64bit PCs and PAE in Puppy World - Facts versus Myths

What works, and doesn't, for you. Be specific, and please include Puppy version.
Message
Author
gcmartin

64bit PCs and PAE in Puppy World - Facts versus Myths

#1 Post by gcmartin »

Whether anyone of us has
  • AMD 64bit
  • Intel 64bit
  • Single Core
  • Multicore
  • Puppy Linus 32bit distro(s)
  • FATDOG or LightHOUSE 64bit distro(s)
You may/may-not have had a passing conversation about its ability.

This thread
  • is an appeal for useful information about 64bit platforms and the Puppy Linux that runs on this hardware-motherboard combinations.
  • intends to dispel much emotion from the reality of the hardware and what the OS does in using it.
  • educate Puppy Forum Discussion members thru collaboration and collection of facts versus "much emotion" that can cause distress when it comes from the uninformed.
1994
Firstly, lets go all the way back to Pre-PentiumPro when the Intels, AMDs, IBMs, HPs, Fujitsu, MSs, etc look out over the horizon and planned for the advent of Multimedia processing needs as well as the scientific community's data management needs with the burgeoning communities of humans that will be moving mankind forward.

They planned an interim and a longer range solution to address those needs. It became known as PAE in shorten time span and 64bit for the further-out time span.

PAE, either AMDs or Intels, basically allows the OS to use a new hardware reference to memory pages existing in RAM. It is a replacement for the "normal" hardware memory reference of RAM pages (called frames). That's it, folks. Nothing fancy, no mirrors, no sorcery, just giving the OS page frame reference to getting "stuff" stored in the RAM frames to the CPU(s).

There has been many cries to needing or using RAM during our lifetime by people as they expound about whether there is need or etc. This is emotional, to me, as the FACT is that the hardware vendors have paved the way, no matter whether you want to use it or not. And, further, since its in hardware, they did it so such that the time to reference a RAM frame is the same...whether you use it or not!

That's the fact. So for about 16 years, the ability to reference RAM on a single platform in 32bit mode has been there whether the OS used the old way or the 16 year old newer way!

2005
Lets fast forward 10 years later when AMD, first, Intel, followed, began generally providing business and customer 64bit PCs. In their implementations, each of them did the exact same thing that they did 10 years earlier. Gave you and I a way to use EITHER the 32bit way or the 64bit way. Very few people stopped long enough to compare the similarity between how this was delivered to us. Shame, because, had we done so, we might have learned the obvious.

What they provided here is a fact of life (ur, manufacturing). And they do not care which of these you choose to purchase or use.

So, even today, there are many who continue to ignore the obvious nature of the landscape "as they expound about whether there is need or etc. This is emotional, to me, as the FACT is that the hardware vendors have paved the way, no matter whether you want to use it or not."

I have been fortunate enough to have done comparative testing with customers on several vendor systems over the years in their benchmark centers. What's significant about the PC arena is that there has never been shown to have encountered a negative impact on the use of the 32bit PAE OSs when comparing it against non-PAE.

Further, there has NOT been a single report that has shown a general degradation in overall performance of a 32bit versus 64 bit OS on the same 64bit platform.

So we are here in the 21st Century battling the same demons from the last century. ....Emotions!

My question for all of us, is whether we feel its right for us to make attempts to persuade or dissuade a distro developer to not use new technology if there is little to no impact on either behavior or performance on his/her's Puppy deliverable.

There was a report earlier last month on the forum about a remark that Linus Tovalds made. When reading and taking his remark into context, he was making a matter of fact statement that was intending to discourage those who have been trying to get him to continue 32bit life of Linux....when to him, the world is going 64bit and greater. I AGREE WITH HIM! So asking that he continue to do more with32bit via PAE with the limited resources and the current direction, I have to take his side and say something along the same lines. .."I have already addressed 32bit. I do intend to use my precious resources on the location of where the world is and the direction where its going."

