Amiga.org
Amiga computer related discussion => Amiga Hardware Issues and discussion => Topic started by: takemehomegrandma on February 14, 2012, 10:27:50 PM
-
@Iggy
(A break-out from the X1000 dnetc benchmark thread)
@ Iggy
A Cortex-A8 (Efika MX) beats a Sam440.
A Cortex-A9 beats a G4.
You like long winded posts, but you never offer any proof for your statements.
I'd really like to see a benchmark on the second one since it seems implausible.
"A Cortex-A8 (Efika MX) beats a Sam440"
PPC440EP (http://myapm.apm.com/MyAMCC/retrieveDocument/PowerPC/440EP/PPC440EP_PB2001.pdf): 2.0 DMIPS/MHz, meaning 1,334 DMIPS @ 667MHz of a Sam440ep
Cortex-A8 (http://arm.com/products/processors/cortex-a/index.php): 2.0 DMIPS/MHz, meaning 1,600 DMIPS @ 800MHz of an Efika MX
NBench (http://www.morphzone.org/modules/newbb_plus/viewtopic.php?forum=3&topic_id=7675&viewmode=flat&sortorder=0&start=164) pictures the Efika MX as generally faster, even after upscaling the Sam results to 800MHz. In addition to this, the CPU used in the Efika MX has several HW accellerators that will offload and speed up many multi-media tasks from the CPU, like audio/video encoding/decoding, OpenVG, it has a SIMD (like Altivec, but not as powerful) called Neon, etc. None of this is present in the PPC440EP, which doesn't even have a L2 cache (the Efika MX CPU has 256KB). For example, the Efika MX can decode and display 720p HD video streams (it can actually decode 1080p streams, but not display it at 1080p resolution, due to some bottle necks that was removed in the i.MX53 version). The Efika is the winner.
Oh, and for the sake of comparison:
Efika MX: $129, giving you a complete, fully working system for most uses (but yes, rather limited and not expandable)
Sam440: ~$600, giving you a *motherboard only*, system components of choice *on top* of this, plus the mandatory OS4, all in all "north of" $1,000.
VAT not included.
"A Cortex-A9 beats a G4"
MPC7447A (http://www.freescale.com/files/32bit/doc/data_sheet/MPC7447AEC.pdf?fpsp=1): 2.3 DMIPS/MHz:
This should mean about 3,266 DMIPS @ 1,420 MHz (a *fast* G4, used for example in a Common Mac Mini)
Cortex-A9 (http://arm.com/products/processors/cortex-a/cortex-a9.php): 2.5 DMIPS/MHz (*per core*, can have 1-4 cores):
A *generic* Cortex-A9 from ARM (used in Apple A5 and Tegra 2 for instance) delivers 2,500 DMIPS per core @ 1,000 MHz, meaning 5,000 DMIPS per CPU in dual core configuration.
Then there is the upcoming (summer 2012?) Qualcomm Krait, that isn't simply just using the generic core, but optimized and tuned to improve the per-core performance to 3.3 DMIPS/MHz (built using 28nm process), resulting in a staggering 9,900 DMIPS @ 1,500 MHz.
The Tegra 3 is a *quad-core* Cortex-A9, running up to 1.300 MHz (times 2.5 DMIPS/MHz, times 4 cores). It's "up to 3 times faster than Tegra 2 (http://en.wikipedia.org/wiki/Tegra#Tegra_3_.28Kal-El.29_series) (measured with GLBenchmark 2.0 Egypt)", still being Cortex-A9, still being built in a 40nm process. Actually, the Tegra 3 is a "five core" CPU: "The Tegra 3 is the world's first variable symmetric multiprocessing processor (vSMP). Variable SMP is the brain that makes the 4 main + 1 companion core setup tick. Basically, it smartly distributes the workload so the platform uses four cores humming at up to 1.3GHz when needed (think games), but switches to two for lower loads (think browsing with Flash), and goes down to one for casual activities (browsing, no Flash). Finally, it can also use only the companion core running from 0 to 500MHz for active stand-by, video and music." (Link: An Interesting Article on the Tegra 3 (http://www.phonearena.com/news/NVIDIAs-quad-core-Tegra-3-the-more-the-better_id23752))
The scaled NBench numbers (http://www.morphzone.org/modules/newbb_plus/viewtopic.php?forum=3&topic_id=7675&viewmode=flat&sortorder=0&start=165) (measuring *one* of the cores on an "old" OMAP4 based 1GHz PandaBoard) suggests that the Cortex-A9 is about on par with the G4, but again, much has happened to the Cortex-A9 CPU's since that one was made, and since most Cortex-A9 based CPU's come in dual core configuration (or quad), I actually feel my statement that Cortex-A9 beats a G4 is water proof. In addition to this, the Cortex-A9 processors (like most other ARM processors) also come with various HW accelleartors on the chip, freeing the core CPU from the most common heavy multi-media tasks, vector graphics/GUI/Flash, hardware Java/JIT acelleration, etc, things that the G4 simply don't have.
And while we are at it: Some speculations about Cortex-A15 and G5
While not being a 64-bit CPU (although it will at least be able to address up to 1TB of memory thanks to the 40-bit Large Physical Address Extensions (LPAE) addressing), the Cortex-A15 will probably be giving the old G5 a match, with its 3.5 DMIPS/MHz (http://www.morphzone.org/modules/newbb_plus/viewtopic.php?forum=3&topic_id=7675&viewmode=flat&sortorder=0&start=211) in configurations suitable both for mobile, high-end home entertainment, or low-power servers, up to quad core running at 2.5GHz (or for servers for example, octo-core or larger configurations through multiple coherent SMP processor clusters through "AMBA 4").
