Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.sys.mac.hardware.misc > #170 > unrolled thread

Memory lock failed

Back to article view | Back to comp.sys.mac.hardware.misc

Contents

  Memory lock failed Francois <xanrov@numericable.fr> - 2011-08-19 22:10 +0200
    Re: Memory lock failed philo <philo@invalid.not> - 2011-08-20 07:45 -0500
      Re: Memory lock failed Francois <danzas@9online.fr> - 2011-08-20 17:08 +0200
        Re: Memory lock failed philo <philo@invalid.not> - 2011-08-20 16:11 -0500
      Re: Memory lock failed Howard Brazee <howard@brazee.net> - 2011-08-20 18:06 -0600
        Re: Memory lock failed Tom Stiller <tom_stiller@yahoo.com> - 2011-08-20 21:04 -0400
          Re: Memory lock failed Howard Brazee <howard@brazee.net> - 2011-08-21 16:10 -0600
            Re: Memory lock failed Tom Stiller <tom_stiller@yahoo.com> - 2011-08-21 19:04 -0400
      Re: Memory lock failed dempson@actrix.gen.nz (David Empson) - 2011-08-21 12:56 +1200
        Re: Memory lock failed philo <philo@invalid.not> - 2011-08-20 21:37 -0500
          Re: Memory lock failed dempson@actrix.gen.nz (David Empson) - 2011-08-21 17:17 +1200
            Re: Memory lock failed nmassello@yahoo.com (Neill Massello) - 2011-08-21 02:32 -0600
              Re: Memory lock failed dempson@actrix.gen.nz (David Empson) - 2011-08-22 00:38 +1200
                Re: Memory lock failed Tom Stiller <tom_stiller@yahoo.com> - 2011-08-21 09:18 -0400
                  Re: Memory lock failed dempson@actrix.gen.nz (David Empson) - 2011-08-22 01:34 +1200
            Re: Memory lock failed philo <philo@invalid.not> - 2011-08-21 07:15 -0500
              Re: Memory lock failed dempson@actrix.gen.nz (David Empson) - 2011-08-22 01:04 +1200
                Re: Memory lock failed philo <philo@invalid.not> - 2011-08-21 08:27 -0500
    Re: Memory lock failed Jolly Roger <jollyroger@pobox.com> - 2011-08-20 15:13 +0000
      Re: Memory lock failed Francois <danzas@9online.fr> - 2011-08-21 10:14 +0200

#170 — Memory lock failed

I had recently memory problems on my Mac Pro (early 2008), and I had to 
replace some 4GB DIMM's to recover my original 24 GB RAM. This Ram 
consists now of 4 slots with 2GB ( risers A & B) and 4 with 4GB ( also 
half on A , half on B). The makers are Crucial (2x4 GB) and "elixir" for 
the rest. Of course I ran memtest  ( in single user mode ) after each 
memory change ( there were several). Now ,every time, memtest log said :
        "Attempted memory lock... ERROR Memory lock failed - reason unknown
WARNING : Testing with unlocked memory may be slower and less reliable"

The 3 test sequence was every time successful, and the computer behaves 
normally.
Anyway I'm baffled : why is my memory not "lockable" ?
Has this something to see with the placement of the memory sticks on the 
risers ?
Thanks for any hint, although that doesn't prevent me to sleep.

[toc] | [next] | [standalone]

#171

On 08/19/2011 03:10 PM, Francois wrote:
> I had recently memory problems on my Mac Pro (early 2008), and I had to
> replace some 4GB DIMM's to recover my original 24 GB RAM. This Ram
> consists now of 4 slots with 2GB ( risers A & B) and 4 with 4GB ( also
> half on A , half on B). The makers are Crucial (2x4 GB) and "elixir" for
> the rest. Of course I ran memtest ( in single user mode ) after each
> memory change ( there were several). Now ,every time, memtest log said :
> "Attempted memory lock... ERROR Memory lock failed - reason unknown
> WARNING : Testing with unlocked memory may be slower and less reliable"
>
> The 3 test sequence was every time successful, and the computer behaves
> normally.
> Anyway I'm baffled : why is my memory not "lockable" ?
> Has this something to see with the placement of the memory sticks on the
> risers ?
> Thanks for any hint, although that doesn't prevent me to sleep.



More than likely you are booting to single mode using the 32bit kernel
which cannot address more than 4 gigs of ram.