32bit Puppy does a significant contribution to all. It provides the best overall general purpose desktop solution that anyone can have on a 32bit platform. It satifies 32bit PCs from 64MB RAM (older versons) to 64GB RAM (Barry's Drake and Barry's RACY). It uses older 32bit RAM frames references and PAE RAM frame references as it intends to satify all 32bit PC needs no matter how little or how much RAM you have. You can find a PUP in this community to address your platforms peripherals, CPU, and RAM!

64bit versions; namely FATDOG and LightHOUSE64, are addressing those PCs built since 2006, which have 64bit CPUs on their motherboards. This class of PCs comes with DVD burners, and RAM sizes beginning at 512MB RAM and can grow to a size that I cannot conceive. These platforms are capable of providing file system performance that is unparalleled.

ISO size
Many people in the Puppy community are unaware of the ISO size issue and why it exist. The ISO size has being in the amount of data that is transferred from a website to your PC so that you can burn a CD or a DVD to start your Puppy. In many parts of the world this is VERY very important because their connection with the internet is extremely slow. Working off a dial-up modem has many challenges to uploading and downloading information from the internet.

Next, the CD/DVD when booting will need to "expand" the OS from that media into the system's RAM + Swap. But, in every case, thus far, I have found that once the system is booted on desktop, the running OS has never been more than the ISO's size.

This is the fact. Everyone of us can go look for themselves at the summary screen provided by the Hardinfo Tool that is provided in almost every Puppy built.

Puppy does one GREAT-GREAT thing with the remaining RAM that exist in your PC. He uses it for the "FILE SYSTEM" of the running Puppy. Thus Puppy gives you about the fastest "disk" drive that you could only die for. I am amazed at the many persons I've found (including a couple distro developers) who are not aware of this. Further, they sometimes are unaware of the beauty in how the OS works with SWAP for a truly wonderful user experience.

ISO size has little to do with your system's performance. Your system's performance is not (and never has been determined by the size of ANY program). The SIZE of a program is NOT the determinant. The determinant is the path length necessary to carry out a useful operation! Many also are not aware of this as they merely point at the size of something as if it is bad. Its NOT if it gives every user a satisfactory level of useful experience This is the same as your deciding if you want the comfort of a Bentley or you want the teeth rattling ride-speed of an Ferrari. The user experience is different for each and its up to the distro designer to select the products that meets their OTB reference point(s). They do all of us a favor by getting us to a starting point so that we can enjoy and expand the delivered desktop experience (assuming RAM availability to do so)

Another item on note is that for any system to reference a frame at the lower end of RAM, in the middle of all of your RAM, or the very last frame of your RAM, the access time is EXACTLY THE SAME! This is a fact that I think most understand, but, its important to note.

A Summary
  • Every 64bit system has basically 2 benefits over most 32bit PCs that were purchased; namely bus (64bit) and RAM.
  • Though, very rare, some 64bit and some 32bit PCs were built without PAE
  • Except for Itanium (Intel) all 64bit PCs can run 32bit OSs AND 64bit OSs.
If there are ANY mistakes in what is shared here, please, please share it with us so that this is a "true" collaboration of information that is useful to everyone in Puppyland.

This thread intends to be a collaboration. All information you contribute should intend to help us all. And, if referencing other website's info, please summarize with your reference.

When we understand the history and the facts, we are in a much better position to provide assistance to the Distro Developers who work very hard to provide their services to us. They need us to understand as much as we can about what they are trying to do for us.

Hope this helps.

P.S. This thread intends, neither, to discourage or encourage how the community exercises their PCs. This thread merely attempts to separate some of the emotion from the facts of the distros we see in Puppyland. It should give anyone a realistic understanding on history and simple usage of our PCs as we select what to pay attention to when selecting a PUP.

I hope these intent parameters, mentioned, makes clear this thread's OTB objective.
Last edited by gcmartin on Wed 21 Sep 2011, 18:55, edited 5 times in total.

