Have more than 3GB of RAM? You need one of these Puppies

Using applications, configuring, problems
Message
Author
gcmartin

#31 Post by gcmartin »

I updated the thread to reflect ATOM processors.

This thread is as accurate as possible base upon what is seen in the industry and remembering what we did in Puppy testing of PAE,

Puppy does something rather unique in its operation that feeds the performance benefit from PAE use. PAE is NOT some special kind of Puppy. It is a Linux build that has the smarts to take advantage of built-in CPU features that were added after the original Pentium design. Intel never retooled the old RAM methodology or the newer PAE methodology. This design model continues and was embraced when the manufacturers moved from 32bit to 64bit. We continue to see the PAE RAM model in many/most of the 64bit CPUs, even today if you should run a 32bit distro on it.. If you have a 32bit PC and it can suport PAE, none of our testing has surfaced Linux degradation no matter how much RAM you throw at your PC. It just does NOT degrade performance. In the greater IT community, its been shown that under certain workloads, in "Tom's Hardware Comparison's" measurements, to provide positive performance results. This matches Intel's own decades old internal report. And, This is consistent with what we found in the testing here in Puppyland. PAE can and does have a positive impact by comparison..

This thread intends to shed light on community findings of PAE benefits. It is to be viewed as helpful in understanding the developer's PUPs that are being offered and the benefits that users will acquire in the use of PAE.

No one who has most of today's PCs should expect or will see any loss of performance in the operations of your desktop use with a Puppy PAE Distro. In fact, you should expect benefit in your use. All that Puppy LInux does is exploit the hardware for your use.

No offense anyone. This is for understanding. It is NOT designed to build camps of people for or against. It is information for person to understand the benefit being presented by developers.

It is hoped that it is seen as such. Here to help

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

#32 Post by Q5sys »

gcmartin wrote:We continue to see the PAE RAM model in many/most of the 64bit CPUs, even today if you should run a 32bit distro on it.. If you have a 32bit PC and it can suport PAE, none of our testing has surfaced Linux degradation no matter how much RAM you throw at your PC. It just does NOT degrade performance. In the greater IT community, its been shown that under certain workloads, in "Tom's Hardware Comparison's" measurements, to provide positive performance results. This matches Intel's own decades old internal report. And, This is consistent with what we found in the testing here in Puppyland. PAE can and does have a positive impact by comparison.
I'm trying to track down the Red Hat White paper, waiting to hear back from a few people about it, but there can be a performance hit. Thats even stated in the BSD docs
To access more than 3.2 GB to 3.7 GB of installed memory (meaning up to 4 GB but also more than 4 GB), a special tweak called PAE must be used. PAE stands for Physical Address Extension and is a way for 32-bit x86 CPUs to address more than 4 GB of memory. It remaps the memory that would otherwise be overlayed by address reservations for hardware devices above the 4 GB range and uses it as additional physical memory. Using PAE has some drawbacks; this mode of memory access is a little bit slower than the normal (without PAE) mode and loadable modules are not supported. This means all drivers must be compiled into the kernel.
When might this be an issue? Well here is one that I can think up... If a user grabs the PAE version of Slacko (for example), and then downloads a virtualbox pet that was built on a non-PAE slacko system. I'm not quite sure if there would be any issues in loading the vbox kernel driver. Will there be a problem? I dont know. Someone would have to test it out. If someone is having trouble running vbox this may be the reason and not be aware of it. (this is actually a good example of why its best to compile vbox on your system instead of trying to grab a pet for it)
Same also goes for TrueCrypt since it can utilize kernel crytpo services. I have tested neither so I dont know.


Does this render PAE a bad choice. No, it doesnt. I'd honestly doubt if most people could even notice a 5% performance issue in their ram. But for those that want to push things... its something to take into account. Toms Hardware isn't really a reliable source for fact... to many people on there making claims and not enough facts to back it up.

Benchmark: http://www.phoronix.com/scan.php?page=a ... _pae&num=5
Only a very small drop in performance can be found with the PAE kernel in the PostMark disk test, but the 64-bit kernel was immensely faster.

