AmigaONE benchmark - Important

[Karlos]

Does any of this x86 vs PPC matter?
It always seems to come up.

What matters is, what is the A1 / Pegasos going to do for you?

Is your MorphOS / OS4 experience, compared to OS3.x running on your trusty 680x0, going to kick arse?

With a minimum G3 600MHz behind it, which lets face it is far more powerful than any existing Amiga hardware, what sort of apps can we look forward to?
Will the audiophiles get realtime soft synths?
Will the arty types get more pixel mutating power?
Will the gamers get to blow new stuff up?

All this fretting over what the x86 can do faster and vice versa is just silly.

Lets get smart, have fun and bash out some nice shiny new apps for these machines, rather than running synthetic benchmarks and whingeing ;-)

[BlackMonk]

You have it the opposite way round. x86 needs modern DDR-RAM for speed because of (bad!) legacy CPU design, while PPC is the opposite and is not sped up much by fast external RAM but is sped up greatly by extra cache. This is also why PPC is so good at cache-based calculation like RC5.

All CPUs need a fast interface to the rest of the system.  Bad legacy CPU design?  Uh, no.  x86 CPUs run at a relatively high speed compared to main system RAM and you want to keep the CPU fed with instructions.  Fast caches help with this problem.

Do you know why the G4 is sped up more by extra integrated cache as opposed to faster main system RAM?  It's simple--Motorola has crippled the main system bus to 166 MHz SDR.  So while the Athlon is at, what, 166 MHz DDR (333 MHz effective) and the P4 is a, what, 133 MHz QDR (533 MHz effective), the lowly G4 is forced to use 166 MHz or 133 MHz.

Already the vaunted Altivec units are hampered by being starved for instructions.  This has been a complaint by coders on the Mac platform for a while.  It's funny, the Mac has DDR DRAM but it only gets accessed through an SDR bus... the CPU is crippled.

That's the only reason that programs that can basically fit in cache (data and instructions for the critical loops) perform so well with the PPC... and heck, that's an ideal case for any CPU.

and it completely buries the Athlon per MHz. The 604e has a better FPU than both chips as well.

So?  I don't see any 1.2 GHz 604e CPUs laying around, who cares?  The CPU couldn't scale much higher and the manufacturers moved on.  The relative efficiency of a 280 MHz CPU compared to modern CPUs of different architectures, even, has little to no practical relevence.  I mean, a 68040 might have even better per-MHz FPU performance than even the new PPC970 but who would use that over the new CPU?  

Anyway, the whole "x86 needs more bandwidth because it's poorly-designed" is a false assertion.  ALL modern CPUs benefit from higher bandwidth to the main system memory.  With the G4 there is no option for increasing that speed which is why y'all should be happy that there's a huge L3 cache included.

I like PPC but there are some shortcomings that can't be ignored for the current CPUs.  That should all go away with the PPC970, a CPU I'm excited about.  

[BlackMonk]

BTW, I'm not trying to bash the G4 and say, "omg, x86 is better, ur so dum!"  Heck, I'd LOVE to have a dual-G4 system!  It's possible to like something but also be aware of its shortcomings.  

The Pentium 4 and the Athlons have their own shortcomings, as well, they just aren't really relevent to a discussion about overall G4 performance since it is more of a worst-condition exercise of the mind.  All I wanted to do was clear up some misconceptions and whatnot.

AMIGA all the way.
AMIGA rulez!

Anything else sucks.

[samo79]

Hei Ikir ma come faranno sti poveretti a leggere i topic italiani nei nostri siti ??  :-D  :-D ....

Mi sa che dobbiamo cominciare a tradurre le pagine! :-?  :-?

Ciao

Hello!  I am not much of a power user, nor do I consider myself a knowledgeable computer user, but I found this intersting bit of information on Digita's Organizer for PC, on an article entitled "Evolution of the CPU"  'The Pentium ...  'interestingly,  the awkward and old architecture is such a barrier to improvement, that most of the Pentium-compatible CPU's (NextGen Nx586, AMD K5, Cyrix M1, and even the Pentium's replacement, a 2-chip-14-stage CPU/Cache module named 'P6', do not even clone the Pentium, but emulate it with specialized hardware decoders which convert Pentium instructions to RISC-like instructions which are executed on specially-designed superscalar RISC cores, actually faster than the Pentium itself.  One rumour has it that IBM is developing hardware to translate Pentium instructions to run on a PowerPC CPU  called the 615.'  So maybe all the "modern" enhancements to x86 are just workarounds to the CPU's limitations.

