From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
To: David Laight <David.Laight@ACULAB.COM>
Cc: Xie Shaohui-B21989 <B21989@freescale.com>,
Liu Qiang-B32616 <B32616@freescale.com>,
Zang Roy-R61911 <r61911@freescale.com>,
Timur Tabi <timur@tabi.org>,
"tiejun.chen" <tiejun.chen@windriver.com>,
Fleming Andy-AFLEMING <afleming@freescale.com>,
Bhushan Bharat-R65777 <R65777@freescale.com>,
linuxppc-dev@lists.ozlabs.org
Subject: Re: fsqrt
Date: Fri, 07 Jun 2013 22:14:05 +1000 [thread overview]
Message-ID: <1370607245.3766.386.camel@pasglop> (raw)
In-Reply-To: <AE90C24D6B3A694183C094C60CF0A2F6026B7279@saturn3.aculab.com>
On Fri, 2013-06-07 at 11:48 +0100, David Laight wrote:
> > For those interested, this is the Quake3 sqrt from Carmack ...
> there's
> > plenty of literature about it one or two google clicks away :-)
>
> I guess that is a rough enough approximation for graphics.
>
> However it will be miscompiled unless i and y are put in a union.
It won't in the kernel disables strict aliasing :-)
Anyway, that was just a hack and plenty enough to get anaconda going,
the bloody thing only uses fsqrt because it's python crappola does
something like exp(1.0) / sqrt(2.0) as part of its random number stuff.
Honestly I could have made it just return 1.0 and it would probably have
worked :-)
However, my point remains, it would be very much worthwhile for the
kernel to have some reasonable emulation of those missing instructions
(afaik only a handful) like we have for isel, popcnt* etc... especially
since distros seem to be keen on enabling the use of them in their
toolchain.
I don't personally have the bandwidth to do a clean implementation (that
handles FP exceptions, NaNs, FPSCR, etc...) but I believe it would be
valuable if somebody else did (hint hint hint :-) since without this,
Fedora ppc64 is basically going to be a non-started on those chips.
BTW. Did you guys (ie. FSL) finally add fsqrt to e6500 or it's still
out ?
Cheers,
Ben.
next prev parent reply other threads:[~2013-06-07 12:14 UTC|newest]
Thread overview: 78+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-05-16 4:47 SATA FSL and upstreaming Benjamin Herrenschmidt
2013-05-16 5:45 ` Benjamin Herrenschmidt
2013-05-16 5:55 ` tiejun.chen
2013-05-16 6:06 ` Benjamin Herrenschmidt
2013-05-16 5:59 ` Zang Roy-R61911
2013-05-16 6:01 ` Bhushan Bharat-R65777
2013-05-16 6:05 ` Zang Roy-R61911
2013-05-16 6:09 ` Benjamin Herrenschmidt
2013-05-16 6:17 ` tiejun.chen
2013-05-16 6:20 ` Zang Roy-R61911
2013-05-16 6:25 ` tiejun.chen
2013-05-16 7:20 ` Zang Roy-R61911
2013-05-16 6:26 ` Benjamin Herrenschmidt
2013-05-16 6:21 ` Benjamin Herrenschmidt
2013-05-16 6:35 ` tiejun.chen
2013-05-16 6:37 ` Zang Roy-R61911
2013-05-16 6:40 ` Benjamin Herrenschmidt
2013-05-16 6:43 ` tiejun.chen
2013-05-16 6:48 ` Bhushan Bharat-R65777
2013-05-16 6:49 ` Zang Roy-R61911