In the fourteen tests for this article we did not find using Ubuntu's 32-bit PAE kernel to have a dramatic performance impact whether it be positive or negative. Granted, we were using just 4GB of system memory that is common to many desktops, but if using 8GB, 16GB, or even a greater memory capacity the performance penalties are perhaps higher. By far though exhibiting the best performance was the Ubuntu 64-bit kernel that often ended up being leaps and bounds better than the 32-bit kernel. Unless you have technical or business reasons for not migrating to 64-bit Linux with compatible hardware, there is no reason to stick around with a 32-bit kernel and worrying about physical address extension.
The simple fact is that on a fundamental basis PAE has to be a bit slower because of the mapping requirement to be able to access all the ram. There have to be more access requests for memory space to find exactly what you're looking for.

I asked a friend of mine for some other test results and he gave me these: Showing PAE as slower than non-pae
Image
Image
Image


But here's one showing them the same
Image


And here's one showing mixed results.
Image


I also remembered another possible drawback with PAE... not all drivers support higher than 4gb memory addressing, so in the case of non typical hardware, some drivers wont work. This is a minimal instance, and probably wont affect many people but its still a possibility. The only possible issue that springs to mind right now has to do again with Vbox.

Does all of this support not using a PAE-kernel, not really. But facts are facts. People should at least be aware that there may be a minimal performance difference. It's a personal choice. <5% ram performance vs greater ram space.
I would be willing to bet most would desire the larger memory space, but I'm sure there are others who wouldn't. In any case, users have the right to know and make that decision for themselves. Which way do I go? Well both actually. For a laptop I'd opt for speed over space, for a desktop I'd opt for space over speed.

User avatar
01micko
Posts: 8741
Joined: Sat 11 Oct 2008, 13:39
Location: qld
Contact:

#33 Post by 01micko »

Q5sys wrote:... If a user grabs the PAE version of Slacko (for example), and then downloads a virtualbox pet that was built on a non-PAE slacko system. I'm not quite sure if there would be any issues in loading the vbox kernel driver. Will there be a problem? I dont know. Someone would have to test it out. If someone is having trouble running vbox this may be the reason and not be aware of it. (this is actually a good example of why its best to compile vbox on your system instead of trying to grab a pet for it)
Same also goes for TrueCrypt since it can utilize kernel crytpo services. I have tested neither so I dont know.
For that particular circumstance that is why I patch the Makefile of the kernel such that PAE or 4g is appended to the kernel version. That is, typing uname -r in console returns the kernel version with the 4g or PAE appended. That means that anything compiled in the other is not available for the native kernel.

While this could potentially cause issues with some woof scripts it hasn't so far. I tested it with Barry's version comparison utility vercmp and it seems to handle that scenario fine.

This is a common practice (patching kernel version). All the major distros (apart from Slackware) do this.
Puppy Linux Blog - contact me for access

gcmartin

#34 Post by gcmartin »

PAE is not some special "BIG" performance impacting hardware feature. Its a simple replacement model for providing access to RAM versus the old (and current) method of accessing.

It is exercised at the system level on behalf of system/subsystem requests.

It merely carries out request for RAM on behalf of the OS.

Here is also a definitive report by a respected group on benefits. It kinda mirrors what we have seen with PAE use in the Puppyland community testing that has been done.

Here is a simple explanation of PAE hardware which also has been echoed previously published in threads.

Hope anyone who might be confused can get a good understanding from this information as well.

Be aware, that I am consistent in suggesting that should anyone have a 64bit PC, they may find very good packaging from FATDOG64 and from LightHouse64.

For 32bit PCs, most come with PAE and should you have one, a PAE version of a Puppy, when offered will give excellent performance which equals what you would get from its non-PAE conterpart. This is what the testing has shown. And, you can be assured that it will make use of ALL RAM you can throw at it for your workloads.

And non-PAE will perform on all platforms. But be aware of its limitation.

Hope these references brings clarity.

