Can 32-bit Puppy use >3 GB of RAM? (Yes, with PAE)

What features/apps/bugfixes needed in a future Puppy
Message
Author
p310don
Posts: 1492
Joined: Tue 19 May 2009, 23:11
Location: Brisbane, Australia

#31 Post by p310don »

I see this

Computer
Processor 2x AMD Athlon(tm) 64 X2 Dual Core Processor 5600+
Memory 3115MB (208MB used)
Operating System Puppy Linux 0.52
User Name root (root)
Date/Time Fri 18 Mar 2011 11:33:26 EST

I am not running as pfix=ram as you are (?) but I did try it as such, and get the same ram figure. The filesystem shows my save file, as opposed to my memory size.

gcmartin

Problem with RAM use on a 4GB system

#32 Post by gcmartin »

p310don wrote:I see this

Computer
Processor 2x AMD Athlon(tm) 64 X2 Dual Core Processor 5600+
Memory 3115MB (208MB used)
Operating System Puppy Linux 0.52
User Name root (root)
Date/Time Fri 18 Mar 2011 11:33:26 EST

I am not running as pfix=ram as you are (?) but I did try it as such, and get the same ram figure. The filesystem shows my save file, as opposed to my memory size.
You did NOT post your filesystem report for some reason.
Observation: You may want to download and run FATDOG on your platform to see results that you obtain. I get different results when using FATDOG....but that is 64bit versus 'your call for 32bit".

Check this: Do you have a SWAP partition or are you running a swap file that resides in your system's file system? We can't help you if you don't share enough information that we can use to pin-point your problem.
Flash wrote:Here's what I see, in a computer with 4 GB of RAM but no hard disk, running Puppy from a multisession DVD.
...
You peaked my interest, so I shutdown and booted WARY. Here's what I got so, maybe, there is a bug (but, certainly, this needs more investigation.).

Heres my report which should match what you have. I am running latest WARY LiveDVD on same system which has a SWAP partition on the local HDD. I am missing storage, same as you.

Code: Select all

-Computer-
Processor		: 2x AMD Athlon(tm) 64 X2 Dual Core Processor 4600+
Memory		: 3634MB (135MB used)
Operating System		: Puppy Linux 0.50
User Name		: root (root)
Date/Time		: Thu 17 Mar 2011 11:04:07 PM UTC
and

Code: Select all

-Mounted File Systems-
rootfs	/	0.28 % (3.7 GiB of 3.8 GiB)	
tmpfs	/initrd/pup_rw	0.28 % (3.7 GiB of 3.8 GiB)	
tmpfs	/initrd/mnt/tmpfs	99.09 % (900.0 KiB of 96.4 MiB)	
/dev/loop0	/initrd/pup_ro2	100.00 % (0.0 B of 95.5 MiB)	
unionfs	/	0.28 % (3.7 GiB of 3.8 GiB)	
shmfs	/dev/shm	0.00 % (859.1 MiB of 859.1 MiB)

Code: Select all

# df -l
Filesystem           1K-blocks      Used Available Use% Mounted on
tmpfs                  3933876     10972   3922904   1% /initrd/pup_rw
tmpfs                    98704     97804       900 100% /initrd/mnt/tmpfs
/dev/loop0               97792     97792         0 100% /initrd/pup_ro2
unionfs                3933876     10972   3922904   1% /
shmfs                   879764         0    879764   0% /dev/shm

Code: Select all

# fdisk -l
Disk /dev/sda: 500.1 GB, 500107862016 bytes
255 heads, 63 sectors/track, 60801 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Disk identifier: 0x000b935d

   Device Boot      Start         End      Blocks   Id  System
/dev/sda1           59918       60801     7100730    5  Extended
/dev/sda2               1        6374    51199123+  83  Linux
/dev/sda5           60275       60801     4233127+  82  Linux swap / Solaris

Partition table entries are not in disk order
I'm going to report this as a potential bug.

Hope this helps

gcmartin

Got the same results on other 32bit PUPs

#33 Post by gcmartin »

I got, virtually, the same results on following 32bit Pups:
  • Luci256
    MacPUP 520
    WARY5.1.1
    QuickSet-WARY 505q
    LightHouse 5.03G
Hope this helps
Last edited by gcmartin on Wed 23 Mar 2011, 05:55, edited 1 time in total.

scsijon
Posts: 1596
Joined: Thu 24 May 2007, 03:59
Location: the australian mallee
Contact:

#34 Post by scsijon »

1/ A part of the problem can be hardware chipsets, a lot of the 32bit chipsets used on motherboards have saved components by using some of the addressing matrixes for their internal device addressing.

