PDA

View Full Version : How fast is MorphOS?


Dan
08-02-2003, 07:52 AM
How fast is the 68k emulation in MorphOS compared to a 060-50Mhz?
UAE on my PC is 6 times faster than my A1200 060-50Mhz, I clocked it when decoding jpgs in Collector, is Pegasos/MorphOS faster than that?
How about the PPC JIT for MorphOS?

KennyR
08-02-2003, 08:14 AM
Believe me, MOS on Pegasos feels a lot faster than UAE on any PC, gigahertz CPU or not. Pure emulators just don't come close.

Triumph
08-02-2003, 08:36 AM
In emulation speed with a 'outdated' Trance (nov '02) it is sligtly slower that a Amitlon on a Athlon 1700+.
I did test this with the use of Pagestream4.1 x .
And also slightly slower on UAE on my laptop 2.4 Ghz P4.
It is just a 'feeling' test. As I don't belive in 'numbers' but more in everyday use.

Dietmar
08-02-2003, 09:42 AM
Believe me, MOS on Pegasos feels a lot faster than UAE on any PC, gigahertz CPU or not. Pure emulators just don't come close.

Below are the compilation times of a program (GoldED) compiled with SAS/C under UAE, UAE with JIT and MOS without JIT. Compilation is a reasonable test for raw performance. Integer performance obviously. The compiler is run with precompiled headers to take HD speed out of the equation.

UAE 0.8.17r3 is used, which is not the latest version, for obscure personal reasone: it works better with a tablet. The PC@1.4 GHz has 512 MB DDR-RAM, 40 GB Maxtor HD and 160 GB RAID; Maxtor used for test. The Pegasos@600MHz has 256 MB SD-RAM and a 40 GB Barracuda IDE.

UAE /w JIT: 1: 59 min
UAE no JIT: 11:52 min
MorphOS 1.3 no JIT: 4:19 min

Given those figures, the PC with UAE+JIT is about 2.5 times faster than MorphOS. However, the PC's CPU is also clocked 2.3 times faster. A crude guesstimate would be that a Pegasos/MOS1.3 and UAE/JIT are running emulated code equally fast on CPUs with the same clock frequency.

Now the interesting part: JIT massively contributes to the speed of UAE. JIT makes UAE compile seven times faster. If JIT for Pegasos (which will start shipping with MOS1.4?) has an effect anything like that, it is quite possible that a Pegasos at 600MHz will, with respect to emulated code, meet or exceed the performance of very fast PCs.

seer
08-02-2003, 09:47 AM
And offcourse it will be even better if/when native MOS programs are used.. IMHO that's far more interesting then emulated programs.

Dietmar
08-02-2003, 09:53 AM
To put numbers above a bit into perspective, it took 3.5 hours to compile the four different versions (68000-68040) on an A3000/25 ;)

Cymric
08-02-2003, 10:07 AM
@Dietmar:

Dear lord. And then to think we once thought the A3000 to be the hottest thing since sliced bread... How did we ever manage with those machines?

Targhan
08-02-2003, 10:19 AM
Gyah!!! 3.5 hours... geez, that is some compile time! (the A3000).

I think everyone is sitting on pins and needles waiting for MorphOS1.4 to make it's appearance. :-)

But to the original question, MorphOS (1.3) is very quick. I suspect 1.4 with the Trance JIT will scream. One can simply go to the AmiGOD website and confirm this. A G3 Pegasos and some 2+ Ghz Amithlon based PC have been banging it out for the top spot there. I even submitted my numbers to the site :-)

Targhan
08-02-2003, 10:24 AM
The AmiGOD 2 Benchmark Page (http://www.volny.cz/luky-amiga/Benchmarks.html) is a good place to look at system performance for Amiga(like) systems.

Have fun! :-D

redrumloa
08-02-2003, 10:30 AM
Yes! MOS1.3 on my 600Mhz Peggy is the fastest Amiga-like experience I have ever used, and this is before JIT. I used to own a 060/66-Voodoo3 based A3000, and it wasn't even close. I can only imagine what MOS1.4 with JIT will bring. I await not so patiently:-D

Kronos
08-02-2003, 11:15 AM
@seer

You don't really need native SW for that !