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

#2 Post by Flash »

I would say that there are several versions of 64-bit Puppy that present an almost unique opportunity to compare 32-bit to 64-bit applications and operating systems. As far as I know, every Puppy, 64- or 32-bit, can boot from CD and run entirely in RAM, eliminating the hard disk drive from affecting the results.

If anyone cares to do it, 32-bit applications such as Gimp or OpenOffice, running in a 32-bit Puppy could be compared to their 64-bit counterparts running in a 64-bit Puppy for speed of completing typical tasks while running entirely in RAM. Then we'd know for sure if 64 bits is worth all the hoopla.

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

#3 Post by p310don »

Further, there has NOT been a single report that has shown a general degradation in overall performance of a 32bit versus 64 bit OS on the same 64bit platform.
From reports I have read from Phoronix, and quoted elsewhere on this forum, it seems that applications that are made to specifically take advantage of the benefits of 64bit architecture will yield significant performance improvements over their 32bit counterparts. On the other hand, 32bit applications compiled for use in a 64bit environment don't show significant differences in performance.

As flash suggested, who wants to do the tests for our own results??

User avatar
Aitch
Posts: 6518
Joined: Wed 04 Apr 2007, 15:57
Location: Chatham, Kent, UK

#4 Post by Aitch »

Great post, gcmartin

Unfortunately I haven't acquired/been given a 64bit PC yet, but do use/take advantage of a couple of PAE puppies, wary511pae, and slacko530pae, on an old server box that's crying out for another 9gb of ecc ram, to add to its current 3gb, ....when I can afford it

Discussion seems to have continued here..

http://murga-linux.com/puppy/viewtopic.php?t=72776

Aitch :)

gcmartin

PAE comments made elsewhere on PAE and Linux.

#5 Post by gcmartin »

In another thread, I wrote:For many in this forum the topic of PAE is an emotional one.

I hope its not one with you. I am NOT trying to address any emotion,I am not trying to increase discourse, nor am I trying to justify something. This post is an FYI for any newbie or experienced user for their understanding.

PAE has been around since 1995. It is built-in on almost every 32bit platform built since then. There were only a very few Netbooks, initially, that did not have PAE; AND, there were an even fewer number of Laptop models, ever built, without PAE. Later netbooks (as well as the quality early models) Netbooks did have PAE (in fact, i have an ASUS eeePC 900A which came with PAE).

In my experiences with so many PC running MS/Apple/Linux, I personally, have not seen or owned any PC without PAE. I know that some members in this community have some systems which do NOT have PAE. There are many many PUPs which are designed without PAE. But I acknowledge and know that some of the Puppy forum membership get upset when they see that a PAE kernel exist. In fact, some have tried to infer a comment made by Tovalds when he indicated, years ago, that Linux has done all that is necessary for 32bit development and nothing more is necessary or would be done in the memory model of which PAE addresses. In fact, as he had stated many times before that specific comment, PAE "works" ... why change? I agree with his point of view. Linux is designed to use both the Pentium initially built RAM model or the 1995 PAE add-on RAM model's capability in its system operations. Lastly,its important to note what was happening: There were pundits that wanted Linux extended to automatically detect the 32bit PAE hardware component and to automatically use it when found. Personally, I see nothing wrong with the approach he used and to not devote resources to do such in the rise of 64bit processing. And, 64bit PCs don't need this in 64bit operations.

PAE testing among Puppy Members are responsible, today, for the ability for PAE enabled PUPs to run on everything from 256MB to 64GB (impossible to configure 64GB in any desktop) without either negative issues or negative impacts. The member community did not write anything which suggested that PAE did not work. The development members showed "How to" generate kernels and distros for PUPs which takes advantage of the PAE's hardware component.

Yet, this PAE topic continues to this day with members tossing it around with negative overtones and info contrary to facts.