For example the announced dual core: "ST-Ericsson Nova A9600 (Dual Core @ 2.5 GHz Over 20K DMIPS (http://en.wikipedia.org/wiki/ARM_Cortex-A15_MPCore))"
Iggy, this is a quote from a post you made here on amiga.org back in 2010 (http://www.amiga.org/forums/archive/index.php/t-53802.html):
e300: 1.9
PPC440: 2.0
PPC460: 2.0
Titan: 2.0 (presumably PPC450 based (http://www.morphzone.org/modules/newbb_plus/viewtopic.php?topic_id=7186&forum=11&start=20#74317))
PA6T: 2.2
PPC750: 2.3
e600: 2.3
PPC470: 2.3 or 2.5 (varying with information source)
e500mc: 2.5
PPC970: 2.9
e5500: 3.0
As a reference to year 2012 ARM (http://www.morphzone.org/modules/newbb_plus/viewtopic.php?forum=3&topic_id=7675&viewmode=flat&sortorder=0&start=211) (some in quad core with clock frequencies up to 2.0-2.5GHz):
ARM Cortex-A7: 1.9
ARM Cortex-A8: 2.0
Qualcomm Scorpion: 2.1
Marvell Sheeva PJ4: 2.4
ARM Cortex-A9: 2.5
Marvell Sheeva PJ4B: 2.6
Qualcomm Krait: 3.3
ARM Cortex-A15: 3.5
This is current or immidiate future (like this year), not even mentioning the not-too-far-away 64-bit "x86-killers" from nVidia and others, based on ARMv8.
-
Be interesting to see how newer PPC processors stack up - certainly there's more memory bandwidth to go around than with the G4s, which never got past 166MHz (not even DDR-capable.)
-
Interesting numbers, a good read, Takemehomegrandma. Cheers for posting it.
As for why things haven't migrated to a more modern platform, God only knows. Niche market, no financial reward for the amount of work, who knows.
Some people say if they wanted a x86 or other modern chipped Amiga, AROS is there. A valid point. I've often griped about HW issues with AROS, but the truth is it runs very well on supported HW and that HW list is getting bigger day by day.
The MOS camp have the luxury of the OS running on extremely cheap, readily available PPC Mac's. This has served them very well, and the OS is great, the old PPC Mac's are still easy to find. It won't be that way forever, though - and I imagine they are looking towards more modern platforms for future releases.
AmigaOS? God only knows why they have stuck with it (PPC). My only guess is because PPC is what OS4 runs on now, and perhaps there's simply no manpower or financial rewards to port the whole OS over to ARM or something for them. Then again, the first thing I'd do with OS4 is make it run on PPC Mac's if I ran the show there atm. It's a niche market, and OS4 limited to expensive PPC embedded boards is a true shame.
-
@takemehomegrandma
And now to the *actual point* of my initial post:
No, I'm not necessarily advocating ARM as *the* path to glory for *miga OS's. While ARM without doubt have its significant benefits, and this architecture indeed has a lot of focus from the whole industry, there are some issues as well. Mainly (IMHO): "Consumer Electronic Devices", i.e. not having a broad base for general, open "desktop" systems is probably the main one. Genesi and Pandaboard are two, are there more? Maybe it will be? Genesi is now working on producing a Cortex-A9 design based on the Freescale i.MX61 though, and may very well continue developing new systems whenever Freescale releases a new CPU...
But I wouldn't object to a x86 MorphOS system either! No, no! :lol: That's not the point I'm trying to make! ;)
Some people say that one of the most important things of the X1000 is that it is dual core, and thanks to this, SMP can be incorporated in the *miga OS from Hyperion. (The PowerMac G4 has dual CPU configurations as well BTW)
But here is the point:
Judging from what people have requested over the last decade, and also judging from what seems to be the ambition from some OS developers, "Moving Forward" may at some point mean the incorporation of some of the "modern" OS features, like true SMP, true MP, 64-bit, etc. And since this *will* require a clean-slate break from the Amiga legacy anyway (it must happen if you decide to go there), with a clearly defined border line marking the "before" and "after", the seemless "Amiga compatibility" scrapped post that line, starting anew, I must ask the question:
At that point, why continue the PPC path?
If you are to break the "Amiga" anyway, why not do it on some other architecture. My point is that even ARM seems to beat PPC. Performance wise, and from a desktop Point of View, x86 is even more attractive.
Isn't a platform migration the natural thing to do at that point? There is no "Power" in PowerPC, not in the year 2012 and beyond! ARM and/or x86, but not PPC!
Right?
:)
-
Be interesting to see how newer PPC processors stack up - certainly there's more memory bandwidth to go around than with the G4s, which never got past 166MHz (not even DDR-capable.)
AFAIK, the MPC8610 (the last of the "G4"?) from Freescale had an on-chip DDR/DDR2 SDRAM memory cotroller @ 533 MHz (1066 DDR)...?
-
AFAIK, the MPC8610 (the last of the "G4"?) from Freescale had an on-chip DDR/DDR2 SDRAM memory cotroller @ 533 MHz (1066 DDR)...?
You might be right, though either way it never saw use in Mac or Amiga-like systems...
-
You might be right, though either way it never saw use in Mac or Amiga-like systems...
Exactly like...
"Be interesting to see how newer PPC processors stack up"
...then? ;)
-
THGM
While ppc is not the saviour of us all it has the benefit that MorphOS/AOS4 runs on it already. That is the *only* benefit of ppc over other architectures that holds some significance.
Hence the real question is not "ppc or ARM?", but "ppc or not ppc?". And in case of not ppc - why on earth should you choose ARM over x86 which is _for GP computers (aka PC)_ factors more widespread and very powerful and will not vanish in the foreseeable future.
ARM holds _zero_ significance fo Amiga today. There are virtually no powerful general purpose computers based on ARM available, but a gazillion of x86 pcs.
-
THGM
While ppc is not the saviour of us all it has the benefit that MorphOS/AOS4 runs on it already. That is the *only* benefit of ppc over other architectures that holds some significance.
Hence the real question is not "ppc or ARM?", but "ppc or not ppc?". And in case of not ppc - why on earth should you choose ARM over x86 which is _for GP computers (aka PC)_ factors more widespread and very powerful and will not vanish in the foreseeable future.
ARM holds _zero_ significance fo Amiga today. There are virtually no powerful general purpose computers based on ARM available, but a gazillion of x86 pcs.
