Prev: Found a solution
Next: where to define unicode
From: Peter Olcott on 14 Apr 2010 19:04
I finally got what you were saying about the reason why
Lambda approaching Mu results in an infinite queue length.

I thought about it some more and suddenly had an epiphany
that you were right, it must be the idle time that would be
generated by the exponential distribution of arrivals.

I also am leaning towards your idea of not turning of drive
write buffers, the weight of reasoning seems to support
this.

Respectfully,


Peter Olcott


From: Hector Santos on 14 Apr 2010 19:39
Why don't you move this topic to the Microsoft Win32 kernel
forum or the proper Linux forum iff this is for linux development with
no consideration whatsoever for Windows.

This question is redundant and has nothing to do with MFC unless you
wish to utilized the MFC collection classes and other WIN32 related
features and need help in using them. In that case, I'm sure there
were many to help. But you been told numerous times about every
concept there is about this, including IOCP, Worker Pool concepts,
optimized interlocked singly linked lists queues, leveraging threads
and memory maps, including if you wish to use named pipes! Classes
were provided, links to the pertinent topics on Google.Net were
provided, even links to source code were provided.

Enough!

Please lets not feed the troll again.


Peter Olcott wrote:

> I finally got what you were saying about the reason why
> Lambda approaching Mu results in an infinite queue length.
>
> I thought about it some more and suddenly had an epiphany
> that you were right, it must be the idle time that would be
> generated by the exponential distribution of arrivals.
>
> I also am leaning towards your idea of not turning of drive
> write buffers, the weight of reasoning seems to support
> this.
>
> Respectfully,
>
>
> Peter Olcott
>
>



--
HLS
From: Peter Olcott on 14 Apr 2010 19:59

"Hector Santos" <sant9442(a)nospam.gmail.com> wrote in message
news:euYP0uC3KHA.4964(a)TK2MSFTNGP05.phx.gbl...
> Why don't you move this topic to the Microsoft Win32
> kernel
> forum or the proper Linux forum iff this is for linux
> development with no consideration whatsoever for Windows.
>

Since the question of exactly why queue length approaches
infinity when Lambda approaches Mu had a simple answer, I am
now seeking an equivalent simple answer of why SQMS is more
efficient than MQMS.

Exactly what is it about a SQMS that makes it much more
efficient than MQMS?


From: Peter Olcott on 14 Apr 2010 20:25

"Hector Santos" <sant9442(a)nospam.gmail.com> wrote in message
news:euYP0uC3KHA.4964(a)TK2MSFTNGP05.phx.gbl...
> Why don't you move this topic to the Microsoft Win32
> kernel
> forum or the proper Linux forum iff this is for linux
> development with no consideration whatsoever for Windows.

I am where I get the answers that I need. I have just
widened my search to other groups. I have been speaking with
the Linux/Unix people for quite a while. So far this group
has the most technically adept people, plus I have been here
for many years.

I finally have the essence of Lambda approaching Mu.

>
> This question is redundant and has nothing to do with MFC
> unless you wish to utilized the MFC collection classes and
> other WIN32 related features and need help in using them.
> In that case, I'm sure there were many to help. But you
> been told numerous times about every concept there is
> about this, including IOCP, Worker Pool concepts,
> optimized interlocked singly linked lists queues,
> leveraging threads and memory maps, including if you wish
> to use named pipes! Classes were provided, links to the
> pertinent topics on Google.Net were provided, even links
> to source code were provided.
>
> Enough!
>
> Please lets not feed the troll again.
>
>
> Peter Olcott wrote:
>
>> I finally got what you were saying about the reason why
>> Lambda approaching Mu results in an infinite queue
>> length.
>>
>> I thought about it some more and suddenly had an epiphany
>> that you were right, it must be the idle time that would
>> be generated by the exponential distribution of arrivals.
>>
>> I also am leaning towards your idea of not turning of
>> drive write buffers, the weight of reasoning seems to
>> support this.
>>
>> Respectfully,
>>
>>
>> Peter Olcott
>>
>>
>
>
>
> --
> HLS


 | 
Pages: 1
Prev: Found a solution
Next: where to define unicode