To access the full amount of ram you will need to boot to the 64bit kernel



Supposedly you can boot the the 64 bit kernel by holding down the "6" 
and the "4" keys while booting

I am no mac expert though

[toc] | [prev] | [next] | [standalone]

#173

On 20/08/2011 14:45, philo wrote:

>
>
>
> More than likely you are booting to single mode using the 32bit kernel
> which cannot address more than 4 gigs of ram.
>
> To access the full amount of ram you will need to boot to the 64bit kernel
>
>
>
> Supposedly you can boot the the 64 bit kernel by holding down the "6"
> and the "4" keys while booting
>
> I am no mac expert though

You are right on, philo. When I ran those memtest tests, I was under 
Snow Leopard, and maybe the kernel that appeared in Single User Mode 
(SUM) was 32 bits. Nevertheless memtest took nearly all the memory fro 
testing it (about 23 GB) but could not, of course lock it.
  Now I am under  Lion ( OS X 10.7.1 ) and, when under SUM, it is well 
specified that the kernel is 64 bits : "Kernel is LP64"
When running memtest, it says : "Attempting memory lock ... locked 
successfully".
Thanks to you this "lock mystery" is solved and I'll be much more 
confident in memtest results in the future.
Very best regards.

[toc] | [prev] | [next] | [standalone]

#175

On 08/20/2011 10:08 AM, Francois wrote:
> On 20/08/2011 14:45, philo wrote:
>
>>
>>



<snip>
>> I am no mac expert though
>
> You are right on, philo. When I ran those memtest tests, I was under
> Snow Leopard, and maybe the kernel that appeared in Single User Mode
> (SUM) was 32 bits. Nevertheless memtest took nearly all the memory fro
> testing it (about 23 GB) but could not, of course lock it.
> Now I am under Lion ( OS X 10.7.1 ) and, when under SUM, it is well
> specified that the kernel is 64 bits : "Kernel is LP64"
> When running memtest, it says : "Attempting memory lock ... locked
> successfully".
> Thanks to you this "lock mystery" is solved and I'll be much more
> confident in memtest results in the future.
> Very best regards.



Thanks for the feedback...

Though I have been working on PC's for years
my knowledge of macs are not that great...
but I'm learning.

Glad to know I got this one right

[toc] | [prev] | [next] | [standalone]

#176

On Sat, 20 Aug 2011 07:45:51 -0500, philo <philo@invalid.not> wrote:

>Supposedly you can boot the the 64 bit kernel by holding down the "6" 
>and the "4" keys while booting

I've got one program which won't run because of my default set-up.
It's an add-in to iTunes called Jewel Case.     When I start iTunes, I
get a message saying:

"JewelCase" could not Load in 64-bit mode.

Try selecting "Open in 32-bit mode" in the Finder's info window for
iTunes.

I haven't been able to find that option.

-- 
"In no part of the constitution is more wisdom to be found,
than in the clause which confides the question of war or peace 
to the legislature, and not to the executive department." 

- James Madison

[toc] | [prev] | [next] | [standalone]

#178

In article <fqi0571emjm8idsl6t7iceogr6uvoihb2d@4ax.com>,
 Howard Brazee <howard@brazee.net> wrote:

> On Sat, 20 Aug 2011 07:45:51 -0500, philo <philo@invalid.not> wrote:
> 
> >Supposedly you can boot the the 64 bit kernel by holding down the "6" 
> >and the "4" keys while booting
> 
> I've got one program which won't run because of my default set-up.
> It's an add-in to iTunes called Jewel Case.     When I start iTunes, I
> get a message saying:
> 
> "JewelCase" could not Load in 64-bit mode.
> 
> Try selecting "Open in 32-bit mode" in the Finder's info window for
> iTunes.
> 
> I haven't been able to find that option.

Display iTunes' Finder info window (Cmd-I) and look toward the bottom of 
the "General" pane.

-- 
PRAY, v.  To ask that the laws of the universe be annulled in behalf
of a single petitioner confessedly unworthy. -- Ambrose Bierce

[toc] | [prev] | [next] | [standalone]

#189

On Sat, 20 Aug 2011 21:04:06 -0400, Tom Stiller
<tom_stiller@yahoo.com> wrote:

>> "JewelCase" could not Load in 64-bit mode.
>> 
>> Try selecting "Open in 32-bit mode" in the Finder's info window for
>> iTunes.
>> 
>> I haven't been able to find that option.
>
>Display iTunes' Finder info window (Cmd-I) and look toward the bottom of 
>the "General" pane.