Lot of SW spends most of the time in OS-functions, and those are PPC-native
even if the SW is only 68k. GoldED/MorphED is a perfect example for that,
and feels faster than on Amithlon (AthonXP1600) :-D

downix
08-02-2003, 12:45 PM
Now the interesting part: as you can see, JIT massively contributes to the speed of UAE: JIT makes UAE compile seven times faster in my tests. If JIT for Pegasos (which will start shipping with MOS1.4?) has an effect anything like that, it is quite possible that a Pegasos at 600MHz will, with respect to emulated code, meet or exceed the performance of very fast PCs.
Part of the reason for this is because under UAE one is not just emulating the program, but all of the libs and modules used as well. Under MorphOS, all of those are running natively, so the only thing being emulated is the program itself. As a result, the performance jump for JIT will not be as dramatic on the Pegasos as on the Athlon, but it will still be dramatic.

Merko
08-02-2003, 12:54 PM
I would expect MorphOS with JIT to be faster than this, definitely.
Btw exactly how did you do the test? I could do it, but I'm not sure
what you mean. Decoding a single jpeg doesn't take enough time to
clock it.. I'm not familiar with Collector, maybe there's some batch
mode that you used. I could try it if you'd like some hard figures.

Cymric
08-02-2003, 02:01 PM
Kronos wrote:
... Lot of SW spends most of the time in OS-functions, and those are PPC-native even if the SW is only 68k. GoldED/MorphED is a perfect example for that, and feels faster than on Amithlon (AthonXP1600) :-D
I think that GoldED, like any other event-driven program, spends most of its time waiting for user input rather than in OS-functions. I also think that much of speed gain you feel is simply the fact of having a 600 MHz CPU at your disposal, with a flat Amiga-like memory space. (I'm not sure though, the feature list of MorphOS is inconclusive as to what kind of virtual memory/memory protection it offers. Can some kind soul please explain?) But it's definitely nice to see that old programs just breeze along. Now hope and pray that once people start developing seriously, and taking advantage of CPU speed, programs continues to breeze ;-).

IonDeluxe
08-02-2003, 02:16 PM
Morphos is exactly as fast as terminal velocity allows from my hand to the trash heap, which is to say about 3msec.

downix
08-02-2003, 02:32 PM
@cymric

the MOS VM system is not yet in the public release. When it is included, more detail will be provided.

Hammer
08-02-2003, 04:53 PM
KennyR wrote:
Believe me, MOS on Pegasos feels a lot faster than UAE on any PC, gigahertz CPU or not. Pure emulators just don't come close.

What about against a ~2.2Ghz Athlon XP, Radeon 9800 Pro, nForce 2 400 Ultra and 1Gb 400FSB DDR SDRAM (dual channel 512Mb x2)?

downix
08-02-2003, 05:00 PM
@Hammer

Quoting that you're paying a lot more for the PC and no guarantee you'll be faster.

Hammer
08-02-2003, 05:45 PM
Deleted double post…

Hammer
08-02-2003, 05:47 PM
@downix


Quoting that you're paying a lot more for the PC

That factor was not taken into an account in regards to Kenny's post...

Memory and video card factors are variable factors (these can change with minimal impact on the emulation speed) .

Athlon XP 2700+ (~2.16Ghz)($~126), 2x PC3200 256Mb ($~36 each) and Leadtek/Chaintech ($~59) nForce2 should be cheaper than $299 USD mark.


and no guarantee you'll be faster.

Just post some benchmarks…

KennyR
08-02-2003, 06:02 PM
hammer wrote:

What about against a ~2.2Ghz Athlon XP, Radeon 9800 Pro, nForce 2 400 Ultra and 1Gb 400FSB DDR SDRAM (dual channel 512Mb x2)?