Please zylesea! I guess you didn't read post #4 (http://www.amiga.org/forums/showpost.php?p=680532&postcount=4)? ;)
Again, I wasn't really trying to promote ARM in front of x86! I would actually prefer a x86 port myself! :) The initial post was merely a "prelude"! ;)
Rather it was a question to those who (with the advent of X1000 etc) begin to look for features like SMP, MP, 64-bit, etc, and the question I was trying to rise was: "Since that would require a break from the past, why continue with PPC at all? Why not bring the new platform onto a modern architecture?"
For example, if MorphOS developers would want to start exploring 64-bit support for the future (as suggested by Fab in his presentation of MorphOS future (http://www.morphzone.org/modules/newbb_plus/viewtopic.php?forum=9&topic_id=8121&viewmode=flat&sortorder=0&start=27)), why would they do that on a *dead* G5 platform (or PA6T for that matter)? I mean, if the legacy is to be broken anyway, why not migrate to greener pastures while you are at it?
And indeed, there are many reasons to choose x86 over ARM...
:)
-
I would think ARM is attractive in the sense that it is getting tiny and very efficient compared to x86. I am not too sure what modern PPC platforms or chips you are referring to. Is there a link somewhere to cheap modern PPC chips?
-
Since getting PPC performance up to snuff requires compilers with cache hints and other tricks, I see no reason to continue the PPC path. It was a temporary kludge.
ARM is the cuts a lot of cruft out of other instruction sets including the x86. Though the AMD64 has better developed software that Amiga cannot use, ARM looks more promising.
Going back to 68k is more promising, for that matter. Since the N68050 is slated to contain some of the same features as modern ARM processors, it just is a matter of finishing it up and getting it on the market so we can work out more bugs and get it mass-produced someday. If the N68070 cannot beat the performance of a 68060 per clock cycle, it might make sense to switch to AROS for ARM.
I look forward to the Raspberry Pi, even if it doesn't run much Amiga software.
-
If ARM is the CPU to go around, then use it? The original Amiga idea was to use the best available technology for a given (low) price. That in turn makes it possible to grab a huge market section, which is why there is still Amiga fans around. The other is just that "the best" choice you can make for hardware within a given budget. Which will attract persons that appreciate sound technology choices.
Next objective, which graphics chip will give the best bang for the buck?
And will this translate into the best low price multimedia platform, which seems to been the nische for Amiga ..?
-
I have 2 question,
1 is ARM better price/mips than the cheapest i7?
2 is there an ARM CPU for sale today as powerful as the fastest i7
If the answer is not yes to both then OS4 needs to go x86 and be optimised for i7 multiple cores and hyperthreading IMO.
My $0.02 worth.
-
Very good arguments, takemehomegrandma.
I don't really disagree that ARM is an attractive option.
I'd just like to see MorphOS on a G5 before the transition to ARM or X86 begins.
Moving to another ISA will take a considerable amount of time and effort.
A G5 port would be much easier/quicker and would give us some more powerful equipment (available at a low cost).
Oh, and the DMIPS figures are from Andreas Wolf.
-
You answered your own question I think as to why AMIGAOS4.x will never be ported to x86 OR ARM.
The amiga grave robbers LIKE robbing you for hardware that has specs my 50$ phone beats. Why would they want to port it to something inexpensive when people are paying ridiculous prices for hardware thats so ****ty?
They will NEVER port it to an inexpensive platform while people are paying such high prices for such old spec junk.
Steven
-
Is the most powerful G4 the 1.67Ghz from the Powerbook G4?
-
Is the most powerful G4 the 1.67Ghz from the Powerbook G4?
The fastest G4 is the 7448 dualcore cpu with a frequency of 2Ghz.
I would prefer to see some real tests instead of "Meaningless Information Processor Speed".
Show me some ARM machine I can buy that is faster than a Powermac G5 QuadCore from 7 years ago. If there are no faster ARM machines available in the mass market then ARM is 7 years behind PPC.
Those complains about cpus usually come from ex-amiga users or people who doesn't use amiga at all. My powerbook G4 at 1.67Ghz runs MorphOS nicely and MorphOS would fly on a 2.5-2.7Ghz G5.
If MorphOS had to switch architecture in 6 years I would always choose x86, there are millions of computers available and all them are way faster than anything ARM that exists.
-
(next) new arcitecture without new software .. why ?
-
Show me some ARM machine I can buy that is faster than a Powermac G5 QuadCore from 7 years ago. If there are no faster ARM machines available in the mass market then ARM is 7 years behind PPC.
Yeah, this is really the crux of the issue: there just isn't any good ARM platform out there, not for general-purpose computing. Don't get me wrong, I'm looking forward to the Raspberry Pi and everything, but I've spent the last two years watching semiconductor manufacturers pour quadzillions of dollars into ever-better ARM cores for tablets and smartphones and screaming "Give it to me in a laptop already!" at my screen. (And no, tablet + keyboard is not the same thing. Give me something I can put 4GB RAM and a 320GB hard drive in, damn it.) Even the couple attempts at an ARM netbook I've seen are based on old cores that don't come close to matching my old, janky P4 laptop from 2004 for horsepower, let alone my trusty Eee.
Those complains about cpus usually come from ex-amiga users or people who doesn't use amiga at all. My powerbook G4 at 1.67Ghz runs MorphOS nicely and MorphOS would fly on a 2.5-2.7Ghz G5.
Quite. My PowerBook/Linux attempt may have been abortive, but one thing I did find is that a 1.67GHz PowerBook is perfectly satisfactory for everything except streaming video; if there were one based on a newer chip with a faster FSB I can't imagine I'd have any complaints at all.
-
What graphics should go with it?
-
What graphics should go with it?
If I'm right, just about anything (but the MOS team would probably prefer an ATI solution - me too).
BTW - Thank you all for the support on retaining PPCs (for the time being).
I can get a G5 mac cheap. They're fast and still fairly up to date.
AND, they'll run my current software without recompilation or emulation.
ARM or X86 eventually?
Why not?
But, for the time being, why not continue follow the course which has brought a lot of new users to MorphOS?