I would like to share, once again, that PAE has been shown to run positively when available and does not produce negative system impacts. FURTHER, in almost every distro where PAE has been made available, no one with a PAE capable PC has written in to share that PAE somehow broke or crashed the PC. It doesn't break things, it provides the ability to run any RAM configuration one chooses, it doesn't add path-length to system operations, it just doesn't ... In fact, no one can tell by looking at your PC whether the distro (operating system) is using the PAE hardware of not. Its not something that can be touched as it carries out all system's RAM request.

Again over 99% of all PCs, laptops, etc that have 32bit capability, has PAE built-in. Only a very very very few did NOT. If your PC has PAE, everyone finds that "It just works!" And, this distro works without issue.

This current Precise version by Barry WORKS on the only Netbook I have....very well. And, Precise also works, as well. on the 8 other PC configurations I've tested to provide feedback on its behavior.

I have found this same impartial operations on Slacko and all of the other 32bit PAE PUP distros I have tested and run over the years....no matter how much RAM is on the motherboard: "It just WORKS!" The most important piece of the understanding of PAE is that it works no matter how LITTLE or how much RAM is installed. Linux sees it all, productively.

I find no bugs in its operations

Here to help

gcmartin

#6 Post by gcmartin »

The issue again, here, is that Linux does and continues to operate very well on PCs.

The Linux community, at large, has published performance measurement of non-PAE versus PAE counterparts. This info is PUBLIC information and charts and graphs accompany in those reports we often see.

Tovalds never said the Linux implementation doesn’t address RAM or degrades performance. In fact, it has been shown to perform equivalently.

In our own forum, members have published their finding and reports when PAE was initially introduced in this community. There has been No information anywhere by anyone, including Tovalds, that suggest Linux’s PAE kernel causes negative system operations or impacts.

My reason for writing this thread is to try to point out the the facts found by industry leaders dispelling any notion that PAE will cause you problems, no matter how MUCH RAM you have.

If you have a 32bit PC that is built with the intel/AMD CPU, you will most likely have PAE built-in. You can use a PAE distros on your system and it will perform equally as well as its non-PAE version, no matter the RAM amount you have on your PC.

This is not an opinion. It has been tested and retested and comparisons done for the past half-decade. Linux works here and it works extremely well.

No emotion, no opinion. Best advice has already been given in this thread. Use either 32bit PAE, 32bit without PAE or use 64bit for any reasons one might have.

This thread is intended to provide information on the CPUs and distros built to run on your PCs.

Here to help

User avatar
Terryphi
Posts: 761
Joined: Wed 02 Jul 2008, 09:32
Location: West Wales, Britain.

Re: PAE comments made elsewhere on PAE and Linux.

#7 Post by Terryphi »

gcmartin wrote:
This current Precise version by Barry WORKS on the only Netbook I have....very well. And, Precise also works, as well. on the 8 other PC configurations I've tested to provide feedback on its behavior.

I have found this same impartial operations on Slacko and all of the other 32bit PAE PUP distros I have tested and run over the years....no matter how much RAM is on the motherboard[/u]
I have no interest in the debate about PAE which I consider is of no practical relevance to the average user. However, to set the record straight it needs to be pointed out that the latest Precise versions, currently 5.6.11beta and 5.6.1, are NON-PAE so their excellent performance on a range of hardware cannot be used in support of PAE. Apologies if I have misunderstood your reference to Precise.
[b]Classic Opera 12.16 browser SFS package[/b] for Precise, Slacko, Racy, Wary, Lucid, etc available[url=http://terryphillips.org.uk/operasfs.htm]here[/url] :)

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

#8 Post by James C »

Since the content of the post I answered was moved to this thread for some reason might as well move my replies as well.
http://www.murga-linux.com/puppy/viewto ... 728#715728

gcmartin wrote:
This current Precise version by Barry WORKS on the only Netbook I have....very well. And, Precise also works, as well. on the 8 other PC configurations I've tested to provide feedback on its behavior.
After your little discourse on the virtues of pae I almost hate to mention this but....... you do realize that the latest version of Precise, as well as other recent versions, features a non-pae kernel.