If you buy me one I'll be sure to compare it fairly and honestly. ;-)

Oh, I didn't get a chance to thank you for finding my PC motherboard details earlier (the Quake2 thread). With your help I was able to bite the bullet and quickly pick a 512MB DIMM for the PC. It's much more responsive now (namely it no longer takes 3 minutes to load IE). Thanks!

Athlon
08-03-2003, 01:55 AM
I can't wait to build my Pegasos PC in September when it is supposed to be available in The U.S.A... I have heard that it boots up in five seconds.... :-D Oh yes when booting up Moph OS ! :-P

Dietmar
08-03-2003, 03:07 AM
>I think that GoldED, like any other event-driven program, spends most of its time waiting for user input rather than in OS-functions.

Doesn't help much. Event-driven applications may idle along most of the time but when user input arrives, it still has to be processed quickly. In the case of GoldED, scrolling is a demanding scenario: while a user scrolls through a C source code, each word is compared against 16000 known words before it is displayed. If spellchecking is on, it is additionally compared against megabytes of dictionary content. There are other CPU-banging functions on top of that. A fast computer is not wasted on it. A G3@600 can run it reasonably fast but it's not fast enough to avoid getting friendly with gcc in the long term (to have room for more CPU-banging features :)

Dan
08-04-2003, 04:19 AM
I would expect MorphOS with JIT to be faster than this, definitely. Btw exactly how did you do the test? I could do it, but I'm not sure what you mean. Decoding a single jpeg doesn't take enough time to clock it.. I'm not familiar with Collector, maybe there's some batch mode that you used. I could try it if you'd like some hard figures.
It´s a thumbnail program so I made a index from a directory with 33 jpgfiles. Also note that it is made for 256 color screens it´s only half speed at 24bit screens. The libaries it uses to decode jpg is available for ppc, i think.


Since the 600Mhz G3 is half the speed of UAEJIT then the G4 1GHz will be the same speed as UAEJIT?
When will MorphOS get JIT?

Cymric
08-04-2003, 05:38 AM
downix wrote:
the MOS VM system is not yet in the public release. When it is included, more detail will be provided.
Well, that's the not-yet-included VM subsystem---what does MorphOS do currently? Is it flat, or paged? Or perhaps a combination? (While it's too bad that there isn't any more information available, I respect that decision. There has been too much flaming going on regarding VM on the Amiga anyway.)

Cymric
08-04-2003, 05:48 AM
@Dietmar:

Yes, of course, that's the challenge of an editor: do as much as possible in as little time as you can. But as I understood the original statement is that it was claimed---or at least hinted at---that the efficiency of the OS (functions) was the root cause for GoldED's speed. While in fact it is simply the 600 MHz of computing power, as you so clearly explained. I must say that what you wrote down surprised me: a 600 MHz CPU not yet fast enough for an editor...? Wow. You did a helluva job on the original then, which ran very well on my A4000. Then again, I suspect the program has evolved considerably since I last used it.

Prod
08-04-2003, 06:27 AM
@Athlon
Hmm, what do you mean by 'Pegasos PC'? :-)
But if you are getting a real Pegasos1 then drool not to fast, althogh it is a fantastic piece of machinery that 5 second bootup time is a silly joke that should never have been mentioned in public. First time I booted my Pegasos there were PC guys in the room, how embarrasing for me, luckily it was still faster than an average PC so they did not notice something was not as it should have been. A typical Pegasos boot is around 28 seconds, but the OS itself will indeed boot in a very decent 2.6 seconds, the rest of the time is spent doing some nasty hardware tests or whatever, this should be fixed in Pegasos2. The 68k emulator in 1.3 is supposed to be an exact emulation of a 060, no slower or faster, but like Kronos said, programs spend a lot of time in OS-functions and those are PPC, even if it is a 68k program, and since this is all happening on modern hardware, especially the gfxcard, then everything goes much much much faster than an elderly classic Amiga. :-)

Dietmar
08-04-2003, 06:43 AM
>Since the 600Mhz G3 is half the speed of UAEJIT then the G4 1GHz will be the same speed as UAEJIT?