-
I see a couple of pros and cons.
It takes time to move to a new platform. Look at how hard it is to get drivers for existing systems, even 68k. As good as they are at our specific platform, our OS guys are slow compared to Linux, BSD or even Haiku teams.
It's not easy to emulate PPC yet, PPC isn't quite that outdated but it's getting there, so that software would probably be a loss unless it is recompiled. Many of these are either ports already or still actively developed, so it might not be terrible. Many users never moved to PPC, so it doesn't harm 100% of the userbase.
That said...
When you see that the platform is essentially dead, you've got to move asap or get caught without hardware available. PPC has been done for since Apple moved to x86 whether we want to admit it or not.
Personally I'm tired of scrounging for old hardware and always being a decade or more behind the performance curve just because I like a certain OS.
Any current platform would be a plus.
-
What graphics should go with it?
The Raspberry Pi has a SoC so it comes with the OpenGL-ES graphics core on the same chip as the CPU. Since OpenGL-ES can support fixed-point maths, it might be tricky getting some Mesa support since that requires floating-point maths. We'll see how it goes.
-
What graphics should go with it?
Exactly, why add yet an other architecture, on an already small and divided market?
-
Exactly, why add yet an other architecture, on an already small and divided market?
In a word: Cost!
The original post of this thread reveals that ARM based netbook can plow the streets with a SAM 460 yet cost much less.
Personally, I'm holding out for the NatAmi MX and the Raspberry Pi. I've got a Micro-A1c but I'm not going to shell out that kind of money ($1400 for a complete system) again.
-
@CommodoreJohn
There was BBRV's blog post about a new Efika MX model (http://bbrv.blogspot.com/2011/12/holiday-surprise.html) coming sometime this year. It might not be so powerful as what you're looking for, but it should be a bit faster than the existing Efika MX.
-
There was BBRV's blog post about a new Efika MX model (http://bbrv.blogspot.com/2011/12/holiday-surprise.html) coming sometime this year. It might not be so powerful as what you're looking for, but it should be a bit faster than the existing Efika MX.
Huh, that's definitely an improvement over the current model. If they'd go with one of the multi-core variants they might actually be getting somewhere...
Of course, there's still the problem that the Efika has zero RAM/storage expandability.
-
...In a word: Cost!
Which is exactly why I've argued for support of more PPC Mac models (like the G5).
They are extremely inexpensive and still offer competitive performance (doubt me? go look at the X1000 benches again).
And now that I think about it, the Natami is going to be rather expensive (not to mention not that great a performer - FPGAs have there limits).
The 68K had its day and that day is gone.
-
Your problem is that this question is simply wrong.
PowerPC is a technology that we Amiga users use in our Amiga 15 years.
15 years of history, 15 years of memories.
PowerPC is a technology that we know, technology that we love.
For many of us, Amiga users only Amiga with PowerPC is a true Amiga.
68k is (was) ridiculously slow.
And a PC with x86 will never be Amiga.
You should think for themselves.
We Amiga users will not resign from 15 years of history for peanuts.
-
PPC history? don't make me laugh.
-
PPC on the Amiga was a cobbled together mess in an attempt to keep up and nothing more, even when it was brand new in the 90's.
It was the only option back in the day that we had, and we were sunk long before it came out. Just because it's what we are used to doesn't make it something to admire, and I say that as a person who's main Amiga is strictly PPC with no 68k onboard.
PPC is going nowhere.
-
@takemehomegrandma
not having a broad base for general, open "desktop" systems is probably the main one. Genesi and Pandaboard are two, are there more?
Yes, several:
http://www.linaro.org/low-cost-development-boards
..and of course the very inexpensive Raspberry Pi.
-
PPC on the Amiga was a cobbled together mess in an attempt to keep up and nothing more, even when it was brand new in the 90's.
It was the only option back in the day that we had, and we were sunk long before it came out. Just because it's what we are used to doesn't make it something to admire, and I say that as a person who's main Amiga is strictly PPC with no 68k onboard.
PPC is going nowhere.
Agreed.
AmigaOS has painted itself into a corner, obviously going PPC back in the 90's was a mistake, if they'd just hung on to 68k for a few more years not only would have OS 3.9 been a better OS but it would have become clear that PPC was not the solution.
Anyway the problem now of course is that so much software exists for PPC it will be hard to switch to any other CPU. Just think what a mess a new OS would be if it had to intergrate emulation of both PPC and 68k...yuk.
I guess the only real option would be only support PPC emulation on an OS level and let some 3rd party application like WinUAE take care of the older 68k stuff which would be workable but still a bit of a mess :)
-
No need to emulate PPC, the source code for anything worthwhile will still exist (and will be in C, as no one in their right mind coded in PPC ASM) and can be recompiled for x86/ARM etc :)
-
Your problem is that this question is simply wrong.
PowerPC is a technology that we Amiga users use in our Amiga 15 years.
15 years of history, 15 years of memories.
PowerPC is a technology that we know, technology that we love.
For many of us, Amiga users only Amiga with PowerPC is a true Amiga.
68k is (was) ridiculously slow.
And a PC with x86 will never be Amiga.
You should think for themselves.
We Amiga users will not resign from 15 years of history for peanuts.
I'm not sure I agree about the 68K as I have a lot of fond memories from the '80s and '90s (when Motorola processors were so much better then Intel's offerings it was painful).
But RISC was and is still a better solution then CISC (like the X86).
I hope all of you that are advocating ARM realize that ARM and PPC have far more in common then either does with X86.
And PPCs are still more advanced then most ARM processor (although ARM is catching up quickly).
Personally, I'd just like to see Amiga (and Amiga like OS') stay away from X86. RISC was and is the way to go.
MorphOS>G5>ARM
-
Source may exist, but isn't available :p
But forget the hardware, think about the concept. Best tech solution for a budget that a mass market can afford.
It should make a difference too.. like an iPhone ;)
-
There is no "Power" in PowerPC, not in the year 2012 and beyond! ARM and/or x86, but not PPC!