2/ It hasn't been that long since the standard memory was stepped up from 256meg to 1gig and with 4 memory slots it was planned for 1 or 2gig (using alternate slots) just not 4gig. The possability of 4gig or even 16gig cards was in the address space for the memory socket, just not expected to be used by all motherboards.

User avatar
bigpup
Posts: 13886
Joined: Sun 11 Oct 2009, 18:15
Location: S.C. USA

#35 Post by bigpup »

James C wrote:All 32 bit Linux kernels (without PAE enabled) are limited to using about 3.2 to 3.4 Gb of ram ........minus reserved ram for onboard graphics, etc. It's a kernel limitation that applies to all distros...not just Puppy.
As an example, I just booted Ubuntu 10.10 in a box with 4 Gb ram.... it only showed 2.7 Gb ram .....the 3.2 minus 512 Mb for the graphics equals the 2.7 Gb that shows.
The solution is a PAE enabled kernel or run a 64 bit version.

See here
http://www.linuxquestions.org/questions ... ost3458590

gcmartin

PAE or 64bit seems to be the answer

#36 Post by gcmartin »

Thanks #BIGpup. 32bit PUPs do not support modern hardware like 64bit FATDOG does. That's why I am getting differing readings using differing PUPs vs FATDOG on the same system.

Request: If anyone is running a 32bit PUP on any PC that has 6GB/8GB, Please post your Hardinfo summary for a review. Thanks in advance.
Edited Today, at 4:15 pm EST - Don't bother to post on the "Request" I ask..
I just found out that this is an issue that BarryK or 32bit PUP system builders are going to have to address. It appears to be an oversight by them. Appears to have been an oversight for quite a while.

Thanks to @p310don for bringing this to the forum's attentions. Thanks again, BIGpup, for your expert help.

User avatar
Flash
Official Dog Handler
Posts: 13071
Joined: Wed 04 May 2005, 16:04
Location: Arizona USA

Re: PAE or 64bit seems to be the answer

#37 Post by Flash »

gcmartin wrote:Appears to have been an oversight for quite a while.

Thanks to @p310don for bringing this to the forum's attentions. Thanks again, BIGpup, for your expert help.
This is not the first time the question has come up in the forum. Perhaps a year or two ago there was a thread discussing why neither Linux nor Windows see more than about 3.3 GB of RAM.

If it's a bug in Linux, or at least in Puppy, then Windows seems to have it too.

User avatar
James C
Posts: 6618
Joined: Thu 26 Mar 2009, 05:12
Location: Kentucky

Re: PAE or 64bit seems to be the answer

#38 Post by James C »

Flash wrote:
gcmartin wrote:Appears to have been an oversight for quite a while.

Thanks to @p310don for bringing this to the forum's attentions. Thanks again, BIGpup, for your expert help.
This is not the first time the question has come up in the forum. Perhaps a year or two ago there was a thread discussing why neither Linux nor Windows see more than about 3.3 GB of RAM.

If it's a bug in Linux, or at least in Puppy, then Windows seems to have it too.
It's a 32 bit limitation, period. Windows and all 32 bit Linux distros without the PAE kernel will only show about 3.2 to 3.4 Gb of ram.I'm not going to get into all the physical memory addresses and virtual memory addresses stuff .......it's not a bug or oversight...it's just a reality.

p310don
Posts: 1492
Joined: Tue 19 May 2009, 23:11
Location: Brisbane, Australia

#39 Post by p310don »

To summarise, and try to bring things back on track - standart puppies DO NOT see more than 3.1ish gig of ram, because they are 32bit, NOT A BUG WITH PUPPY. This is a limitation as a result of being a 32bit operating system. Google it for further explanation.

The point of this thread isn't to imply a bug with puppy, but I was hoping that a puppy dev genius, might be able to address the issue with an ingeniously creative solution.

User avatar
Flash
Official Dog Handler
Posts: 13071
Joined: Wed 04 May 2005, 16:04
Location: Arizona USA

#40 Post by Flash »

All I can say is, 2^32 = nearly 4.3 GB.

amigo
Posts: 2629
Joined: Mon 02 Apr 2007, 06:52

#41 Post by amigo »

I'll say again that the kernel does see all 4GB. But, it reserves ~1GB for itself leaving only ~GB available to userspace. It is a bit of a waste because the kernel will never need all that for itself, but the only way to change those proportions is to patch the kernel. With PAE enabled, the kernel can see above 4GB and also allows a way to change the proportion dynamically (IIRC).

jamesbond
Posts: 3433
Joined: Mon 26 Feb 2007, 05:02
Location: The Blue Marble

#42 Post by jamesbond »

Time for a test then :)

I just checked, Barry's kernels don't have HIGHMEM64GB enabled (I think this is what is needed to enable PAE). All that is needed to do is to get Barry's kernel package, enable this config item, and re-build the kernel, then supplant the new kernel into existing puppy (e.g. latest Wary for example). If we're lucky then we don't even need to rebuild the modules, it would be just straightforward swap of vmlinuz with the new bzImage.