Question makes no sense ;) The G3 can not be "half speed of UAE JIT" because any speed difference obviously depends on the PC you throw at it, not just UAE's JIT: A G3-based Pegasos without JIT may very well be much faster than UAE/JIT. And vice versa. Faster or slower is very much a matter of Ghz. Also, please keep in mind that we are talking about how fast classic m68k code is executed, ie. "old software". This thread is not about the speed of the G3 but about the speed of the emulator. That said, some comments on speed:

I've previously posted some of my own timed tests. As far as I can tell from those, both emulators reach about the same speed on "similiar" hardware (that's of course a very loose, crude, questionable term). Compilation was used as benchmark because it primarily tests the emulation rather than the OS. The result is surprising because MorphOS does not yet have JIT. Given the effect JIT had on UAE - it made my UAE benchmarks seven times faster - a G3 with JIT should be a very fast platform for running classic software.

>When will MorphOS get JIT?

JIT has been announced for MorpHOS 1.4, making it a matter of days, I guess. As to the original question, I don't know what effect G4 will have on the emulator. 400 MHz isn't a massive difference and the CPU is only one of many components.

Dietmar
08-04-2003, 06:56 AM
> I must say that what you wrote down surprised me: a 600 MHz CPU not yet fast enough for an editor...?

There are editors and there are editors, it's all a matter of features. Old Amiga editors obviously were designed for the hardware at that time. They run fine with 8 MHz. Hardware gets better, features get better, CPU takes some hits. Compared to good developer-quality PC editors like Source Insight, GoldED is harmless in that regard: those editors compile code in the background (incremental compilation), analyze and tag source codes, build class trees and do a lot of other amazing things.

Btw, I didn't write that a G3/600 is not yet fast enough but that it runs GoldED (ie. emulated software) reasonably fast and that more feature would require a gcc port, ie. a native version.

Dan
08-04-2003, 11:33 AM
>Since the 600Mhz G3 is half the speed of UAEJIT then the G4 1GHz will be the same speed as UAEJIT? Question makes no sense ;) The G3 can not be "half speed of UAE JIT" because any speed difference obviously depends on the PC you throw at it, not just UAE's JIT: A G3-based Pegasos without JIT may very well be much faster than UAE/JIT. And vice versa. Faster or slower is very much a matter of Ghz. Also, please keep in mind that we are talking about how fast classic m68k code is executed, ie. "old software". This thread is not about the speed of the G3 but about the speed of the emulator. That said, some comments on speed:
Sorry, what I meant was that Morphos without JIT is equal in speed to UAE with JIT on a MHz for Mhz basis, meaning that a 1Ghz Pegasos would be as fast as my 1Ghz PC?

Targhan
08-04-2003, 12:51 PM
@Dan
Sorry, what I meant was that Morphos without JIT is equal in speed to UAE with JIT on a MHz for Mhz basis, meaning that a 1Ghz Pegasos would be as fast as my 1Ghz PC?

First, you must understand that speed is dependant on many things beyond just MHz. As a member of the amiga community you must realize this. A 40Mhz 040 based A4000 will smoke any 40Mhz 486 system. Unfortunately, in today's world this is further complicated beyond simple differences in the CPU.

My 600Mhz/G3 Pegasos smokes my 950Mhz Celeron PC for *most* tasks. When running Amithlon (w/JIT) against my non-JIT install of MorphOS (same PC, just amithlon instead of windows) the speeds are comparable in most instances, with one or the other getting the nod of better performance depending on what I'm doing. Unfortunately, my mileage may not match with your mileage.

The reason? Speed isn't about MHz anymore. You have memory speed, bus speed, harddrive speed, the graphics card speed, and all of that before getting into the CPU's cache sizes, which can make a world of difference depending on the application. (Try encoding a video on a 1ghz celeron vs a 1ghz PIII. Remember, they ARE the same CPU with different cache sizes.)

Thus, it becomes very difficult to make any blanket statements. With that said, UAE has to sit on a host OS (windows, linux, etc), so it has some drawbacks right out of the gate. However, in the end, it will still depend on who built the system, how careful they were with selecting parts, etc. I have seen 1Ghz PIII's outperform 1.7 Ghz P4's due to the configuration of the other parts in the system. In the end, my earlier statement stands. Visit the AmiGOD2 statistics site if you want to see test results.

