Amiga.org
Amiga News and Community Announcements => Amiga News and Community Announcements => Amiga Magazines => Topic started by: Daff on August 11, 2009, 03:05:39 PM
-
The magazine Obligement published today a comparative about performances between AmigaOS 4.1 and MorphOS 2.3 on Pegasos II.
This comparative is based on 20 groups of benchmarks including boot time, processor, 3D, USB, IDE hard disk, WarpOS emulation, emulation 68k, etc.
No subjective comparison was made, all benchmarks are based on numbers.
The article is in french but very easy to understand, and it's available on http://obligement.free.fr/articles/amigaos41_vs_morphos23.php
-
Can not read frenchm, but from the figures would appear that AOS4 is much much slower at nearly everyhting.
What gives
-
@JJ
The article is in french, but you can get the page translated.
http://66.196.80.202/babelfish/translate_url_content?.intl=us&lp=fr_en&trurl=http%3a%2f%2fobligement.free.fr%2farticles%2famigaos41_vs_morphos23.php
x303 :D :D :D
-
my life is over, it is ruined...... aos4.1 lost and morphos win..... :(
-
OS4 is completely run over by MorphOS. A good testemony to the MorphOS developers, and a big congratulations is in place! :-)
MorphOS has more and better features.
MorphOS has better Amiga compatibility.
MorphOS has better performance.
Now, is there a way of comparing overall system stability in general usage?
-
my life is over, it is ruined...... aos4.1 lost and morphos win..... :(
Why?
Get a Mac Mini G4 1.5GHz with 64M graphics memory and wait for the MorphOS release for that one! This will be one heck of experience, the fastest Amiga experience ever! :-)
-
IIRC the AOS4 Pegasus port uses a BIOS API for the hardware, can't remember it's name, that's probably one of the reasons why it's slower. As for the rest, dunno, they seem to have sacrificed compatibility for new features. But for some stuff MOS guys seem to have managed to do both at the same time:)
-
will there be any updates for the classic amigas?
-
@takemahomegra
"MorphOS has more and better features.
MorphOS has better Amiga compatibility.
MorphOS has better performance.
"
I see no compatibility or feature comparison there it's mostly about peformance.
2 things are not taken into account, the optimization of the software and that API thing I mentioned above.
And as the article says, which I think applies to most things:
"L'exemple de ZoneXplorer montre d'ailleurs que le Pegasos II G4 est plus lent qu'un AmigaOne G3. On peut aussi noter que la performance de certains tests est plus due à la façon dont a été compilé le logiciel qu'au système lui-même."
-
Interesting to see someone taking the time to test both but considering OS4.1 is not optimized for this hardware somehow kills the "fairness" of the experiment.
Whatever the results, its how much I pay that will make me choose at the end....isn't it why I use a PC now? :P
-
@takemahomegra
I see no compatibility or feature comparison there it's mostly about peformance.
Indeed *this* test is about performance, the other things are known since before. Now, if we could find a way to compare the overall stability...
-
my life is over, it is ruined...... aos4.1 lost and morphos win..... :(
Morphos have an active support and amiga os4 thats is the difference.
-
Interesting to see someone taking the time to test both but considering OS4.1 is not optimized for this hardware
It isn't? Any link or so to back up such a serious claim? It surely hasn't been marketed and advertised as an "unoptimized" beta level OS to the paying customers, you know, rather the opposite. BTW, can I quote you on this on other forums?
And if your testimony would be true and correct, who is to blame? Surely not the MorphOS team. If Hyperion can't deliver a quality product that can give the competition a match, then it's *their* problem!
*This is* a fair comparison of what the paying customers of today get for their money. To bad if you don't like the results, but please don't make up lame excuses.
AFAIK, OS4 cost less, and I guess you get what you pay for...
-
@takemahomegrandma
The article states at the end that a G3 A1 performs better at some tests than the Peg2 (using AOS4), shouldn't that be enouph ?
That said I think it is a fair comparison if one takes into account what is actually compared, and that is the combination of the current Peg2 version of the OS and the current versions of the applications tested.
One thing that I think MOS would win anyway is on 3D, IIRC there was a design flaw somewhere because of Picasso RTG or something...
-
@takemahomegrandma
The article states at the end that a G3 A1 performs better at some tests than the Peg2 (using AOS4), shouldn't that be enouph ?
I have not the slightest idea of what the text in the article say, since it for some strange reason is written in an none-English language, that only a fraction of the world can understand. But numbers are numbers anywhere!
That said I think it is a fair comparison if one takes into account what is actually compared
Yes indeed, the released (and sold) versions of OS 4.1 for Pegasos2, and MorphOS 2.3 for Pegasos2.
MorphOS wins.
-
Well yeah, I supose if features, speed and compatibility are your top criteria, I suppose then Morphos is better.
OS4 is completely run over by MorphOS. A good testemony to the MorphOS developers, and a big congratulations is in place! :-)
MorphOS has more and better features.
