Prev: Warning -- CLAX is *UNDER CONSTRUCTION*
Next: Position : Business Analyst at Iowa City, IA
From: Harold Aptroot on 19 Jan 2010 11:56
It now appears to be failing silently (or can it take multiple hours?)
I didn't get a confirmation email either (are you supposed to get it?)

Anyway, I'll just dump my post here on the off chance that someone cares
about it..
I hope you don't mind :)

Re: Instruction placement
Qyz wrote:
> Hi all,
> Is there anything significant about instructions that fall across 32-
> byte boundaries?
> For instance, do they slow things down potentially, or are they in
> some way a security risk?
> Thanks.

What architecture?
Since crossing a 32byte coundary means crossing a 16byte boundary, on
Core2's they would usually cause a small slowdown, and in some cases a
big slowdown, such as when the instruction that crosses the boundary has
a length changing prefix (66h or 67h) or when a "TEST-like" instruction
NEG, NOT, DIV, IDIV, MUL or IMUL) with a single 16bit operand is
seperated from its ModRM byte by the 16byte boundary (because it looks
like a TEST with a length changing prefix and the CPU cannot know that
isn't without seeing the ModRM byte)

Or at least that's how I remember it, if you want to know for sure, look
in Agner Fog's microarchitecture document

-- harold

From: Nathan Baker on 19 Jan 2010 13:51

"Harold Aptroot" <harold.aptroot(a)gmail.com> wrote in message
news:e4ec1$4b55e437$5355b09d$11807(a)cache100.multikabel.net...
> It now appears to be failing silently (or can it take multiple hours?)

It is in the message cache and I will post it ... um.. sometime. Thank
you for your... um... patience.

> I didn't get a confirmation email either (are you supposed to get it?)

No. Did you get one for every a.l.a post that you have made???

>
> Anyway, I'll just dump my post here on the off chance that someone cares
> about it..
> I hope you don't mind :)

*So important* that it can't wait, huh??? Thank you for being the example
of how technology spoils adults and turns them into children. ;)

Sorry for the sloppy service at this bar, but we did hang some warning signs
on the front door. If the hammering and sawdust "rattle your nerves," then
I suggest you visit the bar across the street.

Nathan.


From: Frank Kotler on 19 Jan 2010 14:12
Harold Aptroot wrote:
> It now appears to be failing silently (or can it take multiple hours?)

Sorry. Took me several retries (fussy about that header, apparently!).
Should be there now.

> I didn't get a confirmation email either (are you supposed to get it?)

We've never done that. If your post were rejected "for cause", you would
supposedly get a mail with the reason. The only time this has happened
to me, Rod was "arguing with the moderators". Wasn't his fault, we
started it, but it still shouldn't have continued. Since we don't have
Rod's email, we couldn't "take it to private mail", as should have been
done... So, sorry again, Rod.

Some groups do send a "your comment is awaiting moderator approval"
mail. Perhaps we could do this, if people wanted it... but not right now...

> Anyway, I'll just dump my post here on the off chance that someone cares
> about it..
> I hope you don't mind :)

Beautiful part of the "alt" hierarchy - if anybody does mind, there's
nothing they can do about it! :)

Best,
Frank

From: Harold Aptroot on 19 Jan 2010 14:50
Ah I'm sorry.. From here it looked like nothing was happening with no
indication of anything at all, it might as well have failed like it did all
other times. I'm not super impatient, I posted here just in case posting to
clax would fail alltogether. Without confirmation there was no evidence to
the contrary, and due to the failure of my ISP earlier I just didn't trust
any posts to come through anymore.
I really do apologize..

Harold

From: Frank Kotler on 19 Jan 2010 16:22
Harold Aptroot wrote:
> Ah I'm sorry.. From here it looked like nothing was happening with no
> indication of anything at all, it might as well have failed like it did
> all other times. I'm not super impatient, I posted here just in case
> posting to clax would fail alltogether. Without confirmation there was
> no evidence to the contrary, and due to the failure of my ISP earlier I
> just didn't trust any posts to come through anymore.
> I really do apologize..

No need to apologize, Harold. Nathan's been tearing his hair out trying
to make this go, and it's makin' him irritable. :)

I'm afraid we're all going to have to be patient for a while longer yet,
but OTOH, we need your feedback when things are going wrong, too. In
this case, your original problem was solved - your post made it as far
as clax-submit(a)inspiretomorrow.net and I "approved" it (after a
while)... but I screwed it up and it didn't appear... I waited a while
to see if it would... it didn't... tried it again with CRLF pairs where
they need to be, no this one was the "Newsgroups:" header was mangled, I
think. Anyway, it finally posted... I think it went right... *I* can see
it, if you still can't, please *do* complain!

You wouldn't *believe* the cobbed-together cluster of tools we're using
to do this! I hope service will become speedier... and more
convenient... as we improve. If we improve. :)

Best,
Frank
 |  Next  |  Last
Pages: 1 2
Prev: Warning -- CLAX is *UNDER CONSTRUCTION*
Next: Position : Business Analyst at Iowa City, IA