Hmm, before I posted, I opened Applications with Finder (my default is
Path Finder), but I didn't look at the info window from there.  

Apparently the info window is different in Finder and in Path Finder,
and this time I found your setting.

So I changed the setting, quit iTunes, and opened it again.   I no
longer get that error, but I also don't get JewelCase.

-- 
"In no part of the constitution is more wisdom to be found,
than in the clause which confides the question of war or peace 
to the legislature, and not to the executive department." 

- James Madison

[toc] | [prev] | [next] | [standalone]

#190

In article <360357dqtp3fv3lgq1843dr73ujtr99pcg@4ax.com>,
 Howard Brazee <howard@brazee.net> wrote:

> On Sat, 20 Aug 2011 21:04:06 -0400, Tom Stiller
> <tom_stiller@yahoo.com> wrote:
> 
> >> "JewelCase" could not Load in 64-bit mode.
> >> 
> >> Try selecting "Open in 32-bit mode" in the Finder's info window for
> >> iTunes.
> >> 
> >> I haven't been able to find that option.
> >
> >Display iTunes' Finder info window (Cmd-I) and look toward the bottom of 
> >the "General" pane.
> 
> Hmm, before I posted, I opened Applications with Finder (my default is
> Path Finder), but I didn't look at the info window from there.  
> 
> Apparently the info window is different in Finder and in Path Finder,
> and this time I found your setting.
> 
> So I changed the setting, quit iTunes, and opened it again.   I no
> longer get that error, but I also don't get JewelCase.

Possibly the APIs have changed in the 4 years since JewelCase was last 
updated.

-- 
PRAY, v.  To ask that the laws of the universe be annulled in behalf
of a single petitioner confessedly unworthy. -- Ambrose Bierce

[toc] | [prev] | [next] | [standalone]

#177

philo <philo@invalid.not> wrote:

> On 08/19/2011 03:10 PM, Francois wrote:
> > I had recently memory problems on my Mac Pro (early 2008), and I had to
> > replace some 4GB DIMM's to recover my original 24 GB RAM. This Ram
> > consists now of 4 slots with 2GB ( risers A & B) and 4 with 4GB ( also
> > half on A , half on B). The makers are Crucial (2x4 GB) and "elixir" for
> > the rest. Of course I ran memtest ( in single user mode ) after each
> > memory change ( there were several). Now ,every time, memtest log said :
> > "Attempted memory lock... ERROR Memory lock failed - reason unknown
> > WARNING : Testing with unlocked memory may be slower and less reliable"
> >
> > The 3 test sequence was every time successful, and the computer behaves
> > normally.
> > Anyway I'm baffled : why is my memory not "lockable" ?
> > Has this something to see with the placement of the memory sticks on the
> > risers ?
> > Thanks for any hint, although that doesn't prevent me to sleep.
> 
> 
> 
> More than likely you are booting to single mode using the 32bit kernel
> which cannot address more than 4 gigs of ram.

Not exactly. The 32-bit kernel on Mac OS X cannot directly address more
than 4 GB at a time, but through the use of PAE it is able to manage up
to 32 GB of RAM, and selectively access all of it.

You can have 64-bit applications running on Mac OS X with 32-bit kernel,
and in those conditions the 64-bit applications can directly address
more than 4 GB of RAM.

The "lock" mechanism used by memtest is very likely to be related to the
4 GB directly addressable limit of the 32-bit kernel but I'm not
familiar enough with that level of detail to explain further.

-- 
David Empson
dempson@actrix.gen.nz

[toc] | [prev] | [next] | [standalone]