:-)

KennyR
08-04-2003, 01:07 PM
A typical Pegasos boot is around 28 seconds, but the OS itself will indeed boot in a very decent 2.6 seconds, the rest of the time is spent doing some nasty hardware tests or whatever, this should be fixed in Pegasos2.

My Pegasos takes 12 seconds to boot, from the moment I turn it on until Ambient is sitting front of me ready to use. I hear that OF is delayed by some hardware combinations, such as not plugging in a PS2 mouse, so maybe you have that problem.

Actually when I first installed MOS it only took 8 or so, but the boot has been slightly delayed by 30 datatypes, USB stack, catweasel floppy dosdrivers, TCP/IP stack, and a huge background pic. :-)

Karlos
08-04-2003, 01:19 PM
I suppose the statement 'bloody fast!' is enough, eh?

@Kenny,

My OS 3.0 used to boot in about 20 seconds on my old 040 card from cold. It now takes 2x longer - there is a scsi startup delay on the blizzppc and a reboot as the rom update is installed.

It still boots in a fraction of the time it takes this sodding Win2K machine :lol:

Dietmar
08-04-2003, 03:37 PM
>Sorry, what I meant was that Morphos without JIT is equal in speed to UAE with JIT on a MHz for Mhz basis

Obviously not, MorphOS is faster because the OS runs directly on the CPU and does not need an emulator. UAE needs an emulator to run the OS. Big difference. But I guess/hope what you really mean is how the speed of emulators compare, ie. the speed at which "old" software is executed? If you mean the emulator in MorphOS, yes, in my experience, with my own PC, a crude MHz comparison vs a typical PCs (of the last year or two) seems to be a reasonable rule of thumb. While waiting for JIT, that is.

Dan
08-04-2003, 03:42 PM
Ok, if boot time is included a pegasos will always be faster than a pc. And with the morphos-jit things will get even faster. Then there is also the silence of the Pegasos. I´m almost convinced to buy a pegasos but does my favorite programs work?
I know PPaint and Photogenics works but does DPaint and PerfectPaint work?
And what about Visage (http://se.aminet.net/~aminet/dirs/aminet/gfx/show/Visage.lha), DiamondBOX (http://se.aminet.net/~aminet/dirs/aminet/gfx/edit/DiamondBOX.lha) and Collector (http://se.aminet.net/~aminet/dirs/aminet/gfx/misc/Collector33.lha) ?
And ArtPRO (http://se.aminet.net/~aminet/dirs/aminet/gfx/conv/ArtPRO1.20.lha) ?

Hammer
08-04-2003, 05:02 PM
KennyR wrote:
Oh, I didn't get a chance to thank you for finding my PC motherboard details earlier (the Quake2 thread). With your help I was able to bite the bullet and quickly pick a 512MB DIMM for the PC. It's much more responsive now (namely it no longer takes 3 minutes to load IE). Thanks!
No problems…

Just note that, a VIA KLE133(e.g. MSI-6378) will have to share its memory bandwidth with the GPU (in this case, an “Integrated Trident Blade 2D/3D video accelerator”) and the CPU.

From Sisoftware’s Sandra 2003, a VIA KLE133's memory throughput is around ~586Mb/s. A MSI-6330 V3.6 (VIA KT133A) has a memory throughput of ~898~910Mb/s (with BIOS V3.6).

PS; A VIA KT133A reference board has a ~1012MB/s (aggressive memory timings, BIOS's settings permitting).

To illustrate the impact speed penalty in regards to memory bandwidth refer to the following scenario.
Try 'MPEG 2 (DVD resolution and quality) to Divx 5.0(~DVD resolution and quality)' conversion (via Flask) on VIA KT133A (with 512Mb 133Mhz SDRAM) vs nForce II (with 512Mb 266Mhz DDR SDRAM) (both was installed with a similar CPU i.e. Athlon XP 1800+(~@1.53Ghz)). The VIA KT133A (MSI-6330) can only reach ~12~14 Fps, while nForce II reaches +26 Fps.