Atle
Posts: 596
Joined: Wed 19 Nov 2008, 12:38
Location: Oslo, Norway
Contact:

#35 Post by Atle »

Thanks for a really wanted and needed crash course in uppers and downers on PAE. Now I even know what it means :D

Despite getting dragged trough complex information so hard to understand that the radio needed to be turned off for a while here, I still want to point out that the issue for me is that i feel Puppy is like a "Light linux", always listed on "mini distro" articles and recommended for small spec hardware.

And its there things goes wrong in my eyes. The current flagship is PAE and that is not so good for getting new users. At least there should be a better set of information on the major download site,

Once the flagship became PAE, i have noticed that Puppy is dropping at Distrowatch and that might be the reason. People THINK they will get a sleek nice minidistro for their wifes computer or their old one etc, and they end up with not being able to boot the live CD they burn. THAT is the problem. Not PAE or 64 bits this and that.

Most people i know would not be able to tell you what PART of technology PAE derives from. Is it shortterm for some rocketscience stuff? A vaccine? Those batteries burning inside the Dreamliner? What what what?

I fancy Puppy for its simplicity and HORRIBLE ability to just do things that are ABSOLUTELY stunning within computing. I love the idea of PAE. Its great.

But its not so great that a certain percentage of new users that has burned their CD and awaits a AWESOME experience are left with nothing but a message... This PC is not PAE compatible...

Do they then come back to puppyland to search the forum for a workaround or just jump to another distro or even WORSE... stick to Windowz.

PAE is here to stay and its good news. But that said, please change the flagship to the NON PAE and make sure there is CLEAR information about when one should download and try the PAE version.

I think the entire Puppyland will grow stronger under such a presentation and that new users are more likely to LOVE puppy at first sight if the FIRST experience is bootable for ALL.

I honestly felt like left out when i tried this PAE first time... It was like... Huuuuu :shock: Did they leave all of us that think OLD computers are great and are struggling with 512 ram in 2013?

Then... after some time i discover I can try "retroprecise".

While the floor is mine for a second... Retro does mean among other things BACKWARDS... That is not very excellent considering one is going forward.

To conclude... A first time user experience should come before anything else and having non bootable CDs for a certain percentage of new users is not gone do Puppyland any good.

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

#36 Post by Q5sys »

Atle wrote:To conclude... A first time user experience should come before anything else and having non bootable CDs for a certain percentage of new users is not gone do Puppyland any good.
I couldn't agree more. PAE should be an option, but not the default.
(Right now with Slacko that is the case right? Anyone know about Presice?)

Atle
Posts: 596
Joined: Wed 19 Nov 2008, 12:38
Location: Oslo, Norway
Contact:

#37 Post by Atle »

Precise is the current flagship, and the PAE version is the "main option".

Slacko does not have PAE version as a primary choice as far as I have seen.

One can have a look [ur=http://puppylinux.org/main/Download%20L ... lease.html]here[/url]

I guess this is the page where most people have their first landing in Puppyland. I guess most people will end up downloading the first option and that is the Precise with PAE.

gcmartin

#38 Post by gcmartin »

Hi @Atle

Could you run this command and post results here using the PC you had problems booting, please.

Code: Select all

hardinfo -r -f html > report.html.gz
At the bottom of your post, you have the option to attach the "report.html.gz" file the command creates.

Maybe we can spot something
Thanks in advance

gcmartin

#39 Post by gcmartin »

I just went thru the Slacko thread. I am looking for evidence of some large (or small) persons reporting PAE is not working.
Edited: I, also, just went thru Barry's Precise thread, too.

Since in excess of 98% of all desktops/laptops sold have the PAE feature, especially those machines sold before 2006 (which is what is being presented by members recently in this thread), I (and you too) should expect to see complaints that PAE is not working from its ISO.

Where is the evidence???

P310Don, whom has helped us in the past by bringing the early ATOM processor issue to our attention and (I think) another member did also share that they had a rare Intel (newer than 2005) that did not have the PAE. These are two accounts of the many many who have posted and shared their configurations and their boot problems in the forum. Everyone else was able to boot and run without issues. Everyone!

The most prevalent problem with booting is the ISO hash on downloads....not PAE. And those of us who have been around Puppy Linux for awhile and help by testing have, on accasion, gotten a bad download whch borks at boot time.