MorphOS has better Amiga compatibility.
MorphOS has better performance.
Now, is there a way of comparing overall system stability in general usage?
-
So, MorphOS wins! Looks like OS 5 (or whatever) may have some catching up to do :)
I've yet to try either, heh. But just out of interest, is there a huge difference in performance between OS4 on Pegasos and OS4 on Sam? Has anyone tried both?
-
So, MorphOS wins! Looks like OS 5 (or whatever) may have some catching up to do :)
But just out of interest, is there a huge difference in performance between OS4 on Pegasos and OS4 on Sam? Has anyone tried both?
Of course! A *huge* difference! At least it should be! Have you even looked at the specs?
A Pegasos2 with a G4 (with both Altivec and L2 cache) @ 1000MHz *runs circles* around any embedded "G2" processor like the AMCC440/MPC5200b. Even the Peg2 G3 might win. The lack of L2 cache will be especially obvious when running older amiga applications. :-)
-
Of course! A *huge* difference! At least it should be!
That's not an answer to the question, though. :hammer:
-
will there be any updates for the classic amigas?
I would say, probably not :(
-
That's not an answer to the question, though. :hammer:
No?
Please correct me? :-?
-
@x303
Thanks for posting the url for the translation to English. Sadly, I don't know French well enough to follow the original article.
@all
Interesting test results. I think this round goes to MorphOS 2.3.
---
redfox
-
MorphOS p\/\/n3d OS4.1. There it is in plain black n white, the numbers dont lie. Congratulations MorphOS developers, then again, I knew those coders were the best!
-
If I had a Peg2, I think I'd be happy enough that I could run both operating systems in the first place. As it stands, I don't think I'll ever see MOS 2.3 on my A1.
-
No?
Please correct me? :-?
It is obvious that there _should_ be a huge difference between OS4 on peg2 and OS4 on SAM, but the question was if there really _is_ a huge difference. :hammer:
-
It is obvious that there _should_ be a huge difference between OS4 on peg2 and OS4 on SAM, but the question was if there really _is_ a huge difference. :hammer:
OS4 for the SAM is rated beta, OS4 for the Pegasos is *not* beta. Anyway, someone really surprised by the result of this comparision?
-
Of course! A *huge* difference! At least it should be! Have you even looked at the specs?
A Pegasos2 with a G4 (with both Altivec and L2 cache) @ 1000MHz *runs circles* around any embedded "G2" processor like the AMCC440/MPC5200b. Even the Peg2 G3 might win. The lack of L2 cache will be especially obvious when running older amiga applications. :-)
Heh, you caught me ;) I have to admit I knew very little about the Peg2 specs, but for some reason I had it in my mind that they were a similar speed to the Sam :o Dunno where I got that from.
So anyway, yes, that answers my question! It's been quite a long time since I've done anything Amiga-y, so I'm kind of catching up on what's been going on.
-
So, anyone done a comparison of OS4 on peg2 vs. OS4 on SAM440 yet?
-
So, anyone done a comparison of OS4 on peg2 vs. OS4 on SAM440 yet?
I doubt you ever see any except those ones where SAM is faster than Pegasos 2.
-
Semi off-topic: Why so many interesting articles are written in german, french and polish language? This is a waste, imho, but it's better than nothing.
On line translators doesn't give very good results.
-
MorphOS p\/\/n3d OS4.1. There it is in plain black n white, the numbers dont lie. Congratulations MorphOS developers, then again, I knew those coders were the best!
You did?
So the difference is bad code? Have you got any facts backing up this statement? and with facts I do not mean assuming that something is or is not in ones fantasy.
-
@takemehomegrandma: I have no idea what you are talking about and why I have pissed you off that much. (must have hit the sensible Peg2 spot...sorry)
I was commenting from the point of view of a user that has no more Amiga but is hoping for new hardware/software to be available at a fair price with average consumer level standards.
This test made by the french guys is merely stating that MorphOS runs faster on the Peg2 thans OS4.1 with numbers to prove it. They also say at the end:
""Il faut mettre tout de même un bémol à ce comparatif. Il a eu lieu sur un matériel sur lequel l'AmigaOS 4.1 est soi-disant en version "bêta"."""
Which translates roughly to: "Don't take these comparisons too seriously. They were made on hardware that OS4.1 is supposed to be in beta version"
For the buyer (me for example) that means, nothing is final, stay tuned for more.
-
You did?
So the difference is bad code? Have you got any facts backing up this statement? and with facts I do not mean assuming that something is or is not in ones fantasy.
LOLOL!!!!
Here is my post again:
You did?
MorphOS p\/\/n3d OS4.1. There it is in plain black n white, the numbers dont lie. Congratulations MorphOS developers, then again, I knew those coders were the best!