2013-05-16 6:53 ` Benjamin Herrenschmidt
2013-05-16 6:56 ` tiejun.chen
2013-05-16 7:01 ` Zang Roy-R61911
2013-05-16 7:05 ` Benjamin Herrenschmidt
2013-05-16 7:13 ` Bhushan Bharat-R65777
2013-05-16 7:26 ` Benjamin Herrenschmidt
2013-05-16 7:20 ` Xie Shaohui-B21989
2013-05-16 7:25 ` Bhushan Bharat-R65777
2013-05-16 6:59 ` Benjamin Herrenschmidt
2013-05-16 7:17 ` Zang Roy-R61911
2013-05-16 6:52 ` Benjamin Herrenschmidt
2013-05-16 14:56 ` Timur Tabi
2013-06-07 3:52 ` Benjamin Herrenschmidt
2013-06-07 4:39 ` Benjamin Herrenschmidt
2013-06-07 4:45 ` Zang Roy-R61911
2013-06-07 4:47 ` Benjamin Herrenschmidt
2013-06-07 4:50 ` Zang Roy-R61911
2013-06-07 7:41 ` fsqrt Benjamin Herrenschmidt
2013-06-07 7:45 ` fsqrt Zang Roy-R61911
2013-06-07 8:53 ` fsqrt Benjamin Herrenschmidt
2013-06-07 8:59 ` fsqrt Benjamin Herrenschmidt
2013-06-07 10:48 ` fsqrt David Laight
2013-06-07 12:14 ` Benjamin Herrenschmidt [this message]
2013-06-07 19:19 ` fsqrt Kumar Gala
2013-06-07 23:23 ` fsqrt Benjamin Herrenschmidt
2013-06-07 23:25 ` fsqrt Benjamin Herrenschmidt
2013-06-07 23:30 ` fsqrt Benjamin Herrenschmidt
2013-06-08 0:20 ` fsqrt Dan Malek
2013-06-08 0:34 ` fsqrt Benjamin Herrenschmidt
2013-06-08 1:13 ` fsqrt Dan Malek
2013-06-08 4:31 ` fsqrt Benjamin Herrenschmidt
2013-06-09 6:32 ` fsqrt Benjamin Herrenschmidt
2013-06-07 7:46 ` fsqrt tiejun.chen
2013-06-07 8:53 ` fsqrt Benjamin Herrenschmidt
2013-06-07 9:02 ` fsqrt tiejun.chen
2013-06-07 12:07 ` fsqrt Benjamin Herrenschmidt
2013-06-07 7:05 ` FSL 64-bit DMA window question Benjamin Herrenschmidt
2013-06-07 7:58 ` Zang Roy-R61911
2013-06-07 8:55 ` Benjamin Herrenschmidt
2013-06-07 9:44 ` Zang Roy-R61911
2013-06-07 12:09 ` Benjamin Herrenschmidt
[not found] ` <1370642488.6813.15@snotra>
2013-06-07 22:02 ` Scott Wood
2013-06-07 22:09 ` Benjamin Herrenschmidt
2013-06-07 22:34 ` Scott Wood
2013-06-07 22:39 ` Benjamin Herrenschmidt
2013-06-07 23:29 ` Scott Wood
2013-06-07 23:33 ` Benjamin Herrenschmidt
2013-06-07 12:09 ` SATA FSL and upstreaming Timur Tabi
2013-05-16 6:17 ` Zang Roy-R61911
2013-05-16 6:23 ` Benjamin Herrenschmidt
2013-05-16 6:33 ` Bhushan Bharat-R65777
2013-05-16 6:34 ` Benjamin Herrenschmidt
2013-05-16 6:35 ` Zang Roy-R61911
2013-05-16 6:37 ` Benjamin Herrenschmidt
2013-05-16 6:41 ` tiejun.chen
2013-05-16 6:48 ` Zang Roy-R61911
2013-05-16 6:24 ` Xie Shaohui-B21989
2013-05-16 6:31 ` Benjamin Herrenschmidt
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1370607245.3766.386.camel@pasglop \
--to=benh@kernel.crashing.org \
--cc=B21989@freescale.com \
--cc=B32616@freescale.com \
--cc=David.Laight@ACULAB.COM \
--cc=R65777@freescale.com \
--cc=afleming@freescale.com \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=r61911@freescale.com \
--cc=tiejun.chen@windriver.com \
--cc=timur@tabi.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.