So where is the evidence that any PAE is turning off users....old or new and steering them away from Puppy, as is being suggested? i, personally, have been responsible for at least 40-50 people over the last 3 years to boot PUPs and of late, to boot Slacko and Barry's RACY. Not one has reported issues. Somewhere, or somewhere else, there must be evidence of what you share.

This is not a challenge, its a request for assisting information.

Please help.

Atle
Posts: 596
Joined: Wed 19 Nov 2008, 12:38
Location: Oslo, Norway
Contact:

#40 Post by Atle »

I am doing this hardware test as i write, but its been going on for like 25 minutes now. I think the hardware test failed on my machine as its gone near silent now...

But let me make a test as suggested, as I can create a bootable USB as done before, check that the PAE version works on a newer laptop I can access tomorrow and then test it on my own personal stock of crap hardware. I got 4 laptops and that is:

Dell Latitude D505 512 Ram
Acer Aspire One ZG5 512 Ram
Fujitsu Siemens Amilo X1 2 gig Ram
Fujitsu Siemes Lifebook S7010D 1 gig ram

Anyhow... I need to boot into Lucid as i think there is a feature there that can verify the download of precise to be ok or not.

If the ISO is ok, is a USB boot good to go as a test?

I can report on all four machines pretty fast once iso is verified.

I am running Slacko as we speak.

Atle
Posts: 596
Joined: Wed 19 Nov 2008, 12:38
Location: Oslo, Norway
Contact:

#41 Post by Atle »

Tested 3 machines with MD5 verified ISO. precise 5.4.3(current)

Dell Latitude D505 512 Ram NOT working with PAE
Acer Aspire One ZG5 512 Ram Working with PAE
Fujitsu Siemes Lifebook S7010D 1 gig ram NOT working with PAE

Error message given upon boot is as follows:


Loading initrd.gz.........................ok

This kernel requires the following feature not present in CPU
pae

Unable to boot. Please use a kernel appropriate for your CPU


Fujitsu Siemens Amilo X1 2 gig Ram was not tested as it seem to have given up being booted from USB, but a test will also be done and posted as soon as i have a boot CD available.

gcmartin

#42 Post by gcmartin »

Hi @Atle
That makes sense in the messages you show. Thanks.

What is shown is good, in that, the Puppy distro builders have features in the system which alerts the user that the distro will not work on that hardware.

This is a similar message one gets when someone boots a 64bit FATDOG64/LightHouse64 PUP on a 32bit PC.

This message is clear and one would assume it to be useful to the viewer, too.

Admittedly, this can be disheartening,but, certainly should not be a "show stopper". For the right person(s), it would begin a trek to read or enter the forum.

I can see where on a rare occasion, one might end with all of their PCs being non-PAE.

I acknowledge your pain and your concern.

But, now, knowing that most PUP distros which are, both, 32bit and non-PAE will work for your class of PCs. This kind of things should be helpful versus a discouraging event. The Puppy community does NOT try to discriminate against anyone who has non-PAE PCs. It is my observation that most every 32bit PUP developer, including Barry, tries to present a non-PAE version to accommodate those PCs as well.

I have been in IT all of my life, and I have never owned or used a Netbook. And, I do fully recognize these as the planned follow-on to smaller laptops; that is, before the industry switched to tablets.

I hope this has been a help as your experience helps all of us see what you share. Now, should someone present these messages in the forum, we can be a help in sharing useful information.

You inspire enough that this may be a good time to post a thread or a document somewhere for users who would have some typical problems at boot time. And as well, there may be some guidelines that would suggest that if users see a particular message, what they should do about it.

Lastly, I have seen more and more PUP distro developers offering suggestions what platforms they foresee their distro running on. As such, this also is efforts by the development community to guide users in where they can expect some safe, tested operating environments.

These are just observations.

Again, Thanks for taking a moment to share your findings.

gcmartin

#43 Post by gcmartin »

Atle wrote: ... I am running Slacko as we speak.
On this PC, would you mind running the hardinfo command on it and posting your report, please. Is hardinfo hanging on this PC as well?