Really?
http://cache.freescale.com/files/32bit/doc/fact_sheet/T4240T4160FS.pdf?fpsp=1
12 cores, 1.8 GHz, 6.0 DMIPS/MHz per core not good enough for you?
-
Really?
http://cache.freescale.com/files/32bit/doc/fact_sheet/T4240T4160FS.pdf?fpsp=1
12 cores, 1.8 GHz, 6.0 DMIPS/MHz per core not good enough for you?
But you don't get it, mongo! PowerPC is dead! Because they say it, it must be so!
-
From TFA:
T4240 Markets and Applications
• Service provider networking: RNC,
metro networking, gateway, core/edge
router
• Enterprise equipment: Router, switch
services, UTM
• Data centers: ADC, WOC, UTM, proxy,
server appliance
• Storage controllers: FCoE bridging,
iSCSI controller, SAN controller
• Aerospace, defense and government:
Radar imaging, ruggedized network
appliance
• Industrial computing: Single board
computers
-
...Single board computers
Yes, that would do nicely.
But, I'd still prefer the Freescale T5040.
Only four cores, but up to 2.5 GHz.
Not as many PCIe lanes as the PA6T though.
Yeah, PPC must be dead.
Freescale and Applied Micro keep introducing new ones and IBM is even introducing PPCs like the one they just designed for Nintendo that use features from the Power8 family.
Frankly, I just wish Microsoft had allowed third parties to use the Xenon.
-
But RISC was and is still a better solution then CISC (like the X86).
I hope all of you that are advocating ARM realize that ARM and PPC have far more in common then either does with X86.
Modern ARM with Thumb2, x86 and 68060 are a CISC decoder with RISC core. PowerPC and the original ARM are RISC decoder with RISC core. It looks to me like CISC cores are all but dead but so are RISC decoders.
And PPCs are still more advanced then most ARM processor (although ARM is catching up quickly).
PPC has some powerful features but they are also difficult to program. It's easier to get more power out of a CPU that is easy to program. ARM (with Thumb2) isn't bad but N68k is easier and I think will have better code density.
-
1 is ARM better price/mips than the cheapest i7?
ARM's main strengths are efficiency (performance/watt), and price.
Also the ARM business model, which builds on others licensing the core technology for inclusion in their own CPU design is an enabler. Hence we see lots of different CPU's, coming from numerous manufacturers, that includes various accelerator technologies and hardware controllers inside the very CPU (including graphics and video/audio decoding/encoding, etc). ARM CPU's are generally complete "Systems on Chips" that can do what most people expects from a netbook or tablet.
2 is there an ARM CPU for sale today as powerful as the fastest i7
No, no, not by far! :lol:
But it's still quite capable for the "every-day tasks" that most people use their computers for. And it's cheap! :)
I bet that 95% of the Core-i3, i5 and i7 users (i.e. the "common people" (not those really into computing, if you know what I mean), more or less clueless, buying their pre-built systems off the shelves at MediaMarkt or from Dell) spends *most* of their time using Office, Youtube, Facebook, e-mailing, watching holiday photos and videos, video conferencing, chatting, playing music, and watching DVD's and 1080p movies, etc (power gaming, rendering, compiling of huge projects like Linux distributions, etc are excluded), and then current ARM chips works just fine.
But for raw performance, I guess the *current* ARM CPU's would be more comparable to "Atom" kind of x86 CPU's rather than the desktop CPU's...
-
No need to emulate PPC, the source code for anything worthwhile will still exist (and will be in C, as no one in their right mind coded in PPC ASM) and can be recompiled for x86/ARM etc :)
Exactly my point! :)
Some people say that a dual core X1000 (or indeed a Power Mac with dual G4's) will make it possible to develop SMP for Amiga, but if you aim to do that, to develop "modern" features, like real SMP, real Memory Protection, 64-bit with increased addressable memory limit (which would also make "Virtual Memory"/swap-disk meaningful), it will *require* a break-up from the past, it will require a fresh start with a clean slate. And if you are to do that, why on earth continue the PPC path?
And as you said, no need to emulate the PPC. Most of the essential stuff is still "alive" and can be ported/recompiled. It would be a bit like AROS on x86 in that sense, you won't be able to run old 68k Amiga applications as seamless as you do in MorphOS today (would require UAE), but many of the MorphOS native applications could still be made available in a new version. Heck, with MorphOS 3.x, many things will even come bundled with the OS itself, including CD/DVD burning SW, FTP, etc.
-
Really?
http://cache.freescale.com/files/32bit/doc/fact_sheet/T4240T4160FS.pdf?fpsp=1
But you don't get it, mongo! PowerPC is dead! Because they say it, it must be so!
Yes really.
Nobody claims PPC is dead for routers, switches, printers, and similar embedded applications (where CPU's like this one will do just fine, I'm sure). But nobody is developing PPC CPU's for laptop/desktop usage, that stopped 5-6 years ago (whenever it was that Apple went x86) and more importantly, nobody is making viable laptop/desktop motherboards or systems based on PPC CPU's!
It's dead Jim!
-
Really?
http://cache.freescale.com/files/32bit/doc/fact_sheet/T4240T4160FS.pdf?fpsp=1
12 cores, 1.8 GHz, 6.0 DMIPS/MHz per core not good enough for you?
For embedded system, it's great for networking. As far as going head to head with the A15, I have a feeling it's going to come up way short. So what are the prices on the T4240?
-
Yes really.
Nobody claims PPC is dead for routers, switches, printers, and similar embedded applications (where CPU's like this one will do just fine, I'm sure). But nobody is developing PPC CPU's for laptop/desktop usage, that stopped 5-6 years ago (whenever it was that Apple went x86) and more importantly, nobody is making viable laptop/desktop motherboards or systems based on PPC CPU's!
It's dead Jim!
Not to mention that some PPC makers will be introducing very capable ARM designs.
I don't think that PPC will go completely dead that soon, but the writing is on the wall, it seems.
-
For embedded system, it's great for networking. As far as going head to head with the A15, I have a feeling it's going to come up way short. So what are the prices on the T4240?
