linux-kernel - Re: x86_64 and AMD with C1E

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <b70f2f960710010740x56683c38ma2c16f57a128cccf@mail.gmail.com>
Date:	Mon, 1 Oct 2007 18:40:01 +0400
From:	"Mikhail Kshevetskiy" <mikhail.kshevetskiy@...il.com>
To:	"Andi Kleen" <andi@...stfloor.org>
Cc:	linux-kernel@...r.kernel.org,
	"Thomas Gleixner" <tglx@...utronix.de>
Subject: Re: x86_64 and AMD with C1E

2007/10/1, Andi Kleen <andi@...stfloor.org>:
> > No, it boot and work normally. The only thing i bother, is the additional
>
> Ok you want an additional feature; not a bug fixed.
>
> Please make that always clear.
>
> > 260 timer interrupts per seconds.
> > Here is short result:
> >
> > c1e enabled:
> >   -- power consumption about 23 watts
> >   -- there is only C1 power state enabled
> >   -- there are about 260 timer interrupts per seconds
>
> If you want less timer interrupts in a standard kernel
> a simple way is to compile with CONFIG_HZ=100

i use NOHZ enabled kernel for test. I just noticed that i have the
same timer interrupts as if NOHZ was not selected (e.g. 2.6.22/x86_64)

> > tested with x86_64(2.6.22, 2.6.23-rc8, 2.6.23-rc8-hrt1), i386(2.6.21, 2.6.22,
> > 2.6.23-rc5-hrt1)
>
> There is normally a threshold above which you don't save significantly
> more power by doing less timer interrupts. You can test this
> by doing the CONFIG_HZ=100 above.
>
> Also as a warning: most simple methods of measuring power (like
> watching battery consumption) are fairly imprecise and can
> have large systematic errors.
>
> Normally reducing the timer interrupts significantly only helps
> on longer latency sleep modi; which is C2+ where the latency
> of going in/out of the sleep is reaching towards half of
> the timer period. But that is not the case on the AMD C1e
> implementations as far as I know -- they're fairly fast.

AMD say: C1E is a Stop Grant state like C3 supported by dual core
mobile processors.
The difference between C1E and C3 is that transition into C1E is not
initiated by the
operating system. The C1E state is initiated by an I/O transaction to
the chipset when
both cores transition into the C1 state.

> > c1e disabled:
> >   -- power consumption about 27 watts
>
> Don't do this then.
>
> > I want to reduce the power consumption of my notebook. I see the 2
> > possibility:
> >   -- remove 260 additional timer interrupts (c1e enabled case )
>
> It's unclear it will really help all that much. Most likely
> you'll get more bang for the buck by disabling other devices.
> e.g. reducing the backlight or unloading USB/firewire/sound
> can do wonders.
>
> >   -- force enable C1, C2 and C3 states  (c1e disabled case)
>
> If the BIOS doesn't announce it the platform doesn't implement it.

I think, it should support C1 and C2 at least.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/