> From the DMIPS/MHz per core figures you've list earlier

Dhrystone is a very artificial benchmark.

> the e500, e600, G5, and e5500 if run at similar speeds would all outperform the PA6T.
> Of course, not all those cores clock to 1.8 Ghz (or above), but several do.

From my list only e5500, PPC970 and PPC470 would be able to beat the 1.8 GHz PA6T in DMIPS figures at their respective maximum clock rates.

> you better hope you can find a good used Pegasos, because otherwise
> you're stuck running on one of Acube's anemic boards.

I'm eager to know how the PPC460EX performs in real life compared to the equally clocked MPC7447 in the Pegasos II G4. I expect the G4 to perform somewhat better though, even without AltiVec.

I'd expect the 7447 to outperform a similarly clocked PPC460EX as well, and the dhrystone measurement you gave for the Titan don't look too promising either.

A 7448 would outperform a 7447, and apparently the three processors you've mentioned would outperform that (of course, all based on - as you pointed out - a very artificial test).

And Acube's pricing, based on performance, is worse than A-eon's.

Now, as to the max speeds of the processors you've mentioned, they appear to be higher than the PA6T (unless that runs at 2.0 and so far its only been seen running at 1.8).

Wouldn't we still be better off with G5 Powermacs. I can get a fairly well equipped high speed G5 for hundreds less than an Acube based system.

And, when (if), we get Powermac support there are the 7447 and 7448 upgrades. A 1.8Ghz might or might not outperform an X1000, but its going to walk all over anything from Acube.

As HD video decoding with current MorphOS supported hardware can be problematic at really high res, why worry about even slower hardware than the Pegasos?

> apparently the three processors you've mentioned would outperform that
> (of course, all based on - as you pointed out - a very artificial test).

Based on that benchmark, a 1.8 GHz PA6T core would marginally (by 1.3%) outperform a 1.7 GHz 7448 (that's the maximum stock clock rate according to Freescale) as well.

> Acube's pricing, based on performance, is worse than A-eon's.

Taking only one core of the PA6T into account, Dhrystone performance of the 1.8 GHz PA6T is twice that of the 1.0 GHz PPC460EX. I expect a full Sam460ex system to be priced somewhat less than half of the X1000. On the other hand I expect real life performance of one 1.8 GHz PA6T core to be much better than only twice that of the 1.0 GHz PPC460EX, add to that AltiVec capability, faster bus, larger caches and the option to use the 2nd core whenever OS4 will support that. So yes, you may be right in your estimation.

> as to the max speeds of the processors you've mentioned, they appear to be higher
> than the PA6T (unless that runs at 2.0 and so far its only been seen running at 1.8).

Not quite. The PPC470 is clocked at up to 1.8 GHz as well (in the LSI Axxia, while being specified at only 1.6 GHz max by IBM btw).

> Wouldn't we still be better off with G5 Powermacs.

Performance wise, probably yes. But not everyone would be happy with the high wattage of the PPC970 I guess. That's where the PA6T shines in comparison. In real life performance (i.e. not Dhrystone) I expect the PA6T to have significant higher performance/wattage figure than the PPC970 by providing only slightly less performance/clock.

> there are the 7447 and 7448 upgrades. A 1.8Ghz might or might not
> outperform an X1000

In terms of Dhrystone an e600 core above 1.72 GHz would outperform a 1.8 GHz PA6T core. So the overclocked 1.8+ GHz G4 upgrades deliver more DMIPS than a 1.8 GHz PA6T core. In real life performance though I think a 1.8 GHz PA6T core would outperform an equally clocked 7448.

> but its going to walk all over anything from Acube.

Yes, that's for sure.

> As HD video decoding with current MorphOS supported hardware can be problematic
> at really high res, why worry about even slower hardware than the Pegasos?

I'm just curious to the whole thing. Be assured that I definitely won't downgrade from my 1.5 GHz G4 ever :-)

|||


[ Edited by Velcro_SP 20.04.2011 - 07:02 ]

> I spoke to the general idea of making use of (note: not necessarily "adding," esp.
> where already present) dedicated video hardware and more RAM to a system with
> slow processor to improve performance.