A15 is about 40% faster than A9. That gives you about 3.5 DMIPS/MHz.
-
A15 is about 40% faster than A9. That gives you about 3.5 DMIPS/MHz.
That's about as fast a PA6T. But remember, there are 64 bit ARM chips on the way + nVidia's Denver CPU(64bit also, that will be backwards compatible ).
-
Arm already dominates the mobile phone/tablet markets, it's coming to servers and the desktop next. Freescale's iMX series is ARM successor to 68k, why not?
Some people won't call it Amiga. I don't care what it's called, I just want a reasonable performance desktop PC that doesn't have all that crappy Intel x86 legacy inside it.
ARM makes sense. x86's days are numbered. PPC is as good as dead already.
-
Yes really.
Nobody claims PPC is dead for routers, switches, printers, and similar embedded applications (where CPU's like this one will do just fine, I'm sure). But nobody is developing PPC CPU's for laptop/desktop usage, that stopped 5-6 years ago (whenever it was that Apple went x86) and more importantly, nobody is making viable laptop/desktop motherboards or systems based on PPC CPU's!
It's dead Jim!
The difference between a CPU used in an embedded application and a CPU used in a desktop or laptop is what, according to you?
Where can I buy a desktop motherboard based on an ARM CPU?
-
Arm already dominates the mobile phone/tablet markets, it's coming to servers and the desktop next. Freescale's iMX series is ARM successor to 68k, why not?
Some people won't call it Amiga. I don't care what it's called, I just want a reasonable performance desktop PC that doesn't have all that crappy Intel x86 legacy inside it.
ARM makes sense. x86's days are numbered. PPC is as good as dead already.
Power PC is SO dead that IBM has designed all-new, Power (http://en.wikipedia.org/wiki/Power_Architecture)-based microprocessor for the Wii U around it incorporating features found in IBM Power 7 server processors.
I'd love to have that in a personal computer.
I already have PPC and ARM based systems. I like ARM and will continue to use it for Linux (and eventually for Windows 8).
But right now, I'm sticking with PPC.
BTW - While I have quoted them in the past, DMIPS figures are not that great a measure of a processors power.
-
Arm already dominates the mobile phone/tablet markets, it's coming to servers and the desktop next. Freescale's iMX series is ARM successor to 68k, why not?
Of course it dominates the mobile phone/tablet markets. There's not much competition there. Servers and the desktop are another matter entirely.
Some people won't call it Amiga. I don't care what it's called, I just want a reasonable performance desktop PC that doesn't have all that crappy Intel x86 legacy inside it.
Unless you're writing an OS, or programming in assembler, you'll never have to see "all that crappy Intel x86 legacy inside it".
ARM makes sense. x86's days are numbered. PPC is as good as dead already.
ARM makes sense in situations where power saving is more important than performance. x86 will be around for a long, long time still. IBM, Freescale, and AppliedMicro would disagree with you about PPC.
-
Unless you're writing an OS, or programming in assembler, you'll never have to see "all that crappy Intel x86 legacy inside it".
I saw it when I put it in the socket. I KNOW it's there, that's frustrating enough.
ARM makes sense in situations where power saving is more important than performance. x86 will be around for a long, long time still. IBM, Freescale, and AppliedMicro would disagree with you about PPC.
In other words it makes sense in my PC. I build this system with quietness in mind first and foremost. 45W still seems like quite a lot though.
-
Perhaps I'm mistaken, but to me it seems that the latest Freescale should be i7 caliber in performance?
The e6500 core scales up to 2.5Ghz, so I'm sure there will be models with fewer cores, but higher than 1.8Ghz clock.
Will this be the first time when PowerPC is going to use hyperthreading?
http://www.freescale.com/webapp/sps/site/overview.jsp?code=QORIQ_AMP_SERIES
I think AOS will not change CPU architecture anytime soon:
http://obligement.free.fr/articles_traduction/itwgentle_en.php
+ previously
"When it comes to leading edge PowerPC hardware their level of technical expertise is possibly unrivalled. As a result Varisys has contracted to work on several new development projects to improve and extend the AmigaOne product line". Adam Barnes, Varisys Technical Director said, "There is some exciting cutting edge PowerPC technology on the horizon and we are delighted to have the opportunity of bringing this to the AmigaOne platform." Varisys Managing Director Paul Gentle added, "We enjoyed working on the Nemo development which provided our first experience of the global Amiga community. We look forward to extending our relationship with the new projects and will do our best to "Keep this party going". Please refer to the Varisys and A-EON Technology website for future updates."
-
Exactly my point! :)
Some people say that a dual core X1000 (or indeed a Power Mac with dual G4's) will make it possible to develop SMP for Amiga, but if you aim to do that, to develop "modern" features, like real SMP, real Memory Protection, 64-bit with increased addressable memory limit (which would also make "Virtual Memory"/swap-disk meaningful), it will *require* a break-up from the past, it will require a fresh start with a clean slate. And if you are to do that, why on earth continue the PPC path?
And as you said, no need to emulate the PPC. Most of the essential stuff is still "alive" and can be ported/recompiled. It would be a bit like AROS on x86 in that sense, you won't be able to run old 68k Amiga applications as seamless as you do in MorphOS today (would require UAE), but many of the MorphOS native applications could still be made available in a new version. Heck, with MorphOS 3.x, many things will even come bundled with the OS itself, including CD/DVD burning SW, FTP, etc.
I think you may agree to http://via.i-networx.de/q86.htm then. But my conclusion remains that yet for a switch x86 makes more sense than ARM.
-
A15 is about 40% faster than A9. That gives you about 3.5 DMIPS/MHz.
The Tegra 3, is much faster than the Tegra 2, both being Cortex-A9 (higher clock frequency, more cores, better graphics and other accelerators, etc).
Some Cortex-A15 numbers: (http://www.itproportal.com/2011/03/14/exclusive-arm-cortex-a15-40-cent-faster-cortex-a9/) "What's more we already know of at least one manufacturer who has published benchmark figures for the Cortex-A15; Last month at the Mobile World Congress in Barcelona, ST-Ericsson revealed that its new system on Chip, the Nova A9600, would have two Cortex-A15 core, runs at 2.5GHz and, more importantly, breaks the 20,000 DMIPS barrier.
