time, the machine would look and feel almost as slow as
> it does today.
>
	This is one area in which NeXT is abhorantly behind the industry  
standards. Take a look at the sequential read rates from disk or over the  
network for the turbo machines, and you'll understand why NeXTs are slow. The  
turbo machines are actually slower that the 25MHz slabs (even with the ASIC  
chip)! I don't know what NeXT was thinking when they designed this, but I hope  
they fix it in the near future (i. e. on the RISC machines or possibly the  
40MHz version of the slab/cube). If you want to see the specifics check out  
Unix Review for Sept. 92, they review the turbo color and the disk/net times  
are amazing (and I don't mean pleasant...).
--
Robert Seymour				rseymour@reed.edu
Departments of Physics and Philosophy
Artificial Life Project			NeXTmail welcome
Reed Solar Energy Project (Sunrayce 1993)
Reed College Portland, OR