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

#9 Post by James C »

http://www.murga-linux.com/puppy/viewto ... 736#715736

gcmartin wrote:
In fact, some have tried to infer a comment made by Tovalds when he indicated, years ago, that Linux has done all that is necessary for 32bit development and nothing more is necessary or would be done in the memory model of which PAE addresses.

I guess you're talking about me. Since you brought it up......

Linus Torvalds wrote:

http://www.realworldtech.com/forum/?thr ... stid=76973

PAE really really sucks.

The biggest single reason to go 64-bit is exactly because of physical address space. Your virtual address space
needs to bea multiple of the physical one: when you hit 1GB of RAM, 32-bit virtual memory is no longer acceptable. You literally do need more virtual memory than physical.

PAE turned that very simple fact on its head, and screwed things up royally. Whoever came up with the idea was totally incompetent, and had forgotten all the DOS HIGHMEM pains. There’s a damn good reason why we left the 286 behind, and started using 386′s, instead of having HIGHMEM crap with windows into a bigger physical space.

Repeat after me:

Virtual space needs to be bigger than physical space. Not “as big

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

#10 Post by James C »

And lastly.....

http://www.murga-linux.com/puppy/viewto ... 746#715746


Billtoo wrote:
James C wrote:


Been trying for almost 2 years to have you back up this assertion....... I provided a link and a quote to support my position.

gcmartin wrote:

Yet, this PAE topic continues to this day with members tossing it around with negative overtones and info contrary to facts.


I for one am open to consider any contrary documented facts you can provide..... not mere opinions.

Linus Torvalds hasn't changed his opinion of pae in this recent article (which is all geek to me of course Smile)

http://news.softpedia.com/news/Linux-Ke ... 5300.shtml
Thanks for the link. Smile


Sounds like Linus still really likes PAE.....

....the disgusting abortion that is x86 PAE, and a small part of me feels that anybody who hit this problem on such a machine probably got whatever they deserved,

User avatar
rcrsn51
Posts: 13096
Joined: Tue 05 Sep 2006, 13:50
Location: Stratford, Ontario

#11 Post by rcrsn51 »

Here is the question that I have never seen addressed:

Under what specific conditions will a 32-bit Puppy actually require more than 4 GB of memory?

I'm not interested in hypotheticals. I would like to hear about actual experiences where the 4 GB limit kept you from doing something useful. Was there a reason that you could not switch to a 64-bit Puppy?

User avatar
pemasu
Posts: 5474
Joined: Wed 08 Jul 2009, 12:26
Location: Finland

#12 Post by pemasu »

Playing in linux the Steam game. I found one post where 4 Gb ram resulted to swap usage. Emphasis is that the game would have benefited of more ram.

http://www.dedoimedo.com/games/steam-li ... -here.html

And no, I dont think that gaming is something useful. So....it didnt stop doing something useful. My personal opinion, of course.

Linux server usage ? Though Puppy is rarely used as public server under heavy work. Just my hunch.

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

#13 Post by James C »

Another glowing recommendation for PAE.....

http://blog.superuser.com/2011/04/04/64 ... e-and-you/
So far I’ve been ignoring the fact that modern 32-bit processors have features such as PAE, mainly as for a home user it is akin to getting a live grenade, pulling the pin and just holding it in your hand. Sooner or later you’re going to be wondering why your arm hurts and why you can’t pick up your coffee mug any more.

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

#14 Post by jamesbond »

I opened the debate on PAE by creating the first PAE-enabled kernel years ago. So here's something I wanted to say, in no particular order.

1. PAE has performance impact. It may be non-noticeable for newer CPUs, but the further one looks back in time, the bigger the impact is (ie older CPU will be hit the worst).