#179

On 08/20/2011 07:56 PM, David Empson wrote:
> philo<philo@invalid.not>  wrote:
>
>> On 08/19/2011 03:10 PM, Francois wrote:
>>> I had recently memory problems on my Mac Pro (early 2008), and I had to
>>> replace some 4GB DIMM's to recover my original 24 GB RAM. This Ram
>>> consists now of 4 slots with 2GB ( risers A&  B) and 4 with 4GB ( also
>>> half on A , half on B). The makers are Crucial (2x4 GB) and "elixir" for
>>> the rest. Of course I ran memtest ( in single user mode ) after each
>>> memory change ( there were several). Now ,every time, memtest log said :
>>> "Attempted memory lock... ERROR Memory lock failed - reason unknown
>>> WARNING : Testing with unlocked memory may be slower and less reliable"
>>>
>>> The 3 test sequence was every time successful, and the computer behaves
>>> normally.
>>> Anyway I'm baffled : why is my memory not "lockable" ?
>>> Has this something to see with the placement of the memory sticks on the
>>> risers ?
>>> Thanks for any hint, although that doesn't prevent me to sleep.
>>
>>
>>
>> More than likely you are booting to single mode using the 32bit kernel
>> which cannot address more than 4 gigs of ram.
>
> Not exactly. The 32-bit kernel on Mac OS X cannot directly address more
> than 4 GB at a time, but through the use of PAE it is able to manage up
> to 32 GB of RAM, and selectively access all of it.
>
> You can have 64-bit applications running on Mac OS X with 32-bit kernel,
> and in those conditions the 64-bit applications can directly address
> more than 4 GB of RAM.
>
> The "lock" mechanism used by memtest is very likely to be related to the
> 4 GB directly addressable limit of the 32-bit kernel but I'm not
> familiar enough with that level of detail to explain further.
>


Not a function of being a mac...
it's the same for any 32 bit os  vs 64 bit

very few folks use PAE...
at any rate, I am sure the OP does not have PAE enabled
nor do I know if his h/w supports it or not

[toc] | [prev] | [next] | [standalone]

#180

philo <philo@invalid.not> wrote:

> On 08/20/2011 07:56 PM, David Empson wrote:
> > philo<philo@invalid.not>  wrote:
> >
> >> On 08/19/2011 03:10 PM, Francois wrote:
> >>> I had recently memory problems on my Mac Pro (early 2008), and I had to
> >>> replace some 4GB DIMM's to recover my original 24 GB RAM. This Ram
> >>> consists now of 4 slots with 2GB ( risers A&  B) and 4 with 4GB ( also
> >>> half on A , half on B). The makers are Crucial (2x4 GB) and "elixir" for
> >>> the rest. Of course I ran memtest ( in single user mode ) after each
> >>> memory change ( there were several). Now ,every time, memtest log said :
> >>> "Attempted memory lock... ERROR Memory lock failed - reason unknown
> >>> WARNING : Testing with unlocked memory may be slower and less reliable"
> >>>
> >>> The 3 test sequence was every time successful, and the computer behaves
> >>> normally.
> >>> Anyway I'm baffled : why is my memory not "lockable" ?
> >>> Has this something to see with the placement of the memory sticks on the
> >>> risers ?
> >>> Thanks for any hint, although that doesn't prevent me to sleep.
> >>
> >>
> >>
> >> More than likely you are booting to single mode using the 32bit kernel
> >> which cannot address more than 4 gigs of ram.
> >
> > Not exactly. The 32-bit kernel on Mac OS X cannot directly address more
> > than 4 GB at a time, but through the use of PAE it is able to manage up
> > to 32 GB of RAM, and selectively access all of it.
> >
> > You can have 64-bit applications running on Mac OS X with 32-bit kernel,
> > and in those conditions the 64-bit applications can directly address
> > more than 4 GB of RAM.
> >
> > The "lock" mechanism used by memtest is very likely to be related to the
> > 4 GB directly addressable limit of the 32-bit kernel but I'm not
> > familiar enough with that level of detail to explain further.
> 
> Not a function of being a mac...
> it's the same for any 32 bit os  vs 64 bit

Linux and Mac OS X use PAE to allow a 32-bit kernel to manage an address
space larger than 4 GB on a 64-bit processor. Windows doesn't: it has
separate 32-bit and 64-bit versions of the OS.

> very few folks use PAE...
> at any rate, I am sure the OP does not have PAE enabled
> nor do I know if his h/w supports it or not

PAE is a standard feature of Mac OS X when using a 32-bit kernel on an
Intel Mac with a 64-bit processor, and PAE is supported on all 64-bit
processors that Apple has used in Intel Macs. The user has no control
over whether or not PAE is being used by the 32-bit kernel (I'm not
aware of any low level mechanism to disable it). If more than 4 GB of
address space is available, the kernel will use PAE.

Apple has been supplying Intel Macs with 64-bit processors since August
2006 (original Mac Pro). That model supported up to 32 GB of RAM while
running Mac OS X 10.4, which only has a 32-bit kernel (as does 10.5).
The 32-bit kernel in 10.4 and later is able to manage up to 32 GB of RAM
via PAE, and 64-bit applications are able to use all of the RAM.