This means that the ST-Ericsson's tweaked implementation of the Cortex-A15 can reach at least 4.01 DMIPS/MHz, which is itself a 14 per cent improvement upon what ARM's figures.
Given that startups like Caldexa are already developing quad-core iterations of the Cortex-A9, one can easily envision that a 4-core Cortex-A15 will shatter the 40K DMIPS barrier, putting it within touching distance of the AMD Phenom II X4 940 Black Edition (http://www.newegg.com/Product/Product.aspx?Item=N82E16819103471) which reached almost 43K DMIPS running at 3GHz."
The Tegra "4" (http://en.wikipedia.org/wiki/Tegra#Tegra_.28Wayne.29_series) (Wayne) will be quad- or octo-core, made with a 28nm process, thus able to pump up the clock speed simiarly, and again there will be improvements on other aspects of the CPU, like the GPU, etc.
-
The difference between a CPU used in an embedded application and a CPU used in a desktop or laptop is what, according to you?
First, I think Kronos said it in a good way over at AW.net: 1 (http://amigaworld.net/modules/newbb/viewtopic.php?mode=viewtopic&topic_id=35150&forum=33&start=40&viewmode=flat&order=0#652613), 2 (http://amigaworld.net/modules/newbb/viewtopic.php?mode=viewtopic&topic_id=35150&forum=33&start=40&viewmode=flat&order=0#652616), 3 (http://amigaworld.net/modules/newbb/viewtopic.php?mode=viewtopic&topic_id=35150&forum=33&start=40&viewmode=flat&order=0#652621), 4 (http://amigaworld.net/modules/newbb/viewtopic.php?mode=viewtopic&topic_id=35150&forum=33&start=40&viewmode=flat&order=0#652625).
Second, Amiga is single processor. It can only use one of the cores, which will probably make a MPC8610 more attractive than this (the community could have had a readily developed, open sourced and and free to use design based on this (http://www.power2people.org/projects/profile/43) at the price of 20x X1000's. But I guess it was deemed too expensive...), and please note: I'm *not* calling the 8610 attractive in 2012! True SMP is prohibited by Amiga design, can't have real Memory Protection, does only have 31-bit memory space (and no, you can't have "more memory" using swap-disk). If you want features like this, you must break with the past and start all over with a clean slate, a fresh start. Endianness doesn't matter then, choose whatever architecture you want. And my point is that nobody in their right mind would choose PPC.
But most important: Nobody is making *viable* desktops or laptops using these CPU's. There are no viable products, and there won't be any either.
Where can I buy a desktop motherboard based on an ARM CPU?
As you may have seen, I have acknowledged this lack of broad range of motherboards based on ARM, it's focused on "consumer electronic devices" (where an OS like MorphOS potentially *could* find a role to play) so I am not about arguing that, but in this thread, there has been at least 5-6 mentioned, I have both a stationary computer and a netbook based on the Efika MX myself...
Of course the x86 would be much better in this regard! :)
-
I think you may agree to http://via.i-networx.de/q86.htm then.
Some, but some not. Endianness won't matter one bit, for example. And in a micro-kernel environment (like Quark is meant to be), something like the "A-box" is really an abomination; you just don't do something like that, it's way too much going on in one "box" (if you want to call it like that). A more "kosher" way would be having every driver, every device, file system, every application, etc in their own "box" in the Quark memory space. PPC emulation is not important, at least it comes way, way down my personal priority list. I'd be happy to use UAE to run 68k apps and games, like on AROS, in the nice and easy way as MorphOS already offers (which will be more fun on a faster system than PPC, of course)...
But my conclusion remains that yet for a switch x86 makes more sense than ARM.
I agree, while there could be an interesting future on both of these. I guess it would be too much to as to support both? ;)
-
x86's days are numbered.
I have been hearing that since maybe 1998... sigh...
It seems to me Amiga users/devs don't want a true new OS (I mean "OS", we shouldn't care at all about the hardware) but are more interested in running old stuff.
As I see it, the question is simple: do you want something new, taking advantage of new hardware (no matter if it's x86/ppc/68k/arm/big/little endian), powerful, fun to use, future-proof ? Or do you want to run your single-threaded workbench, display your nice png icons, and play with your scrollable screens ?
Just answer the question. If it's the second, no need to even talk about new architectures...
-
A15 is about 40% faster than A9. That gives you about 3.5 DMIPS/MHz.
So what is the price on the T4240? We can go on and on the exact performance that gives one or the other the upper hand, it boils down what is cheap and powerful enough to make the average user happy. Now as far as OS4, I would prefer it to remain in PPCland.
-
I dont know why we are even discussing 2.5 or 3 DMIPS here and there, geez
an i7 x86 has 53.3/MHZ or 177,730 MIPS at 3.33 GHz (6 core).
When I want a new Amiga system, I dont care what CPU it has, I want it CHEAP and FAST!
I just got myself a Dell XPS 17, i7 4 core - 8GB, 3GB geForce 555 GFX laptop and its amazing.
Noone writes in assembler anymore, so we need speed and apps to compensate.
I dont want an 8 core, 2.5 DMIPS CPU that costs an arm and a leg and performs at 1/4 speed.
We need to INCREASE Amiga users, so we need ready made- off the shelf, x86 systems.
price range laptop : 200 euro netbook, 400 euro mid range laptop, 800 euro DELL XPS 17 gaming laptop.
pre built systems : 250 euro, 450 euro, 550 euro, 750 euro high end
be honest, the first thing you wanted to know for an new Amiga system was PRICE !
Amiga company needs :
maybe even 1/3 price for add costs EG:
ADD PROFIT, maybe high as standard 60%
ADD 50-100 euro for monitor, keyboard, mouse.
ADD 25 -50 euro postage
ADD OS4
So a company aiming for 200 euro system HAS to target own costs price to min HALF sell price.
a 1p difference in any cost x 1000 systems sold, costs company an extra 10 euro aka a loss of 10 euro that potentially could be saved.