2. PAE works exactly like Expanded Memory (EMS) of days gone by --- by bank switching. Remember bank-switched VESA video mode? Remember Commodore 64 bank-switched RAM? PAE works *exactly* like that. The annoying part of EMS memory is that programs *must be* specifically coded to use it, otherwise your shiny new 4MB EMS memory will not help that starved Wordstar program. The case is eerily similar with PAE, read on.

3. PAE allows the system (as a whole) to use more than 4GB of RAM, but individual applications are *still limited* to 3GB RAM unless they use tricks like DOS programs that used EMS memory. (One example of such "smart" program --> here). I can definitely say that the majority of open source programs *does not* come with such hacks simple because it is simpler, cleaner and easier to re-compile the apps as 64-bit application rather than making such hacks (read the previous link of the "smart" program that does use such hacks, and read the details of how the system needs to be configured specially for that hack to work ...).

4. Example of case of 3 above: PAE enables two virtual machines, each consuming 3GB memory, to run on 8GB system. PAE will *not* help a single virtual machine that needs to use 6GB to run, even if you have 32GB RAM on your system.

5. PAE's other benefit: it enables the use of NX bit which helps a little it preventing malware. No PAE == No NX bit, and this was actually the *primary* reason why mainline distros ship with PAE kernel.

6. @rsrcn51, use case of needing more than 3GB of RAM:
- Editing very high resolution graphic files.
- Editing very high resolution video files.
- Running virtual machines
- Running large simulations with very large matrices
- etc etc I'm sure there are other cases.
In either case, PAE may or may not help, as the example I give above.

As for why people can't switch to 64-bit OS - well, my guess is that there isn't any Upup-like or Slacko-like or Precise-like version of 64-bit Puppy. Lighthouse64 is working hard at it, I'm hopeful tazoc's creation may appeal to wider audience when it goes out of beta.

7. Lastly - PAE isn't a run-time choice. A kernel is either PAE or not. PAE cannot be turned off by "pfix=no-pae" or "pae=off". One can't run a Puppy CD with PAE kernel on it to issue a "warning: your CPU isn't supported" on first boot on a non-PAE CPU. It will simply display a blank screen or at best some cryptic one liner message.
A Puppy CD with PAE kernel is as good as a coaster for one whose computer isn't PAE capable. Whichever is chosen as the default kernel doesn't matter, but the release notes should contain warning that a Puppy with PAE kernel *may* potentially not work with older CPUs.

cheers!

anikin
Posts: 994
Joined: Thu 10 May 2012, 06:16

#15 Post by anikin »

Thank you, jamesbond for your explanation.
On the one hand, your presentation has more technical details and nuances. On the other hand, it lacks in passion and eloquence, compared to the original post. Besides, in many ways and aspects, these two views do not coincide and directly contradict each other. Before finally making up my mind on PAE, I would like to hear a response from gcmartin.

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

#16 Post by p310don »

Couldn't resist chiming in....

The best post on this thread is jamesbond's. Facts, not opinions.

I have been using PAE Saluki daily for 18months or more now. No problems at all. I have 8gig of ram in my machine. PAE comes in handy when running windows in Virtualbox. Before I used PAE, if I ran Windows with 2 gig allocated to it, and then tried to use Firefox to browse the web, the system crashed. In that situation, 3.2gig just wasn't enough. I have had no problems since switching to PAE and utilising 8 gig. Note, I can still only allocate 3.6gig to windows, which is enough.

The latest Precise 5.7 has PAE.

PAE is in my opinion, an excellent stop gap measure for Puppies designed for MODERN machines. In the current puppy landscape, there is almost no 64bit option. No "official" one anyway. That means almost all the programs are compiled / developed for 32bit. On most modern machines less than 5 years old, PAE works, but they are also 64bit capable.

This PAE debate is tired and old. 32 bit works. 32 bit with PAE works. 64 bit is the future.