Thanks

Atle
Posts: 596
Joined: Wed 19 Nov 2008, 12:38
Location: Oslo, Norway
Contact:

#44 Post by Atle »

Its the same computer and the same Slacko. Its hangs at John Walker Benchmark or so, but yet there might be a useful information in the output given.
Attachments
report.html.gz
(58.6 KiB) Downloaded 1187 times

Atle
Posts: 596
Joined: Wed 19 Nov 2008, 12:38
Location: Oslo, Norway
Contact:

#45 Post by Atle »

This message is clear and one would assume it to be useful to the viewer, too.
I think this message is clear to about 0.3 percent of the world population.

Imagine a young girl that has a little brother whom she would like to help to get started with computer, and she reads about Puppy Linux, THE BEST CHOICE for old computers. She reads this in a online Linux magazine.

She has only one burnable CD left. She uses prepaid "pay pr mb" mobile internet. Downloading nearly 200mb costs the same as she makes washing dishes at the local cafe for one evening. Since so recommended for low ram and low spec computers, she strikes out for the biggest download she ever dear to do.

It took 5 hours to download the file and its ready to boot.
Image

You think this clear message will help her?

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

#46 Post by p310don »

That girl's face above is kind of what mine is looking like reading through this thread...

Having been part of the initial thread a couple of years ago that started this whole PAE for puppy thing, I have done some testing, research and have some experience and observations regarding PAE.

Firstly, despite Q5sys's assertions from what he has read, in my actual user experience, PAE does not have a 5% reduction in performance. Running through hardinfo tests (which aren't exactly definitive) the differences were more along the lines of 0.1%. Hardly enough to care about.

I have tested PAE on at least 6 PCs. It worked on 5. It didn't play nicely with my eeePC with an Atom processor. These things are / were really popular, and Puppy and puppy derivatives are often touted as being great for them. Think Jemimah's puppeee. Very popular computer that puppy won't boot on? Not cool. PAE works on my Celeron 600 something or other, AMD quad core, Pentium 4 something, AMD dual cores etc. For a PC that supports it, there is a theoretical difference, but it is not noticeable.

On my current main system, the Quad core AMD, I run Saluki. Initially I had 4gig of ram, running non PAE, so seeing about 3.2 and no swap. It's puppy after all, who needs all that ram. This was all well and good until I started using virtualbox to run windows whilst multitasking with other stuff, probably firefox and facebook. I ran out of ram, so bought 4 more gig, and changed to the PAE kernel for Saluki. On this PC with it's larger access to RAM provided by PAE, my stability and usability has increased. Virtualbox runs exactly the same as before, without the crashes. No recompilation required. And I can allocate 4gig of ram to windows in virtualbox, which it does need, coz its windows. I did have to compile the driver for nvidia as it didn't work with the PAE kernel. No problems there though.

This alludes to what I said in my previous post, and I think Atle highlighted it as well. When it comes to PAE, its best to know what you're doing and making an informed choice. For the girl in the story and pic who doesn't know what PAE even is, why bother? For the Puppy audience rejuvenating an old PC with less than 4gig of RAM, why bother? For the newer PC with more than 4gig of RAM, its awesome. For the newer PC with more than 4gig of RAM, 64bit should be awesomer.

In other threads dealing with this topic, I have said 64bit is the future, but, in the limited puppyland, 64bit is a tiny portion, with limited developers and limited support. PAE is a great intermediate step, to be able to use many of the 32bit PETs we have, without having to recompile or redesign for 64bit. The best thing for Puppy's future on the Modern desktop is to help Jamesbond et al working on 64bit to get past 32bit of all flavours.

gcmartin

#47 Post by gcmartin »

@Atle

My Dell Inspiron 6000 has the same specs as your 2GB laptop. I have never experienced any issuesi using PAE on its 1.5Mhz Pentium M. This is an old laptop. My Hardindo works well.

I am not trying to take a position to suggest what you should/shouldn't use. But, I just reference this as you may be having some RAM chip problems or there is some other hardware problems which is causing that laptop to not be completing a Hardinfo task.

Further for that laptop, I have and continue to run PAE for the services it provides and to assist in testing when developers are prepping their distros.