Then those having more than 4GB of RAM can do a test, and see whether the ramdisk thing is really working ...

Anyone?
Fatdog64 forum links: [url=http://murga-linux.com/puppy/viewtopic.php?t=117546]Latest version[/url] | [url=https://cutt.ly/ke8sn5H]Contributed packages[/url] | [url=https://cutt.ly/se8scrb]ISO builder[/url]

p310don
Posts: 1492
Joined: Tue 19 May 2009, 23:11
Location: Brisbane, Australia

#43 Post by p310don »

Jamesbond...
Time for a test then Smile
Agreed. Now, I don't have the skill to patch kernels, but if someone wants to do it, I will happily test it.

User avatar
ecube
Posts: 88
Joined: Fri 11 Jul 2008, 17:00
Location: Västerås, Sweden

#44 Post by ecube »

I don't know if it has any bearing on the topic discussed, but I was successful in adapting Fluppy 06 to 4 GB RAM. :D
http://www.murga-linux.com/puppy/viewto ... 557#460557

p310don
Posts: 1492
Joined: Tue 19 May 2009, 23:11
Location: Brisbane, Australia

#45 Post by p310don »

Ecube - how is that working out for you? Are there any stability issues? Did it increase the size of fluppy significantly? Have you noticed any pros and/or cons?

jamesbond
Posts: 3433
Joined: Mon 26 Feb 2007, 05:02
Location: The Blue Marble

#46 Post by jamesbond »

Hehehe ... I compiled Wary kernel with HIGHMEM64G option also, and yes available memory went from 2.9GB to 3.9GB (this is on a 4GB machine).

Bad news though, enabling this option is enough to change the module layout signature, so existing modules won't load. You will also need to re-compile all the modules also (and then install that modules to pup.sfs and initrd.gz). I didn't do that, so I only got as far as the ramdisk shell :D
Fatdog64 forum links: [url=http://murga-linux.com/puppy/viewtopic.php?t=117546]Latest version[/url] | [url=https://cutt.ly/ke8sn5H]Contributed packages[/url] | [url=https://cutt.ly/se8scrb]ISO builder[/url]

User avatar
Karl Godt
Posts: 4199
Joined: Sun 20 Jun 2010, 13:52
Location: Kiel,Germany

#47 Post by Karl Godt »

Hehehe ... I compiled Wary kernel with HIGHMEM64G option also, and yes available memory went from 2.9GB to 3.9GB (this is on a 4GB machine).
You always have to compile all modules I think everytime you even make small changes to the kernel .config

May I ask about the architecture you did choose ?

Common seems to be the CONFIG_M586
or CONFIG_M686

but 4GB+ Machines might be worth _MCORE2 or _MATOM

I have no machine above 1GB and my testbox is a P4 with 448MB RAM and until now I can say that it is no performance difference in _M386, _M486, _M586, _M586TSC, _M686
and _PREEMPT or _PREEMPT_VOLUNTARY

User avatar
technosaurus
Posts: 4853
Joined: Mon 19 May 2008, 01:24
Location: Blue Springs, MO
Contact:

#48 Post by technosaurus »

the biggest difference in the architectures is the stack size. medium sized apps can fit in later architectures whereas only really small apps can fit on older ones. it makes a big difference if for instance you can tweak the kernel and compiler enough to fit your high load web server.
Check out my [url=https://github.com/technosaurus]github repositories[/url]. I may eventually get around to updating my [url=http://bashismal.blogspot.com]blogspot[/url].

gcmartin

Barry produces the base 32bit kernel(s)

#49 Post by gcmartin »

For mainstream 32bit Pups, Barry produces the kernels. Have we taken the time to present this issue to Barry? Anyone? He may have consciously done this for a reason he feels passionate about. We won't understand unless he responds.

Hope this helps

p310don
Posts: 1492
Joined: Tue 19 May 2009, 23:11
Location: Brisbane, Australia

#50 Post by p310don »

I did pm Barry to reference him to this thread, but as yet have not heard back.

Gcmartin is absolutely right, this is something that Barry, or whoever is creating the puppy needs to do at the beginning, rather than as a patch, as shown by Jamesbond.

Perhaps this is something that should be considered when building puppy 6? Something for the future?

Post Reply