User avatar
Q5sys
Posts: 1105
Joined: Thu 11 Dec 2008, 19:49
Contact:

#17 Post by Q5sys »

PAE has been and always will be a hack at accessing more memory.

There are people out there that continue to push it, but those people simply are repeating themselves in the face of fact and intelligent discourse from those in the tech industry.

I think Linus' post says everything it needs to. But some people in the world continue to want to preach the PAE message. This has come up several times in this forum. One of the last times I decided to actually do some testing and find out what kind of difference there is.
http://murga-linux.com/puppy/viewtopic.php?t=84472

I'd venture that pretty much all machines that can take physically handle more than 4gb ram, are 64bit capable. And should run a 64bit OS to take advantages of the x86_64 instruction set. Running a PAE kernel on a newer processor effectively limits your CPU to running like a multicore Pentium 3, at the additional cost of losing out on hardware AES, AVX, etc.

anikin wrote:Thank you, jamesbond for your explanation.
On the one hand, your presentation has more technical details and nuances. On the other hand, it lacks in passion and eloquence, compared to the original post. Besides, in many ways and aspects, these two views do not coincide and directly contradict each other. Before finally making up my mind on PAE, I would like to hear a response from gcmartin.
Lets say you're going in for Brain Surgery. Would you rather have a surgeon who is blunt with the facts, explains things in a highly technical manner, and is very specific in his details. Or a Surgeon who wants to use eloquent speech and tell you how great your life will be after the surgery?
I cant speak for you or anyone else, but for me; when its a technical issue. I'll take a technical explanation over flowery eloquent speaking.
In this specific case, I'd opt for the person (jmesbond) who's actually show his expertise through his lengthy development work in this community, over a person to likes to talk alot. (gcmartin)

gcmartin

#18 Post by gcmartin »

To my knowledge, I know of no one who is pushing PAE. PAE is a hardware addressing scheme for the CPU to get at RAM. It is designed to make RAM memory available to the OS by taking advantage of a hardware standard.

Again, NO ONE, that I am aware of, is sharing or suggesting that PAE distros are to be thought of to be used in place of a 64bit OS.

ALL PAE does is make RAM available to the processor.

PAE has absolutely nothing at all to do with anyone's emotional feeling. I sorry you express such.

Again, either your processor has it or it doesn't. That it. If it does, as more than 99% of all 32bit and most every 64bit x86 processors do, you can use it NO MATTER HOW LITTLE OR HOW MUCH RAM you have.

Its hardware. Its 18-19 years old. It does what its suppose to. It does not inhibit any OS operation. It was tested extensively by Puppy members. Many found and reported an increase in snappiness (but, that was by observations as we have no tools/utilities to measure "snappiness").

Puppy has been tested by Puppy members and has been provided by Puppy developers so that any user can run his OS no matter its RAM content from little all the way to the impossible 64GB in a 32bit platform.

This discussion continues to be mostly an emotional argument over the actual findings of Puppy members. In Puppyland, its been shown to work and to bring benefit in use.

PAE does NOT harm your PCs, does not degrade snappiness, and it does NOT cause operating system software failures in your distro. Your PC's hardware and software will work to your benefit.

That being said, its up to any Puppy user to select the OS he feels comfortable with. If you have PAE, a PAE distro will run. If you do NOT, a non-PAE distro will run. If you have a 64bit PC, most likely any 32bit/64bit OS will run on your PC. Puppy developers make their offerings available for all of our benefit.

I hope Puppy membership understand that this is merely a hardware feature where Puppy developers are making it available as, many PCs are now becoming apparent ,that have much more RAM now than was available when Barry began developing Puppy for world's users. Linux will use the capabilities of your processor's motherboard contents for your processing advantage. PAE functionality in the Linux kernel is NOT a hack when one considers that it is a selectable part of the kernel at kernel generation time.