As said, it is hoped that this experience with PUPs gives insights about your hardware and what can and cannot be used.

But, I still believe that developers make choices and try to address and benefit the largest possible population with the widest possible peripheral degrees, here in Puppyland.

I get no greater satisfaction to try to support them as best I can same as the many other members of this community, with the limited skils I possess.

Hope this experience and dialogue has been helpful to all participants, even in some small way.

Here to help
Edited: 1st line info of PC
Last edited by gcmartin on Tue 19 Feb 2013, 17:49, edited 2 times in total.

Atle
Posts: 596
Joined: Wed 19 Nov 2008, 12:38
Location: Oslo, Norway
Contact:

#48 Post by Atle »

making an informed choice.
Think your spot on there. People deserve to be given a more informed choice, rather than loosing and disappointing a unknown percentage of new users that will probably loose faith if its not booting the first time.

My perspective of this is purely a "marketing" perspective.

One can only guess how many people came to our car shop and the car in the shop did not start. Do you start to make request to the salesman then or do you go over to the next shop with a different brand(could be slitaz, tiny core or lubuntu etc)

One can never ever change the first impression of anything. Its holy ground and I honestly feel we are messing with that ground, not matter if its 50% or 5% percent of the first time puppy users that gets the clear message(boot failure), instead of seeing the desktop arrive at a amazing speed and get hooked on puppy and become a citizen of puppyland.

Unless that is... Puppy land is getting overpopulated :roll:

Atle
Posts: 596
Joined: Wed 19 Nov 2008, 12:38
Location: Oslo, Norway
Contact:

#49 Post by Atle »

I get no greater satisfaction to try to support them as best I can same as the many other members of this community, with the limited skils I possess
I think you do great Martin...

I am not a technical guy, just a refugee from Windows that i dislike so much for political reasons and my skills would more be in the world of marketing.

From a marketing perspective one rather satisfy EVERYONE, rather than a unknown number.

And given that retroprecise has the same amount of packages etc, there is not really any reason to make a PAE version the "default" download.

Or... as said... at least give people a chance to make a decision based upon facts as in informing them better... That is for free i guess.

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

#50 Post by Q5sys »

gcmartin wrote:And non-PAE will perform on all platforms. But be aware of its limitation.
What limitation? Unless you have 4gb ram or greater, there is no limitation. And for those that do, they are most likely running a 64bit processor, so would better be served by a 64bit release. Making a 32bit OS run better for 64bit processors at the cost of negatively impacting older 32bit users is kinda silly in my mind.
p310don wrote:Having been part of the initial thread a couple of years ago that started this whole PAE for puppy thing, I have done some testing, research and have some experience and observations regarding PAE.

Firstly, despite Q5sys's assertions from what he has read, in my actual user experience, PAE does not have a 5% reduction in performance. Running through hardinfo tests (which aren't exactly definitive) the differences were more along the lines of 0.1%. Hardly enough to care about.

I have tested PAE on at least 6 PCs. It worked on 5. It didn't play nicely with my eeePC with an Atom processor. These things are / were really popular, and Puppy and puppy derivatives are often touted as being great for them. Think Jemimah's puppeee. Very popular computer that puppy won't boot on? Not cool. PAE works on my Celeron 600 something or other, AMD quad core, Pentium 4 something, AMD dual cores etc. For a PC that supports it, there is a theoretical difference, but it is not noticeable.
I was curious as to the results everyone here is speaking about. I would be interested in seeing the 'results' that people are basing their opinions on. Because the only results I have seen before show that PAE isnt quite as fast as non-PAE.
So i decided to spend some free time today (which I happen to have) and compare the newest Slacko 5.4 on 4 machines. I burned the PAE and the non-PAE ISOs, and booted them.
Each computer was tested the same way. Boot the system from the CD, load the OS into ram, and run Hard-info. Nothing else was done or changed. Here is the link to the full report, but I'll quote the results here. Anyone interested can go to that link for the graphic results and the hard-info reports. In my mind a result of less than 1% is negligible.
Q5sys wrote:On Modern Hardware the difference is minimal, aside from a very interesting result on a a quad core i5 with 16gb ram. This system was fully able to utilize the benefits of PAE, however did so with an unexpected side effect. Obviously on an modern system with more than 4Gb ram, a user should be using a 64bit release.