The first version of Mac OS X which had a 64-bit kernel is Snow Leopard
(Mac OS X 10.6), which was released in August 2009. Most Macs running
Snow Leopard still use the 32-bit kernel, including those with 8 GB or
more of RAM. The 64-bit kernel is only supported on selected models from
2008 and later, despite several of the older models being able to use
more than 4 GB of RAM.

In Snow Leopard, the only models that default to 64-bit kernel are
recent Mac Pro, Xserve and Mac Mini Server models. Lion defaults to
using the 64-bit kernel on a wider range of models, but it still has the
32-bit kernel as an option for those models, the default for several
more models, and it is the only option for older models.
-- 
David Empson
dempson@actrix.gen.nz

[toc] | [prev] | [next] | [standalone]

#182

David Empson <dempson@actrix.gen.nz> wrote:

> In Snow Leopard, the only models that default to 64-bit kernel are
> recent Mac Pro, Xserve and Mac Mini Server models. Lion defaults to
> using the 64-bit kernel on a wider range of models, but it still has the
> 32-bit kernel as an option for those models, the default for several
> more models, and it is the only option for older models.

The following page lists the older Macs that can boot to the 64-bit
kernel in Snow Leopard. 

<http://www.ahatfullofsky.comuv.com/English/Programs/SMS/SMS.html>

[toc] | [prev] | [next] | [standalone]

#184

Neill Massello <nmassello@yahoo.com> wrote:

> David Empson <dempson@actrix.gen.nz> wrote:
> 
> > In Snow Leopard, the only models that default to 64-bit kernel are
> > recent Mac Pro, Xserve and Mac Mini Server models. Lion defaults to
> > using the 64-bit kernel on a wider range of models, but it still has the
> > 32-bit kernel as an option for those models, the default for several
> > more models, and it is the only option for older models.
> 
> The following page lists the older Macs that can boot to the 64-bit
> kernel in Snow Leopard. 
> 
> <http://www.ahatfullofsky.comuv.com/English/Programs/SMS/SMS.html>

The list on that page is out of date, but it does cover old models. All
Xserve, Mac Pro, MacBook Pro and iMac models introduced in 2008 or later
are capable of running the 64-bit kernel.

This page covers Apple's official position for 10.6:

http://support.apple.com/kb/ht3770

Mac OS X 10.6 non-server defaults to 64-bit kernel for the mid 2010 Mac
Pro, early 2011 MacBook Pro and mid 2011 iMac.

Mac OS X Server 10.6 defaults to 64-bit kernel for the Xserve (early
2008 and later), Mac Pro (early 2008 and later), and Mac Mini Server
(mid 2010).

The same Mac Pro and Mac Mini models can run 64-bit kernel with Mac OS X
10.6 non-server but default to 32-bit. This also applies to the iMac and
MacBook Pro from early 2008.

No MacBook can use the 64-bit kernel, nor can the Mac Mini in the
original body design (late 2009 or earlier models).

I haven't found an official statement from Apple on which models default
to 64-bit kernel with Lion, but my impression is that it has been
extended to more models than with Snow Leopard.

-- 
David Empson
dempson@actrix.gen.nz

[toc] | [prev] | [next] | [standalone]

#186

In article <1k6dj99.1pwqm1d1d892j9N%dempson@actrix.gen.nz>,
 dempson@actrix.gen.nz (David Empson) wrote:

> Neill Massello <nmassello@yahoo.com> wrote:
> 
> > David Empson <dempson@actrix.gen.nz> wrote:
> > 
> > > In Snow Leopard, the only models that default to 64-bit kernel are
> > > recent Mac Pro, Xserve and Mac Mini Server models. Lion defaults to
> > > using the 64-bit kernel on a wider range of models, but it still has the
> > > 32-bit kernel as an option for those models, the default for several
> > > more models, and it is the only option for older models.
> > 
> > The following page lists the older Macs that can boot to the 64-bit
> > kernel in Snow Leopard. 
> > 
> > <http://www.ahatfullofsky.comuv.com/English/Programs/SMS/SMS.html>
> 
> The list on that page is out of date, but it does cover old models. All
> Xserve, Mac Pro, MacBook Pro and iMac models introduced in 2008 or later
> are capable of running the 64-bit kernel.
> 
> This page covers Apple's official position for 10.6:
> 
> http://support.apple.com/kb/ht3770
> 
> Mac OS X 10.6 non-server defaults to 64-bit kernel for the mid 2010 Mac
> Pro, early 2011 MacBook Pro and mid 2011 iMac.
> 
> Mac OS X Server 10.6 defaults to 64-bit kernel for the Xserve (early
> 2008 and later), Mac Pro (early 2008 and later), and Mac Mini Server
> (mid 2010).
> 
> The same Mac Pro and Mac Mini models can run 64-bit kernel with Mac OS X
> 10.6 non-server but default to 32-bit. This also applies to the iMac and
> MacBook Pro from early 2008.
> 
> No MacBook can use the 64-bit kernel, nor can the Mac Mini in the
> original body design (late 2009 or earlier models).
> 
> I haven't found an official statement from Apple on which models default
> to 64-bit kernel with Lion, but my impression is that it has been
> extended to more models than with Snow Leopard.