Here to help

User avatar
Q5sys
Posts: 1105
Joined: Thu 11 Dec 2008, 19:49
Contact:

#19 Post by Q5sys »

gcmartin wrote:To my knowledge, I know of no one who is pushing PAE.
Says the man who has created 6 separate and unique threads about PAE:

1 - http://murga-linux.com/puppy/viewtopic.php?t=77611
2 - http://murga-linux.com/puppy/viewtopic.php?t=77068
3 - http://murga-linux.com/puppy/viewtopic.php?t=72776
4 - http://murga-linux.com/puppy/viewtopic.php?t=67707
5 - http://murga-linux.com/puppy/viewtopic.php?t=84816
6 - http://murga-linux.com/puppy/viewtopic.php?t=71861

This does not count the threads you've had deleted, I know of two off the top of my head, which you requested to be deleted. (I expect you to request this thread be deleted as well.)

gcmartin wrote:PAE has absolutely nothing at all to do with anyone's emotional feeling. I sorry you express such.
No one else in this community has spent as much time and effort as you talking about PAE.

gcmartin wrote:Its hardware. Its 18-19 years old. It does what its suppose to. It does not inhibit any OS operation. It was tested extensively by Puppy members. Many found and reported an increase in snappiness (but, that was by observations as we have no tools/utilities to measure "snappiness").
Citation needed.

Extensive testing has been done by this community which you have refused to accept. http://murga-linux.com/puppy/viewtopic.php?t=84472
Independent testing has been done by professionals in the Computer industry which you have refused to accept. http://murga-linux.com/puppy/viewtopic. ... 446#686079
Experts like Linus Torvalds have discussed at length why PAE is bad and shouldn't be used, which you have refused to accept. http://www.realworldtech.com/forum/?thr ... stid=76973
The Developer in this forum who first presented us with PAE has even talked about its performance degredation which you have refused to accept. http://murga-linux.com/puppy/viewtopic. ... 844#715844

That the technology is 18 years old is pointless. No Pentium1 or Pentium2 Motherboard sold commercially was able to handle more than 4gb of ram. So that the technology was included on the CPU die is meaningless since there was NO way it could be used or even tested outside of Intel's Labs.

It is a fundamental principle that PAE is slower than a non PAE simply by the nature of how it works. It requires two read operations instead of just one.
Never, under ANY circumstances can two reads be accomplished faster than one. This is so fundamental that it should NOT ever need to be explained.
It is an emprical fact that two read opperations take longer than one... because you have to finish the first read operation before you can start the second.
This simple fundamental fact about the technology apparently is not understood by you.
gcmartin wrote:Puppy has been tested by Puppy members and has been provided by Puppy developers so that any user can run his OS no matter its RAM content from little all the way to the impossible 64GB in a 32bit platform.
Citation needed.

Please show a test where a PAE based system has been proven to work with 64gb of ram on Puppy.

gcmartin wrote:This discussion continues to be mostly an emotional argument over the actual findings of Puppy members. In Puppyland, its been shown to work and to bring benefit in use.
Others are presenting facts which can be double checked and verified. You are presenting nothing more than anecdotal evidence. If anyone is emotional about this its you, perhaps thats why you keep creating threads about it over and over.

gcmartin wrote:PAE does NOT harm your PCs, does not degrade snappiness, and it does NOT cause operating system software failures in your distro. Your PC's hardware and software will work to your benefit.
Citation needed. Please provide testing that does not show degradation in performance on older PCs.
Evidence that it does on older low ram PCs has been shown here: http://murga-linux.com/puppy/viewtopic.php?t=84472
Please provide contrasting proof on an older P3 or earlier System with < 512mb ram


The topic is very apt. There are the facts the community keeps presenting and then there is the myth you keep telling.

Jim1911
Posts: 2460
Joined: Mon 19 May 2008, 20:39
Location: Texas, USA

Fact

#20 Post by Jim1911 »

My experience with Slacko 5.6.

Post Reply