Skip to contentSkip to conversation

Combined low latency & performance patches for 2.4.18

DISCUSSION Con Kolivas 3 participants 7 messages Apr 29, 2002 – May 14, 2002

lore
Share & tools
lore
#0 lore raw

I've combined the following patches against 2.4.18:

Scheduler 0(1)
Low Latency
Preemptible
Compressed cache
new IDE subsystem

These are based on fairly recent patches, but not all are the latest.

I've noticed palpable improvements (the feel of using the machine) enabling all of these except for the compressed cache but have no data to support my feelings. The machine I tried them on was up for 3 weeks with heavy loads and proved to be quite stable. I've posted a combined patch at:

http://kernel.kolivas.net

Feel free to test it out and tell me what you think.
Thanks very much to those who put all the effort into each one of these patches.

Con Kolivas

P.S. Please CC me as not subscribed to LKML.

#1 lore raw
Quoted 4 lines · you

After applying it, if you ever do a make mrproper oldconfig dep bzImage it fails to compile sched.c as follows

If I don't do an mrproper it compiles ok. Haven't tested yet.

Hmm
I used a make mrproper && make clean followed by manual configuration without any problems but thanks for your input. I'm not claiming to be a patch or kernel guru. Just offering what worked for me.

Quoted 1 line

It's the O1 sched patch. Not your fault....

Thanks thats kinda reassuring :)

Con.

#2 lore raw
In reply to Con Kolivas
Quoted 12 lines · Con Kolivas
Quoted 4 lines · you

After applying it, if you ever do a make mrproper oldconfig dep bzImage it fails to compile sched.c as follows

If I don't do an mrproper it compiles ok. Haven't tested yet.

Hmm
I used a make mrproper && make clean followed by manual configuration without any problems but thanks for your input. I'm not claiming to be a patch or kernel guru. Just offering what worked for me.

Quoted 6 lines
Quoted 1 line

It's the O1 sched patch. Not your fault....

Thanks thats kinda reassuring :)

Con.

I take that back. It's the low-latency patch. You must have said no to it in your .config.
I do not think all these patches play well together with the O(1)

Signature · 2 lines
Mark Hounschell
[email protected]
#3 lore raw
In reply to Con Kolivas

Mark

Quoted 14 lines · you
Quoted 13 lines · Con Kolivas
Quoted 4 lines · you

After applying it, if you ever do a make mrproper oldconfig dep bzImage it fails to compile sched.c as follows

If I don't do an mrproper it compiles ok. Haven't tested yet.

Hmm
I used a make mrproper && make clean followed by manual configuration without any problems but thanks for your input. I'm not claiming to be a patch or kernel guru. Just offering what worked for me.

Quoted 1 line

It's the O1 sched patch. Not your fault....

Quoted 7 lines
Quoted 3 lines

Thanks thats kinda reassuring :)

Con.

I take that back. It's the low-latency patch. You must have said no to it in your .config.
I do not think all these patches play well together with the O(1)

Here is the relevant section from my .config:

# Processor type and features
#
CONFIG_LOLAT=y
# CONFIG_LOLAT_SYSCTL is not set

CONFIG_PREEMPT=y

As you can see, the low latency is enabled. I didn't see any point in enabling sysctl so I havent tried that. Perhaps that is it? The other thing is, I don't believe it works well with SMP which I am not using.

Con.

#4 lore raw
In reply to Con Kolivas

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Mark (and lkml readers)

On further testing from the feedback I've gotten it seems that this patch works fine for gcc 2.95.3 and gcc 3.0.3+

There is definitely a problem with trying to compile it with gcc2.96

Cheers,
Con Kolivas
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE84RRfF6dfvkL3i1gRAmy0AJ91EYI+juz0W92oQzZsRLkgg1ggTwCfT+EY G/aXt+YD49+3mh2UIDIOlyo=
=bb9b
-----END PGP SIGNATURE-----

#5 lore raw
In reply to Con Kolivas
Quoted 4 lines · Con Kolivas

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Quoted 10 lines

Mark (and lkml readers)

On further testing from the feedback I've gotten it seems that this patch works fine for gcc 2.95.3 and gcc 3.0.3+

There is definitely a problem with trying to compile it with gcc2.96

Quoted 3 lines

As you can see, the low latency is enabled. I didn't see any point in enabling sysctl so I havent tried that. Perhaps that is it? The other thing is, I don't believe it works well with SMP which I am not using.

I'm using 2.95.3 but also SMP. Maybe thats it..

Regards
Mark

#6 lore raw
In reply to Con Kolivas

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Mark

Quoted 1 line · you

I'm using 2.95.3 but also SMP. Maybe thats it..

Bingo. Cant remember which one it was but one of these didnt support SMP.

Con.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE84RxtF6dfvkL3i1gRAnYJAJwKTI6M2fUOZyrzxyKv6r9yl1WjPQCcDzwn kArjn0U/ghc1qXMorgaKji8=
=J2Gy
-----END PGP SIGNATURE-----