[ikir]

samo79 wrote:
Hei Ikir ma come faranno sti poveretti a leggere i topic italiani nei nostri siti ??  :-D  :-D ....

Mi sa che dobbiamo cominciare a tradurre le pagine! :-?  :-?

Ciao

Troppo lavoro :-)

[Hammer]

AmigaPro wrote:
I think its interesting how poorly all of these chips do, both AmigaOne and x86, compared to the venerable old CS-PPC...

http://amigapro.com/Images/280.gif

The old CS-PPC (mine is slightly oc'd there, to 66/280MHz) is just under half as fast in OGR as the AmigaOne G4/800, and it completely buries the Athlon per MHz.  The 604e has a better FPU than both chips as well.

IPC(instruction per cycle) alone doesn’t rescue a CPU’s fortunes…  

One should be focuing on IPC x Mhz = Performance.
PS; Athlon XP with the Barton Core should improve the IPC side of things.  

[Hammer]

>Does any of this x86 vs PPC matter?

Well, somebody made a comparison with Pentium 4 2.4Ghz and X86 in general.

[Hammer]

Focal_Loco wrote:
Hello!  I am not much of a power user, nor do I consider myself a knowledgeable computer user, but I found this intersting bit of information on Digita's Organizer for PC, on an article entitled "Evolution of the CPU"  'The Pentium ...  'interestingly,  the awkward and old architecture is such a barrier to improvement, that most of the Pentium-compatible CPU's (NextGen Nx586, AMD K5, Cyrix M1, and even the Pentium's replacement, a 2-chip-14-stage CPU/Cache module named 'P6', do not even clone the Pentium, but emulate it with specialized hardware decoders which convert Pentium instructions to RISC-like instructions which are executed on specially-designed superscalar RISC cores, actually faster than the Pentium itself.  One rumour has it that IBM is developing hardware to translate Pentium instructions to run on a PowerPC CPU  called the 615.'  So maybe all the "modern" enhancements to x86 are just workarounds to the CPU's limitations.

The K7 and K8 families does that trick (i.e. breaking down the X86 instructions into RISC like instructions before feeding into the main pipeline stages).

IBM's PPC 970 also has decoding/crush stage  for PPC instructions.

[LaBodilsen]

The old CS-PPC (mine is slightly oc'd there, to 66/280MHz) is just under half as fast in OGR as the AmigaOne G4/ 800, and it completely buries the Athlon per MHz. The 604e has a better FPU than both chips as well.

What matters in the end is that the G4 800mhz, is 2.6 times faster than your 604 280mhz.    

better IPC dont always equal better overall performence.

[Atheist]

Karlos wrote:
All this fretting over what the x86 can do faster and vice versa is just silly.

Lets get smart, have fun and bash out some nice shiny new apps for these machines, rather than running synthetic benchmarks and whingeing ;-)

Maybe it's just me, but this makes a lot of sense!

Anyhow, the new AOS4.0 kernal is going to be bigger, because it has more capability AND more instructions, due to being RISC, instead of CISC, So, we really have NO idea of real world performance. Even that benchmark doesn't necessarily reveal anything, because it was using Linux after all. BUT, RC5-72, at a binary level, is probably almost the same in Linux or AOS4.0. Only, the external actions of the OS's determine how many cycles are taken away from the OGR's or RC5s calculated. So we wait.....

Amiga! "Last one out, turn off the monitor," says Amy, the mascot.

[ksk]

This should be interesting:
http://www.math.wm.edu/~cahunter/NASA_G4_Study.pdf

[Agafaster]

Gobsmacked !  :-o

I did wonder when I saw the indifferent scalar results, but when I saw the Altivec enhanced results, me jaw dropped !!
I am now even more excited about my G4-800 now !
I expect 3D-games will bloody well FLY on AmigaOne now ! (when it is an amiga that is - ie: OS4 on AmigaOne) I imagine any G4 equipped Pegasos'll fly too, but thats a little OT here ! ;-)

I cant wait to run a Mandelbrot generator on it ! (small hobby of mine :-P )

[Karlos]

Sorry to be a downer, but not everything can be Altivec optimised.
OTOH, software 3D geometry calculation may be a good canditade, but remember that a good 3D hardware geometry accelerator will probably beat Altivec hands down for this kind of stuff.
Hopefully, Warp3D Nova will address this and we can all look forward to highly accelerated games.

Personally, I'm quite excited at the thought of Altivec optimised softsynths / streaming codecs ;-)

[Helgis75]

AmigaOne really kicks asses!!! :-)  :-)  :-D  :-D  :-o  :-o