I can verify that Lion runs a 64-bit kernel on my iMac7,1 where Snow 
Leopard would not.

-- 
PRAY, v.  To ask that the laws of the universe be annulled in behalf
of a single petitioner confessedly unworthy. -- Ambrose Bierce

[toc] | [prev] | [next] | [standalone]

#188

Tom Stiller <tom_stiller@yahoo.com> wrote:

> In article <1k6dj99.1pwqm1d1d892j9N%dempson@actrix.gen.nz>,
>  dempson@actrix.gen.nz (David Empson) wrote:
> 
> > Neill Massello <nmassello@yahoo.com> wrote:
> > 
> > > David Empson <dempson@actrix.gen.nz> wrote:
> > > 
> > > > In Snow Leopard, the only models that default to 64-bit kernel are
> > > > recent Mac Pro, Xserve and Mac Mini Server models. Lion defaults to
> > > > using the 64-bit kernel on a wider range of models, but it still has the
> > > > 32-bit kernel as an option for those models, the default for several
> > > > more models, and it is the only option for older models.
> > > 
> > > The following page lists the older Macs that can boot to the 64-bit
> > > kernel in Snow Leopard. 
> > > 
> > > <http://www.ahatfullofsky.comuv.com/English/Programs/SMS/SMS.html>
> > 
> > The list on that page is out of date, but it does cover old models. All
> > Xserve, Mac Pro, MacBook Pro and iMac models introduced in 2008 or later
> > are capable of running the 64-bit kernel.
> > 
> > This page covers Apple's official position for 10.6:
> > 
> > http://support.apple.com/kb/ht3770
> > 
> > Mac OS X 10.6 non-server defaults to 64-bit kernel for the mid 2010 Mac
> > Pro, early 2011 MacBook Pro and mid 2011 iMac.
> > 
> > Mac OS X Server 10.6 defaults to 64-bit kernel for the Xserve (early
> > 2008 and later), Mac Pro (early 2008 and later), and Mac Mini Server
> > (mid 2010).
> > 
> > The same Mac Pro and Mac Mini models can run 64-bit kernel with Mac OS X
> > 10.6 non-server but default to 32-bit. This also applies to the iMac and
> > MacBook Pro from early 2008.
> > 
> > No MacBook can use the 64-bit kernel, nor can the Mac Mini in the
> > original body design (late 2009 or earlier models).
> > 
> > I haven't found an official statement from Apple on which models default
> > to 64-bit kernel with Lion, but my impression is that it has been
> > extended to more models than with Snow Leopard.
> 
> I can verify that Lion runs a 64-bit kernel on my iMac7,1 where Snow 
> Leopard would not.

That is interesting. I used to have a mid 2007 MacBook Pro, which
predated that iMac by a couple of months. Despite having 64-bit EFI, it
was not able to run 64-bit kernel on Snow Leopard. I expect the mid 2007
iMac was in the same category, and Lion has extended 64-bit kernel
support to models with 64-bit EFI.

The question is whether that also includes MacBook and Mac Mini models
with 64-bit EFI.

My mid 2007 Mac Mini (oldest supported Mac Mini series) is not able to
run the 64-bit kernel in Lion.

