Gentoo Archives: gentoo-user

From: Grant Edwards <grant.b.edwards@×××××.com>
To: gentoo-user@l.g.o
Subject: [gentoo-user] Re: CPU speed scaling quirk (Intel; Dell i660)
Date: Thu, 05 Mar 2020 19:40:36
Message-Id: r3rkj6$140q$1@ciao.gmane.io
In Reply to: Re: [gentoo-user] CPU speed scaling quirk (Intel; Dell i660) by madscientistatlarge
1 On 2020-03-05, madscientistatlarge <madscientistatlarge@××××××××××.com> wrote:
2
3 > To reduce problems with emitted Radio Frequency Interference, most
4 > processors now use a clock that varies in speed over time. This
5 > doesn't really reduce the emitted energy, but because it is always
6 > changing frequency interference with other devices tends to be
7 > intermittent, and Ideally unnoticeable. Also the oscillators used
8 > in computers are not the most precise, they don't need to be and
9 > precision cost. The bios may let you toggle this deliberate
10 > frequency variation and off, which I suppose could be critical in
11 > some real-time cases, or a varying clock may, in some cases cause
12 > objectionable interference where as the fixed clock, may not, YMMV.
13
14 A clock that varies like that is usually referred to has a
15 "spread-spectrum" clock. If properly implimented it has no measurable
16 effect on software execution (even for real-time cases) because the
17 variation is done so that the average frequency is "constant" and the
18 deviation from that average sums to 0 for any significant period of
19 time (anything over a few hundred microseconds).
20
21 The variation of the average over temperature and supply voltage is
22 usually far more significant.
23
24 --
25 Grant Edwards grant.b.edwards Yow! PIZZA!!
26 at
27 gmail.com