-
Re:x86 vs. ARM performance vs. others
I was planning to use the 32-bit little-endian target being introduced into the Clang compiler by Google for use with their PNaCl virtual machine in conjunction with AROS. It would be a statically compiled virtual machine (install time compile instead of JIT). This way recompiled apps could be installed on ARM or x86-32 bit machines with equal ease.
It also wouldn't be too hard to add a big-endian 32-bit target to the compiler by modifying the other little-endian one. This would allow PPC and any other big-endian machine such as the N68050 will be able to run the same code.
The big limitation is that running the code from these VMs on anything 64-bit would require a 32-bit emulation sandbox. For Google's purposes, that's fine because their PNaCl browser plugin for Chrome will include a sandbox, but running a 64-bit version of AROS will also require a sandbox.
Another option would be to have a third and fourth file format that support 64-bit architectures in big-endian and little-endians respectively. In my opinion this is a pain.
-
When I want a new Amiga system, I dont care what CPU it has, I want it CHEAP and FAST!
Amiga was never about having the fastest CPU. Amiga was a media PC and it had the advantage because of its custom chipset. The Blitter was a proto-GPU. Xbox 360 gets by on a 3-core PPC chip (I wonder if AROS could be made to run on it?) and is perfectly affordable.
Performance-wise I'd like an Amiga that could play media and browse the internet without sweating, maybe play some games, maybe a bit of graphic design and compose some music... it was a computer you could use to do anything, let your creativity out. It was never a workhorse. I still occasionally use my A1200 for these sorts of things even today.
If you want a blazing fast CPU, get a PC, they've already got that market covered. Amiga needs to fill a niche to succeed. The niche exists. There were people installing Linux on Playstation 3 until Sony stopped them, for some reason.
The niche is not tablets, either. The Amiga niche is a computer that goes under your TV and doesn't need to boot the full OS to play a DVD.
At least, that's the way I see it.
-
Amiga was never about having the fastest CPU. Amiga was a media PC and it had the advantage because of its custom chipset. The Blitter was a proto-GPU. Xbox 360 gets by on a 3-core PPC chip (I wonder if AROS could be made to run on it?) and is perfectly affordable.
Performance-wise I'd like an Amiga that could play media and browse the internet without sweating, maybe play some games, maybe a bit of graphic design and compose some music... it was a computer you could use to do anything, let your creativity out. It was never a workhorse. I still occasionally use my A1200 for these sorts of things even today.
If you want a blazing fast CPU, get a PC, they've already got that market covered. Amiga needs to fill a niche to succeed. The niche exists. There were people installing Linux on Playstation 3 until Sony stopped them, for some reason.
The niche is not tablets, either. The Amiga niche is a computer that goes under your TV and doesn't need to boot the full OS to play a DVD.
At least, that's the way I see it.
Thank you very much for that sentiment. Personally, I just want DIFFERENT hardware. I already have an X86. And if I was using an ARM I'd be running Linux.
BTW - Samuraicrow, you must keep us informed on that idea. It sounds fascinating.
BTW - I wasn't interested in the 8610, I was interested in the MPC8640/8641 (and now the T5040).
-
The niche is not tablets, either. The Amiga niche is a computer that goes under your TV and doesn't need to boot the full OS to play a DVD.
At least, that's the way I see it.
They already have something like that. It's called a DVD player.
-
I dont know why we are even discussing 2.5 or 3 DMIPS here and there, geez
an i7 x86 has 53.3/MHZ or 177,730 MIPS at 3.33 GHz (6 core).
I think those were MIPS per core.
So, i7 has about 8.8DMIPS per core?
Latest PPC designs have 6DMIPS per core and 12 cores (24 virtual cores). That's 130,000MIPS at 1.8Ghz for embedded use (2.5Ghz is the maximum clock rate -> 180,000MIPS or so with desktop caliber cooling?)
So, for performance, we "only" need SMP. :-(
The "problem" just seems to be that custom board are more expensive than mainstream boards.
-
They already have something like that. It's called a DVD player.
Zing!
-
I think those were MIPS per core.
So, i7 has about 8.8DMIPS per core?
Latest PPC designs have 6DMIPS per core and 12 cores (24 virtual cores). That's 130,000MIPS at 1.8Ghz for embedded use (2.5Ghz is the maximum clock rate -> 180,000MIPS or so with desktop caliber cooling?)
So, for performance, we "only" need SMP. :-(
The "problem" just seems to be that custom board are more expensive than mainstream boards.
No John,
That's zing!
>we "only" need SMP
There's always Linux
-
Multi-core ARM board? :D
Oh, I mean multi-chip.. I wonder where the soldering iron is....
-
I think those were MIPS per core.
So, i7 has about 8.8DMIPS per core?
Latest PPC designs have 6DMIPS per core and 12 cores (24 virtual cores). That's 130,000MIPS at 1.8Ghz for embedded use (2.5Ghz is the maximum clock rate -> 180,000MIPS or so with desktop caliber cooling?)
So, for performance, we "only" need SMP. :-(
The "problem" just seems to be that custom board are more expensive than mainstream boards.
AMD Llano's 400 stream processors can do 260,000 MIPS at 650Mhz (400 stream processors or 80 VLIW5 math processor array). K markings can overclocked the "Radeon HD cores" to 900Mhz.
-
Amiga was never about having the fastest CPU. Amiga was a media PC and it had the advantage because of its custom chipset. The Blitter was a proto-GPU. Xbox 360 gets by on a 3-core PPC chip (I wonder if AROS could be made to run on it?) and is perfectly affordable.
Most of Xbox 360's heavy compute work is done on AMD Xenos GpGPU i.e. 48 SIMD math processor array.
-
They already have something like that. It's called a DVD player.
Very good. But it would be better to say that it's called an Xbox. The difference is an Xbox doesn't come with a "productivity" OS and a keyboard.
Most of Xbox 360's heavy compute work is done on AMD Xenos GpGPU i.e. 48 SIMD math processor array.
Exactly! That's its "blitter".