Path: csiph.com!x330-a1.tempe.blueboxinc.net!newsfeed.hal-mli.net!feeder3.hal-mli.net!newsfeed.hal-mli.net!feeder1.hal-mli.net!border3.nntp.dca.giganews.com!Xl.tags.giganews.com!border1.nntp.dca.giganews.com!nntp.giganews.com!local2.nntp.dca.giganews.com!news.giganews.com.POSTED!not-for-mail
NNTP-Posting-Date: Fri, 10 Feb 2012 08:37:46 -0600
Message-ID: <4F352BB3.30806@SPAM.comp-arch.net>
Date: Fri, 10 Feb 2012 06:37:39 -0800
From: "Andy (Super) Glew" <andy@SPAM.comp-arch.net>
Reply-To: andy@SPAM.comp-arch.net
Organization: comp-arch.net
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:9.0) Gecko/20111222 Thunderbird/9.0.1
MIME-Version: 1.0
Newsgroups: comp.arch
Subject: Re: Register file splits, a new look.
References: <ggtgp-1DA77E.06332706022012@netnews.mchsi.com> <7z1uq4qpf4.fsf@ask.diku.dk> <21254419.964.1328818372560.JavaMail.geo-discussion-forums@yqad38> <jh2j7v$18n$1@gosset.csi.cam.ac.uk>
In-Reply-To: <jh2j7v$18n$1@gosset.csi.cam.ac.uk>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Lines: 19
X-Usenet-Provider: http://www.giganews.com
X-Trace: sv3-GHl6zZXZ0/EN4tInyeHcggQYOagGYeAW9kuB/kBjqtZwVtDoiGifylBSM/i7yLFiLrvnPbXEzMzkbZI!m6c4ld/rjVTy7Xgtzs/Dlr6P36hBhqPCs0wowscdBwTh5FNcZZs/VF4Bjhhx
X-Complaints-To: abuse@giganews.com
X-DMCA-Notifications: http://www.giganews.com/info/dmca.html
X-Abuse-and-DMCA-Info: Please be sure to forward a copy of ALL headers
X-Abuse-and-DMCA-Info: Otherwise we will be unable to process your complaint properly
X-Postfilter: 1.3.40
X-Original-Bytes: 2318
Xref: x330-a1.tempe.blueboxinc.net comp.arch:5861

On 2/10/2012 12:07 AM, nmm1@cam.ac.uk wrote:
> In article<21254419.964.1328818372560.JavaMail.geo-discussion-forums@yqad38>,
> MitchAlsup<MitchAlsup@aol.com>  wrote:
>>
>> In several machines I have worked on, the register files (plus reorder
>> buffers, and buffering) were bigger than the computational units
>> attached to them! All of these were separate file machines. A lot of
>> power is heading down this hole.
>
> Yes.  It's one of the reasons that I think we should stop digging
> and look for a way out.  My understanding is that most of that is
> to attempt to maximise the performance of serial, in-order code,
> by making the low-level parallelism and out-of-order logic entirely
> transparent.  I think that's a mug's game, and the only solution is
> to move to better ISAs which, in turn, needs us to move to better
> programming paradigms and languages.

If those "better ISAs" require significantly larger register files, then 
they may lose.