Read it. Read it again. Cut-n-paste it into your favorite Amiga text editor (CygnusED is mine) and search for "bad code". Do that a few times. Then re-read my quote again. Rinse, repeat. Come back and then edit your post with a "my bad". :) Nowhere did I say it was bad code. I just said in layman's terms MorphOS beat OS4.1 on the same system.
-
LOLOL!!!!
Here is my post again:
Read it. Read it again. Cut-n-paste it into your favorite Amiga text editor (CygnusED is mine) and search for "bad code". Do that a few times. Then re-read my quote again. Rinse, repeat. Come back and then edit your post with a "my bad". :) Nowhere did I say it was bad code. I just said in layman's terms MorphOS beat OS4.1 on the same system.
I'd go as far as saying there indeed IS bad code. How else can one explain some of those multiple times worse figures.
-
I'd go as far as saying there indeed IS bad code. How else can one explain some of those multiple times worse figures.
Indeed. And code is written by the coders, and the resulting quality is depending on their education, experience and motivation (in other words: competence).
I would also guess that OS4 doesn't make as much use of the Altivec that MorphOS does. It would be kind of interesting to see this test done all over again, but with a Pegasos 2 *G3*. Then you should be able to spot the impact of Altivec, which I definitely think is there.
-
So the difference is bad code?
Seriously what else could it be as the hardware is the same. Or not bad but just not as good / optimized which makes it bad compared to MOS ;)
-
Seriously what else could it be as the hardware is the same. Or not bad but just not as good / optimized which makes it bad compared to MOS ;)
As umisef already pointed out, there are lots of things. MMU setup for example. For all I know AOS4 uses the MMU to invoke the 68k emulator when jumping from PPC into legacy code. This causes a context switch which in turn penalizes performance. MorphOS uses a different approach without causing exceptions. So it's better, while the AOS4 way makes it more "comfortable" in from the OS point of view.
Moreover, while both AOS4 and MorphOS have virtual memory, AOS4 needs to keep far more complex MMU tables for their paging stuff -- also not all TLBs fit into cache at the same time, this causes penalities during memory access. I'm not sure, maybe additional complexity comes from the Supervisor mode MMU tables in this case during context switching.
Finally, and ChrisH won't like this: The Slab allocator is inferior to TLSF. That's a fact. Read the publications, do some research. Slab has higher internal fragmentation (i.e. it wastes memory) and because it wastes memory, CPU and L2 caches are not used optimally because the used memory is more sparse. The external fragmentation on TLSF is also very small, but I don't have a comparing figure to the Slab allocator right now. Also I suspect (consider this not confirmed) that due to the paging that AOS4 allows, external fragmentation might be further increased due to alignment restrictions to MMU page sizes (probably paging is disabled for very small allocations). Allocations and freeing are always O(1) on TLSF with the amount of instructions being not significantly higher than other memory allocators, especially since Slab is very complex and I'm fairly sure that Slab is no O(1) allocator (but probably still sub-linear, so this is certainly an improvement over AmigaOS 3.x).
These are all design decisions that have nothing to do with bad coding style. More with design goals and experience. Live with it. AOS4 allows paging, but this makes the system slower overall. Similar to having memory protection on other operating systems and separate address spaces for every process. You gain something, but you lose other stuff.
I dunno why USB is so slow under AmigaOS4, but having seen that Poseidon under AROS also performs "not optimal" for access to USB sticks due to the non-caching implementation of the FAT file system, I'd rather suspect the CrossDOS file system of is worse than the MorphOS implementation there. One could try "scsispeed" to out-rule the filesystem rather than the USB subsystem.
-
So the code is rotten by design ;) That is even worse as they would have to change the whole design in order to fix it :P
-
So the code is rotten by design ;) That is even worse as they would have to change the whole design in order to fix it :P
Don't think you really understand what was said or what you were saying. What has been said is that the code is different as it has/had different goals, which was not just speed.
-
IIRC the AOS4 Pegasus port uses a BIOS API for the hardware
No. OS4 Pegasos port uses the smart firmware, just like MorphOS. Look for another explaination...
Or just accept the fact OS4 is slower, and live with it :)
-
or chould look how slow new versions of windows work on older machines which werent desined for l guessing as am not a programer or know much about this at all
maybe os 4.1 runs better on hardware was orginally coded on eg A1
and of course sam 440 is designed after this point as well with more modern
sata hd and faster acess which must help alot and faster memory interface
it seem realitive anyway most machines run real good when first install get bit slower more things u have running just a thought.
makes intresting reading
-
maybe os 4.1 runs better on hardware was orginally coded on eg A1
When comparing other OSs on the same platform, we don't use any explainations. It's faster or slower. Period. There's no "well, maybe Windows uses a different way than Linux to initialize the memory controller and this explains it's slower... Linux hasn't been optimized for NVidia boards so 3D speed is inferior",...
It's faster or slower, period.
OS4 is current slower than MorphOS on the same hardware. Now just live with it ! This isn't the end of the world :)
-
i for one don't care which is faster,whats important to me is that it is usable and most important it is stable.