You wrote "dedicated graphics chip", not "dedicated video hardware" (which can be a dedicated chip, a dedicated core inside a SoC or whatever). I advice you to eventually look up what a chip in terms of electronics actually is. So regarding the MPC5121e the usage of a "dedicated graphics chip" (which you subsequently try to turn into the wider concept of "dedicated video hardware") would be equal to adding a second graphics core to the system (one graphics core in the MPC5121e SoC plus one graphics core in the "dedicated graphics chip" makes two graphics cores in the system).
Again it boils down to the question I've already asked you: And now you say that inbetween, i.e. in the 3rd paragraph, you didn't mean to include the MPC5121e when mentioning "slow, low-wattage processor"? Yes or no?

> PowerQuicc is said by Wikipedia to be used in automotive and routers.

We were talking specifically about PowerQUICC II Pro because it's e300 based like the MPC5121e. Automotive is no target application of any PowerQUICC processor (and I doubt it is used in that space either, so I consider Wikipedia wrong in that *) because QUICC is a communications engine:

http://www.freescale.com/quiccengine
http://www.freescale.com/files/graphic/other/NCSG_QUICCENG_VID.wmv
http://en.wikipedia.org/wiki/QUICC (not quite encyclopaedic in its wording)

The MPC8377E I was specifically aiming at has the following target applications according to Freescale: multifunctional printer, small/medium business gateway, IP-PBX systems, network attached storage (NAS), digital media server (DMS).
Admittedly, that may sound like it's hard to build something like a nettop or netbook with it. But technically it has everything you would want for such system, except for (as outlined before) a GPU and AC'97. For these both a dedicated graphics chip (there you go) and a dedicated sound chip connected to the SoC's PCI(e) bus would be needed, or alternatively a combined graphics and sound chip (for instance like the Sam460ex has on-board).

> It's fine with me if someone puts a PowerQuicc processor in a netbook or
> desktop or nettop or smartphone, but I'm not aware that's been done.

Me neither. But that's been the whole point of that part of the discussion: to outline which Power Architecture based processors that are faster than a 400 MHz e300 *could* be used to develop and build such systems.

*) the Wikipedia claim that "Freescale are using PowerQUICC processors as a part of their mobileGT platform" is also false. PowerQUICC and mobileGT (MPC51xx/52xx) do not overlap.

Thanks for making all the careful calculations for speculations I made out of opinion, Andreas.
I actually feel much better about MorphOS' current and proposed future supported hardware.

I didn't downgrade, but my most recent purchase of an Emac doesn't quite have to power of your Mac Mini.

Nice to know that when Powermac support is introduced, I can spend a few hundred bucks and upgrade my Powermac's processor to match the X1000's horsepower.

Blame AIM for the G5's power usage. Since development stagnated and die size is too large (is there anything smaller the 90nm?).

Why are we still discussing the e300? I know you love banter/speculation with a focus on correct information, but this isn't far enough away from the PowerPC 603e for my tastes. Better processors than these have already reached EOL.

[ Edited by Jim on 2010/7/4 16:46 ]

> Nice to know that when Powermac support is introduced, I can spend a few hundred
> bucks and upgrade my Powermac's processor to match the X1000's horsepower.

AFAIK there's still nothing said about support for 3rd party processor upgrades. Do you think they'll work out of the box with MorphOS? Regarding the 7448 specifically: https://morph.zone/modules/newbb_plus/viewtopic.php?topic_id=6910&forum=11#71008

> is there anything smaller the 90nm?

No, PPC970 family is 90 nm all over.

> Why are we still discussing the e300?

I think it's because some in this thread still see a market chance for an e300 based low-power nettop (like C114/C120/LimeBox/X1) or netbook (like LimeBook/Linkbook) running MorphOS. I just say that if it really must be e300 based (for whatever reason) then I'd be all for the MPC8377E. Anything performing below an 800 MHz e300 would be too weak for my own personal taste, in terms of nettops or netbooks that is (real desktop use being another matter altogether where I don't ever want to go below 1.5 GHz G4 again).

> this isn't far enough away from the PowerPC 603e for my tastes.

