On Mon, Apr 09, 2001 at 09:08:37PM +0200, Jaromír Dolecek wrote:
> [...]
> I've put the code to ftp at:
> 	ftp://ftp.netbsd.org/pub/NetBSD/misc/jdolecek/
> 

Hi,
I just tried newpipe-1.4.tgz on various machines. The result is impressive
in some case:

Vax (MicroVAX 3300, KA640 CPU):
text    data    bss     dec     hex     filename
813688  19256   158120  991064  f1f58   netbsd-newpipe
809764  19224   157852  986840  f0ed8   netbsd-oldpipe
old pipes:
cordouan#/home/bouyer>/usr/pkg/bin/lmbench/bw_pipe 
Pipe bandwidth: 0.85 MB/sec
new pipes:
cordouan:/home/bouyer>/usr/pkg/bin/lmbench/bw_pipe 
Pipe bandwidth: 8.85 MB/sec

next68k (68040@25Mhz):
text    data    bss     dec     hex     filename
960176  44964   147856  1152996 1197e4  netbsd-newpipe
955496  44932   147588  1148016 118470  netbsd-oldpipe
old pipes:
barbuda:/users/cao/bouyer>/usr/pkg/bin/lmbench/bw_pipe
Pipe bandwidth: 1.12 MB/sec
new pipes:
barbuda:/users/cao/bouyer>/usr/pkg/bin/lmbench/bw_pipe
Pipe bandwidth: 3.59 MB/sec

Yes, 10x faster for the vax !
"only" 3x faster for the m68k :)

However it fails on a macppc system (kernel compiled from same tree
as the next):
boot device: wd0
root on wd0a dumps on wd0b
root file system type: ffs
Tue Apr 24 11:27:32 GMT 2001
panic: ltsleep: ident == NULL
Stopped in pid 7 (sh) at        cpu_Debugger+0x10:      lwz r0, r1, 0x14,
db> tr
at panic+e8
at ltsleep+11c
at sys_pipe+6b8
at dofileread+d4
at sys_read+98
at trap+35c
at ddblow+e0
at end+14f8be8
at end+14f8448
at end+14f8180
at end+14f5d70
at end+14f5248
at end+14ff3b8
at end+14ff2ec
at end+14f33cc
at end+ffcf3238
db> 

Any idea why ?
Starting bw_pipe from single user fails in the same way.
Hum, both next and vax are diskless, where the macppc boots from IDE ...

--
Manuel Bouyer, LIP6, Universite Paris VI.           Manuel.Bouyer@lip6.fr
--