-- 
David Empson
dempson@actrix.gen.nz

[toc] | [prev] | [next] | [standalone]

#183

On 08/21/2011 12:17 AM, David Empson wrote:
> philo<philo@invalid.not>  wrote:
>
>> On 08/20/2011 07:56 PM, David Empson wrote:
>>> philo<philo@invalid.not>   wrote:
>>>
>>>> On 08/19/2011 03:10 PM, Francois wrote:
>>>>> I had recently memory problems on my Mac Pro (early 2008), and I had to
>>>>> replace some 4GB DIMM's to recover my original 24 GB RAM. This Ram
>>>>> consists now of 4 slots with 2GB ( risers A&   B) and 4 with 4GB ( also
>>>>> half on A , half on B). The makers are Crucial (2x4 GB) and "elixir" for
>>>>> the rest. Of course I ran memtest ( in single user mode ) after each
>>>>> memory change ( there were several). Now ,every time, memtest log said :
>>>>> "Attempted memory lock... ERROR Memory lock failed - reason unknown
>>>>> WARNING : Testing with unlocked memory may be slower and less reliable"
>>>>>
>>>>> The 3 test sequence was every time successful, and the computer behaves
>>>>> normally.
>>>>> Anyway I'm baffled : why is my memory not "lockable" ?
>>>>> Has this something to see with the placement of the memory sticks on the
>>>>> risers ?
>>>>> Thanks for any hint, although that doesn't prevent me to sleep.
>>>>
>>>>
>>>>
>>>> More than likely you are booting to single mode using the 32bit kernel
>>>> which cannot address more than 4 gigs of ram.
>>>
>>> Not exactly. The 32-bit kernel on Mac OS X cannot directly address more
>>> than 4 GB at a time, but through the use of PAE it is able to manage up
>>> to 32 GB of RAM, and selectively access all of it.
>>>
>>> You can have 64-bit applications running on Mac OS X with 32-bit kernel,
>>> and in those conditions the 64-bit applications can directly address
>>> more than 4 GB of RAM.
>>>
>>> The "lock" mechanism used by memtest is very likely to be related to the
>>> 4 GB directly addressable limit of the 32-bit kernel but I'm not
>>> familiar enough with that level of detail to explain further.
>>
>> Not a function of being a mac...
>> it's the same for any 32 bit os  vs 64 bit
>
> Linux and Mac OS X use PAE to allow a 32-bit kernel to manage an address
> space larger than 4 GB on a 64-bit processor. Windows doesn't: it has
> separate 32-bit and 64-bit versions of the OS.
>

Though Windows has both 32bit and 64bit OS's
if the hardware supports it PAE can be enabled in many 32 bit versions 
of Windows

http://msdn.microsoft.com/en-us/windows/hardware/gg487503.aspx

[toc] | [prev] | [next] | [standalone]

#185

philo <philo@invalid.not> wrote:

> On 08/21/2011 12:17 AM, David Empson wrote:
> > philo<philo@invalid.not>  wrote:
> >
> >> On 08/20/2011 07:56 PM, David Empson wrote:
> >>> philo<philo@invalid.not>   wrote:
> >>>
> >>>> On 08/19/2011 03:10 PM, Francois wrote:
> >>>>> I had recently memory problems on my Mac Pro (early 2008), and I had to
> >>>>> replace some 4GB DIMM's to recover my original 24 GB RAM. This Ram
> >>>>> consists now of 4 slots with 2GB ( risers A&   B) and 4 with 4GB ( also
> >>>>> half on A , half on B). The makers are Crucial (2x4 GB) and "elixir" for
> >>>>> the rest. Of course I ran memtest ( in single user mode ) after each
> >>>>> memory change ( there were several). Now ,every time, memtest log said :
> >>>>> "Attempted memory lock... ERROR Memory lock failed - reason unknown
> >>>>> WARNING : Testing with unlocked memory may be slower and less reliable"
> >>>>>
> >>>>> The 3 test sequence was every time successful, and the computer behaves
> >>>>> normally.
> >>>>> Anyway I'm baffled : why is my memory not "lockable" ?
> >>>>> Has this something to see with the placement of the memory sticks on the
> >>>>> risers ?
> >>>>> Thanks for any hint, although that doesn't prevent me to sleep.
> >>>>
> >>>>
> >>>>
> >>>> More than likely you are booting to single mode using the 32bit kernel
> >>>> which cannot address more than 4 gigs of ram.
> >>>
> >>> Not exactly. The 32-bit kernel on Mac OS X cannot directly address more
> >>> than 4 GB at a time, but through the use of PAE it is able to manage up
> >>> to 32 GB of RAM, and selectively access all of it.
> >>>
> >>> You can have 64-bit applications running on Mac OS X with 32-bit kernel,
> >>> and in those conditions the 64-bit applications can directly address
> >>> more than 4 GB of RAM.
> >>>
> >>> The "lock" mechanism used by memtest is very likely to be related to the
> >>> 4 GB directly addressable limit of the 32-bit kernel but I'm not
> >>> familiar enough with that level of detail to explain further.
> >>
> >> Not a function of being a mac...
> >> it's the same for any 32 bit os  vs 64 bit
> >
> > Linux and Mac OS X use PAE to allow a 32-bit kernel to manage an address
> > space larger than 4 GB on a 64-bit processor. Windows doesn't: it has
> > separate 32-bit and 64-bit versions of the OS.
> >
> 
> Though Windows has both 32bit and 64bit OS's
> if the hardware supports it PAE can be enabled in many 32 bit versions
> of Windows
> 
> http://msdn.microsoft.com/en-us/windows/hardware/gg487503.aspx

