Skip to Content.
Sympa Menu

sm-discuss - Re: [SM-Discuss] Smallest Source Mage

sm-discuss AT lists.ibiblio.org

Subject: Public SourceMage Discussion List

List archive

Chronological Thread  
  • From: Phil/CERisE/KG6MBQ <cerise AT littlegreenmen.armory.com>
  • To: sm-discuss AT lists.ibiblio.org
  • Subject: Re: [SM-Discuss] Smallest Source Mage
  • Date: Mon, 29 Jul 2002 17:32:19 -0700 (PDT)

Eric Sandall said:

> The times I posted showed that tmpfs used less system time, but more user
> and real time. So, having no-tmpfs is not an overall winner. And the
> memory issue (I agree, has not been proven either way) will be important
> to those of us using systems with little RAM (I recall someone having 8Mb
> on one posting), and so is not trivial to us.

I'll grant that I haven't looked at them recently, but I recall less
elapsed time. Since there's several components of the system involved,
that's the only field that you can really judge the speed by.
I'm certainly not saying that RAM usage is trivial. Indeed, I think
that since non-tmpfs will use less RAM.

> Yes, you may say, "Well, that's an antiquated machine, you should update,
> computers are cheap!" Yes, they are cheap, but not all of us have the
> resources to buy new machines whenever we need more speed. I use every
> computer I have, and buy "junk" machines for cheap, sometimes even free,
> so that I may add another computer to my cluster. Though it may not add
> much, it can take a little pressure off of my server for the real services
> (Counter-Strike, NeverWinter Nights, SETI@Home, etc. :)).

I'd sooner shoot myself than make that statement.
There's no doubt about the value of efficiency! All the more reason
to use the hard disk because it trivially uses less memory and spends
less time on compilation (thus implying less disk access).
Of course, the claim about memory is unproven, but I think it's
pretty clear. I'd be surprised if it turned the other way 'round.

> The reason I said that is why change it? There are people here who
> believe (even if it's wrong, which also depends on whom you're asking)
> that tmpfs is the ultimate in compilation practice, while others say that
> tmpfs is a waste of resources and slows everything down.
>
> Had you read what I said you would have seen that I mention there be an
> option, clearly visible (it's in the first screen with the Language and
> such) for the user to choose. The default only occurs if the user chooses
> not to change it.

No, I read what you said just fine.
For my part, I haven't seen anyone providing any reasons why tmpfs
is "the ultimate in compilation practice." Although I have no doubt
that they absolutely "believe" that it's better.
As a result, I don't see the value in including it as a choice.
Hence the comment about slow down compilation 5%...

> Chris has already submitted a patch to add the choice to the sorcery menu,
> now all we need is one of the ISO hackers to add the choice to the install
> menu (or have an eager volunteer from the crowd <hint><hint> :)) submit
> the patch to bugzilla.
>
> Somehow, we always come back to tmpfs lately. How 'bout we just close
> this door since the patch is there and have Anders or Nathan (or Chris,
> someone) put this into sorcery so we can end this. Those of you whom do
> not want tmpfs, fine, you can now disable tmpfs usage in sorcery, those
> who do want tmpfs, also fine, just do nothing. :)
>
> (Non important stuff...)
> We do not need to bring in ludicrous statements to this discussion, such
> as your 5% for fun. Do you have something against freedom of choice? If
> the user has the choice, what matters if tmpfs is the default? The user
> can change it if he/she wants. Go ahead and make no-tmpfs default, I
> don't really care, I just want the option of choosing.

It's not ludicrous according to the figures as posted on here...I
figure that 5% is about the average difference in elapsed times between
tmpfs and the hard disk.
I just don't see a lot of reason to including the choice of a
slower, more memory-intensive method which doesn't earn one anything.
Or, at least, if it does earn one anything, there haven't been any
numbers or analyses to that effect.
I don't recall why this one took a turn back. I think it was
related to "These things should happen", but I am not absolute on this.
I was just responding ; )

-Phil/CERisE





Archive powered by MHonArc 2.6.24.

Top of Page