That performance gap widens farther IF the machine employs 333Mhz memory/FSB bandwidth. Of course an Athlon XP 2600+** (@~2.09Ghz) will help in that regard. **333Mhz FSB variant.

FSB and memory bandwidth may have some impact on emulation's speed.

Madgun68
08-05-2003, 03:44 AM
I know PPaint and Photogenics works but does DPaint and PerfectPaint work?
And what about Visage, DiamondBOX and Collector ?
And ArtPRO ?

DPaint won't work as far as I know. All versions are tied to the custom chipsets. Not sure about PerfectPaint, as I haven't tried that yet. As for the others you mentioned:

Visage: Yes. (Kinda slow though. The included MysticView is much faster.)
Collector: Yes
ArtPRO: Yes
DiamondBOX: Sort of. It ran but didn't seem to work correctly.

Dan
08-05-2003, 07:14 AM
Thanks Madgun.

Visage: Yes. (Kinda slow though. The included MysticView is much faster.)
Is it slow when opening the screen or loading the picture?
I run it from DOpus like this: "path/Visage wbmonitor {O}" to stop it from opening a new srceen evertime.
Anyway MysticView is Morphos native right? because its slow on a 060 an even on my uae.

KennyR
08-05-2003, 07:22 AM
Yes, Mysticview (and its libraries, guigfx and render) are fully native.

MysticView is incredibly fast on Pegasos, as is anything that uses datatypes (which are slow on Amiga systems, even emulated ones). If I use Picshow (on Aminet) on a dir full of 1024x768x32 pictures and hold down the downarrow key, the pictures load so fast one after the other that my screen becomes a blur of colour.

sdesros
08-05-2003, 07:56 AM
Actually DPaint V kinda works on Peg+MOS, as it did support RTG. I think lots of people forgot that but then again, the RTG modes were SOOOO SLOW, but it's not much of a problem now. :)

I haven't tested it out fully yet, but I think that the animation features are the only real problem. I think the animation player doesn't work fully in the RTG screen modes.

Madgun68
08-05-2003, 10:23 AM
Is it slow when opening the screen or loading the picture?
I run it from DOpus like this: "path/Visage wbmonitor {O}" to stop it from opening a new srceen evertime.Sorry, I should have elaborated a little more. Opening the screen is quick enough.. What's slow is rendering the pictures.

[Note to self: Don't post first thing after waking up in the morning]

Dan
08-06-2003, 01:49 AM
Sorry, I should have elaborated a little more. Opening the window is quick enough.. .
:-?
Window???
Visage always opens its own screen to display the pictures!
And there is an option to disable the internal 68k assembler optomized jpg loader and use datatypes instead but I haven´t bothered with it yet.

olegil
08-06-2003, 06:30 AM
KennyR wrote:
My Pegasos takes 12 seconds to boot, from the moment I turn it on until Ambient is sitting front of me ready to use. I hear that OF is delayed by some hardware combinations, such as not plugging in a PS2 mouse, so maybe you have that problem.


Heh, I just cut down the boot-time of my AmigaOne by about a minute. The problem? I had fiddled too much with the harddrive so two pins on the connector had been pushed back and was touching eachother on the PCB... IDEACK was tied to GND, so it was a miracle it even worked. Removed the faulty pin, and suddenly I'm in Linux after 15 seconds (Linux isn't impressive when it comes to boot time, I must say).

Still, this here RapidBIOS by Intel on the office computer hasn't started Lilo by the time my AmigaOne is in userland, so I can't complain :-)

woof
08-06-2003, 11:40 AM
Hi i am the author of Woof3D and i've
seen my own prog (a 3D 68020-040 library) run on
a beta of Morphos 1.4 and it's very fly
with the new JIT
(and after all it works !! and simply that astonished me :-))
Dont know if it is faster than a 68060
(never seen running on it) but i look
very comparable to the WinUAE JIT on
a more than 1Goz PC
So it's will be very coool on a Pegasos 2 G4

Alain Thellier - Paris -France
woof3d.make.you.free.fr

Mad-Matt
08-06-2003, 11:54 AM
visage can use datatypes by specifying the -datatype option on the commandline.