The worst result as I expected was on older hardware, where ram is minimal and is precious. The system had enough ram to boot and as such the ISO loaded into ram. The extra memory used for PAE, which is required, impacts older minimal hardware more than modern hardware. On older hardware running PAE is a severe drain of precious resources, and as such should NOT be encouraged.

Overall the results are mixed. On Modern hardware PAE is slower than non-PAE in some tasks but not others. However on older hardware it was slower across the board, and on older hardware a 2% performance decrease is much more noticeable in the user experience than on a modern system.



Toshiba 7200 Cte Laptop
Total RAM on PAE is 2% less than non-PAE.
Free RAM on PAE is 36% less than non-PAE.
Benchmarks -
Blowfish - PAE is 2% slower than non-PAE
CryptoHash - PAE is 1% slower than non-PAE
Fibonacci - PAE is 1.5% slower than non-PAE
N-Queens - PAE is 1% slower than non-PAE
FFT - PAE is 2.7% slower than non-PAE
RayTracing - PAE is 2.3% slower than non-PAE

Acer Aspire One D257 Netbook
Total RAM on PAE is 1% less than non-PAE
Free RAM on PAE is 1% less than non-PAE
Benchmarks -
Blowfish - PAE is <1% slower than non-PAE
CryptoHash - PAE is 2% faster than non-PAE
Fibonacci - PAE is <1% slower than non-PAE
N-Queens - PAE is 2% faster than non-PAE
FFT - PAE is 1% slower than non-PAE
Raytracing - PAE is 2% faster than non-PAE

Lenovo Y510 with 4gb ram
Total RAM on PAE is 1% less than non-PAE
Free RAM on PAE is 3% less than non-PAE
Benchmarks -
Blowfish - PAE is 1% faster than non-PAE
CryptoHash - PAE is 1% slower than non-PAE
Fibonacci - PAE is 7% faster than non-PAE
N-Queens - PAE is <1% slower than non-PAE
FFT - crashes hard-info - so no results
Raytracing - crashes hard-info when run - so no results

Custom Built Intel i5 system with 16gb ram
Total RAM on PAE is 530% greater than non-PAE (what we would expect)
Free RAM on PAE is 603% greater than non-PAE (what we would expect)
Used RAM on PAE is 1% higher than non-PAE (what we would expect due to how PAE works)
Benchmarks -
IMPORTANT DISCLAIMER: PAE system @ 2800Mhz compared to non-PAE running @ 1200Mhz
Blowfish - PAE is <1% faster than non-PAE
Cryptohash - PAE is 9% slower than non-PAE
Fibonacci - PAE is <1% slower than non-PAE
N-Queens - PAE is <1% slower than non-PAE
FFT - PAE is <1% faster than non-PAE
Raytracing - PAE is <1% faster than non-PAE
This is not what I expected. On this sytem I expected PAE to be much faster than non-PAE. The fact that PAE is only marginally faster than non-PAE when run at 233% processor speed is very suprising to me. Moreso considering that when Cryptohash on PAE was running the CPU at 2800Mhz it was 9% slower than the non-PAE running Cryptohash with CPU at 1200Mhz.
(I preformed this test 3 more times and got the same results, so its not a random result)



p310don wrote:In other threads dealing with this topic, I have said 64bit is the future, but, in the limited puppyland, 64bit is a tiny portion, with limited developers and limited support. PAE is a great intermediate step, to be able to use many of the 32bit PETs we have, without having to recompile or redesign for 64bit. The best thing for Puppy's future on the Modern desktop is to help Jamesbond et al working on 64bit to get past 32bit of all flavours.
I agree with you on this, I think our focus should be on 64bit for modern hardware, and 32bit for older hardware. My Lenovo Y510 that I used in that test was made in 2008 and has a 64bit processor. Pretty much any system that can run more than 4Gb ram has a 64bit processor on it. Motherboard Companies arent stupid, they arent going to rely on a software patch they have no control over to make their boards usable by the majority of the market.

Locked