The e300c4 that is built in the MPC8377E (and MPC5121e) has twice the cache size of the 603e and is 2-way superscalar which the 603e (as well as e300 versions before e300c4) is not.

https://morph.zone/modules/newbb_plus/viewtopic.php?topic_id=6196&forum=11&post_id=64270#64270

Add to that the leap in clock rate from 300 MHz to 800 MHz. So I conclude that an 800 MHz e300c4 is at least 3 times as fast as a 300 MHz 603e in real life performance. That might be enough for a low-power nettop or netbook I think.

> Better processors than these have already reached EOL.

Yes, but you should see that there're actually two different things being discussed in this thread in parallel: (1) high-performing desktop systems and (2) low-power nettops/netbooks (where "faster" does often not equal "better").
If you're not interested in discussing (2) then you should simply ignore anything related to that, including the e300 discussion.

Quote:

Yes, you have a point. I have in strong interest in desktops and notebook computers, but the purpose of netbooks confuses me (perhaps longer battery life is the most important point).

I can see the advantages of lower power draw and more convenient compact size, but the low power processors and hard to read smaller displays really put me off.

As I don't have any input that's constructive or informative on the e300 discussion, I'm dropping out of that part of the thread.

And, since I also haven't seen a processor with an on die GPU that's worthy of consideration, I'll refrain from commenting on that concept. Maybe AMD or Intel's future planned releases will change my mind, but I still have a bad taste in my mouth from Cyrix's shot at this idea.

While not intended for desktops or notebooks, an e5500 cored processor seems low draw enough for a good notebook. But the cost and complication of notebook design (and netbook design for that matter) make using anything but a pre-existing design impractical.

> While not intended for desktops or notebooks, an e5500 cored processor
> seems low draw enough for a good notebook.

Yes, you're obviously not the only one to come to that conclusion:

http://blogs.freescale.com/2010/06/23/what%e2%80%99s-up-with-64-bit-embedded-computing/#comments

Unfortunately, Preet Virk (who is strategic marketing director for Freescale's Networking Processor Division), while referring to Roberto's comment in general (and some points specifically), did not comment on his notebook remark.

Quote:

Am I the only one very scared at reading this phrase from him?

Quote:

It sounds like 64 bit processors were at people's homes by the time The Mamas and the Papas were still on the road... Or is it that his phrase ends up forty years later that it starts?

Alright, I also red sometime ago an IT expert saying the Commodore 64 was the first 64 bit computer.

Quote:

the key is "for super computers", so nothing like "people's homes".

Quote:

Ha ha ha! Good one! No.

> Or is it that his phrase ends up forty years later that it starts?

Rather twenty years:

http://en.wikipedia.org/wiki/DEC_Multia (1994)
http://en.wikipedia.org/wiki/Nintendo_64 (1996)

More info:
https://en.wikipedia.org/wiki/64-bit_computing#64-bit_address_timeline

> the key is "for super computers", so nothing like "people's homes".

Huh? "Home computers and gaming consoles" very much sounds like "people's homes".

Quote:

Don't know about the majority, but I liked the idea of an ultra small notebook for years. I am not at home quite often, a netbook is just tiny enough to get carried with you all the time. Think about those ppl who don't drive trucks, SUVs or the likes, but rather go by train or bicycle. Size does matter The pocket mony price is another plus.
I was hoping for an ultracheap mini ppc laptop when I heard about the 5200 years ago, but this was a while b4 Asus came along with their Eee. Manwhile the market has changed a lot and getting a foot between the door became more diffucult. And an 300/400MHz definitely leads nowhere today. An e300/800 could be still okay today, but only if really dirt cheap. I would focus on QorIQ.
I am still a bit sad about the 86xx. The 8610 would have made a nice netbbok, but e600 seems to be EOL - well, who knows maybe Freescale rediscovers Altivec one day. At least ppc got a bit of momentum again recently.

> maybe Freescale rediscovers Altivec one day.

I fear that won't happen. Regarding Power Architecture it's all about VSX today:

https://morph.zone/modules/newbb_plus/viewtopic.php?topic_id=6196&forum=11&start=140#74094
http://en.wikipedia.org/wiki/AltiVec#VSX

But I don't know enough about VSX to make a proper comparison to AltiVec/VMX. Feanor?

Quote:

Yes, like, say, absurd demands like HD video in a four inch screen, powered by batteries. The worst thing is that now it IS actually possible!