Thanks for the correction. I was thinking of non-server editions of
32-bit Windows (up to and including Windows 7), which can't address more
than 4 GB address space.

-- 
David Empson
dempson@actrix.gen.nz

[toc] | [prev] | [next] | [standalone]

#187

On 08/21/2011 08:04 AM, David Empson wrote:
> philo<philo@invalid.not>  wrote:
>
>> On 08/21/2011 12:17 AM, David Empson wrote:
>>> philo<philo@invalid.not>   wrote:
>>>
>>>> On 08/20/2011 07:56 PM, David Empson wrote:
>>>>> philo<philo@invalid.not>    wrote:
>>>>>
<snip>

>>>>>>
>>>>>> More than likely you are booting to single mode using the 32bit kernel
>>>>>> which cannot address more than 4 gigs of ram.
>>>>>
>>>>> Not exactly. The 32-bit kernel on Mac OS X cannot directly address more
>>>>> than 4 GB at a time, but through the use of PAE it is able to manage up
>>>>> to 32 GB of RAM, and selectively access all of it.
>>>>>
>>>>> You can have 64-bit applications running on Mac OS X with 32-bit kernel,
>>>>> and in those conditions the 64-bit applications can directly address
>>>>> more than 4 GB of RAM.
>>>>>
>>>>> The "lock" mechanism used by memtest is very likely to be related to the
>>>>> 4 GB directly addressable limit of the 32-bit kernel but I'm not
>>>>> familiar enough with that level of detail to explain further.
>>>>
>>>> Not a function of being a mac...
>>>> it's the same for any 32 bit os  vs 64 bit
>>>
>>> Linux and Mac OS X use PAE to allow a 32-bit kernel to manage an address
>>> space larger than 4 GB on a 64-bit processor. Windows doesn't: it has
>>> separate 32-bit and 64-bit versions of the OS.
>>>
>>
>> Though Windows has both 32bit and 64bit OS's
>> if the hardware supports it PAE can be enabled in many 32 bit versions
>> of Windows
>>
>> http://msdn.microsoft.com/en-us/windows/hardware/gg487503.aspx
>
> Thanks for the correction. I was thinking of non-server editions of
> 32-bit Windows (up to and including Windows 7), which can't address more
> than 4 GB address space.
>

It looks like it might be a PITA to implement though...
seems Apple has a better handle on it

[toc] | [prev] | [next] | [standalone]

#174

Francois <xanrov@numericable.fr> wrote:
> I had recently memory problems on my Mac Pro (early 2008)

Just curious: what made you think you had memory problems to begin with?

-- 
Posted from my iPhone.

[toc] | [prev] | [next] | [standalone]

#181

On 20/08/2011 17:13, Jolly Roger wrote:
> Francois<xanrov@numericable.fr>  wrote:
>> I had recently memory problems on my Mac Pro (early 2008)
>
> Just curious: what made you think you had memory problems to begin with?
>
The computer could not start. Apple people detected it was due to a 
failed memory stick. When this DIMM was removed, with it's corresponding 
pal (as DIMMs in the Mac Pro go always by identical pairs), I found 
myself with 8 Gb of RAM gone. So I bought new ones and installed them. 
Hence The memtest. This happened again after a few month. Only the 
second time I had not to buy anything as it was again a couple of 4GB 
which was removed. I just reinstalled the good DIMM's which were of the 
same maker, with ,again, a memtest. As I said this was under OS X 10.6.5.

[toc] | [prev] | [standalone]

Back to top | Article view | comp.sys.mac.hardware.misc

csiph-web