May as well use multiview if you want it in a window though.

Dan
08-06-2003, 12:06 PM
visage can use datatypes by specifying the -datatype option on the commandline. May as well use multiview if you want it in a window though.
Thats what I don´t understand, what WINDOW is he talking about ???
Doesn´t MorphOS allow the programs to open their own screens?

Kronos
08-06-2003, 12:08 PM
@Dan

Offcourse it does have screens .....

Dan
08-06-2003, 12:16 PM
Thanks for clearing that up Kronos.
I want to buy a Pegasos but which one should I get the Peg1 G4 or the Peg2 G4?

Madgun68
08-06-2003, 12:32 PM
Thats what I don´t understand, what WINDOW is he talking about ???
Doesn´t MorphOS allow the programs to open their own screens?It was a simple miscommunication between my brain and fingers. In other words, it was supposed to be SCREEN, which is why I edited my post.

Dan
08-06-2003, 12:35 PM
Sorry I didn´t saw that.

Merko
08-06-2003, 05:26 PM
I'll make a speedtest with Collector as soon as 1.4 is officially out.
We're not supposed to spread info about it before then.

Actually I think it would be a good idea to set up a little "test
suite", consisting of widely used Aminet programs, to make real-world
tests with. Things like AmiGOD are nice, but real programs are better
imo. This way we could compare 060 systems, Amithlon, MorphOS, and OS4
for that matter when that's out.

Preferrably, programs which are scriptable should be used, I guess. So
you can just download the stuff from Aminet and run the script,
watch the action and get the results.

downix
08-06-2003, 09:19 PM
I run many apps on their own screens.

Dietmar
08-09-2003, 07:40 AM
Below are the compilation times of a program (GoldED) compiled with SAS/C under UAE, UAE with JIT and MOS without JIT. Compilation is a reasonable test for raw performance. Integer performance obviously. The compiler is run with precompiled headers to take HD speed out of the equation.

[...] The PC@1.4 GHz has 512 MB DDR-RAM, 40 GB Maxtor HD and 160 GB RAID; Maxtor used for test. The Pegasos@600MHz has 256 MB SD-RAM and a 40 GB Barracuda IDE.

UAE /w JIT: 1: 59 min
UAE no JIT: 11:52 min
MorphOS 1.3 no JIT: 4:19 min


With MorphOS 1.4 released today, I'll use the opportunity to update those figures:

UAE /w JIT: 1: 59 min
UAE no JIT: 11:52 min
MorphOS 1.3 no JIT: 4:19 min
MorphOS 1.4 /w JIT: 1:05 min

MorphOS 1.4 is almost four times faster than the previous version in this test (which is a test about running CPU-intensive classic software, not about PPC software). It outperforms my 1.4 GHz UAE PC by a factor of two.

Karlos
08-09-2003, 05:18 PM
Just curious....

So, does the MOS JIT manage keep all the 68K register contents in PPC registers during the bulk of the execution or what?

No matter, it sure sounds impressive.

Targhan
08-10-2003, 12:09 AM
For the sake of comparison (and we knew this was coming), I am posting my AmiGOD2 results from MorphOS 1.3, and the new numbers from MorphOS 1.4 with JIT. (G3/600MHz, Radeon 7000)
----------------------------------

MorphOS 1.3
2D GFX: 50506
Math 68k: 184
Intuition: 811
Memory: 555

MorphOS 1.4
2D GFX: 52433
Math 68K: 645
Intuition: 2011
Memory: 625

--------------

Amiga 4060/50MHz, Piccaso II

2D GFX: 1794
Math68k: 54 (native)
Intuition: 119
Memory: 8

--------------

Your numbers may vary, but that's what I have on my MorphOS machine. The numbers from the 4000 come from the AmiGOD2 page (Don Cox's A4000 tests, number 17). My 1.3 numbers are on the same page (Number 4).