(Am I the only one who still remembers when one had to wait for a computer to finish something?)

Quote:

I replied to "It sounds like 64 bit processors were at people's homes by the time The Mamas and the Papas were still on the road...". I agree that now 20+ years after, 64-bit is the norm.

Quote:

Difficult to say, no access to Power7 hardware, but from the _very_ few docs out there that I could find, it looks like a mix of Cell SPU and AltiVec. Some instructions are the same, others -esp the double fp or the decimal fp support- have no Altivec equivalents. No idea how it performs really. It probably is quite powerful, but there are virtually no benchmarks out there, IBM sure does a crap job of promoting its technologies.

Update: this is the only doc I found with actual specs/performance info for VSX.

[ Edited by feanor on 2010/7/5 16:04 ]

[ Edited by feanor on 2010/7/5 16:10 ]

Quote:

Zylessa, trains and bicycles, man that's so European! I don't drive a truck or and SUV, but train service in my State is extremely limited (blame the US government for short sighted transportation policies) and a bicycle trip to work (for me) would be 80 miles round trip.

And I think I could get used to a netbook, even with a small display and keyboard, if the screen resolution was better (say 1024X768 or at least 800X600) and there was more processing power (you're right 300 to 400 Mhz doesn't cut it these days).

Andreas' minimum works, but again I'll agree with you that a QorlQ processor would be better. And it is a shame that the e600 core is EOL'd. I spent a lot of time studying apps for the 86xx processors.

Still, small Apple G4 notebooks start with a 12" screens, so that's fairly compact (not quite a netbook, but not stupid sized like many laptops on the US market). We should eventually have support for those.

And I'm glad a few people here have noticed to increase in PPC development. A consumer PA Semi based motherboard, AM titan based processors, the new e5500 64bit core - all welcome developments.

Hopefully, I'll still be able to use MorphOS on fairly decent hardware for some time to come.

Quote:

Offtopic perhaps, but, IIRC, AMD owns xenon production lines nowdays (the Power chip of xbox 360).

That I didn't know, but by AMD you now mean Global Foundries don't you (since AMD has spun off production facilities).

And don't worry about Andreas' comment, he knows perfectly well that none of the three companies I mentioned produce PPCs. However, AMD and Intel may actually be the first to introduce processors with decent on die GPUs (but it hasn't happened yet, all current processors offering this feature have horrible GPU performance).

>> ATI/AMD do Power Architecture CPUs?

> IIRC, AMD owns xenon production lines nowdays

With the merger of Chartered into GlobalFoundries last year AMD became *minority* stockholder of the Xenon production business (ATIC being the majority one).
Anyway, with "do" I actually meant "develop", not "manufacture". A chip being manufactured by GlobalFoundries doesn't make it a GlobalFoundries chip, the same way it didn't make it a Chartered chip before :-)

> I replied to "It sounds like 64 bit processors were at people's homes by the time The Mamas
> and the Papas were still on the road...". I agree that now 20+ years after, 64-bit is the norm.

Then you seem to have missed the actual question: "Or is it that his phrase ends up forty years later that it starts?", where the answer would simply be: Yes, it does (with the small correction that it's rather twenty years, not forty) :-)

> from the _very_ few docs out there that I could find, it looks like a mix of Cell SPU
> and AltiVec. Some instructions are the same, others -esp the double fp or the decimal
> fp support- have no Altivec equivalents.

Thanks for information. My main question would be if VSX is backwards compatible to VMX/AltiVec by itself or if POWER7 supports both by having both implemented separately. Or similarily put: Does a CPU implementing VSX automatically support VMX/AltiVec, i.e. is VSX a strict superset?

I think it would be pointless for AMD and Intel to "do" anything other than what they're doing right now.
Currently, they're providing virtually all processors used in personal computers (unless you want to include iPad and other compact devices in with PCs).
While some of use seem to be worried about ARM, X86 is the real competitor. Upgraded constantly, originating with a small company building calculator processors, and very luckily picked for IBM first off the shelf consumer oriented computer.
This unlikely architecture has matured to dominate the market.
Would that it weren't true, but AMD and Intel have no reason to "do" anything else.