> POWER

Just a small hint at terminology: POWER is IBM's server CPU microarchitecture implementing the Power ISA (instruction set architecture) launched in 2006 as successor to the PowerPC ISA. See:

http://en.wikipedia.org/wiki/Power_Architecture#Glossary

> That URL doesn't show me the post you want. It might be that I chose
> "most recent at top" in my MorphZone forum preferences.

Yes, that's the reason. Thanks for pointing it out. I'll make my internal links more compatible from now on, using your URL pattern (plus fragment identifier).

> See that "75659" now as the parameter "post_id", instead of the hash (#).

The "#<number>" is needed nevertheless because it works as a fragment identifier.

Quote:

> Freescale has just released a 64bit successor to the e500, thr e5500.

It's not released yet, just announced.

> IBM has a processor based ion the P7

That's what The Register claims at least. As I said, I'm not really convinced that PowerEN is really POWER7 based (see in-order vs. out-of-order execution).

> AMCC is carrying forward a 40nm variant of their Titan core without
> Intrinsity's Fast14 technology.

As far as I understand, the new 40 nm core that is supposed to substitute the scrapped Titan core won't be a Titan variant but another core altogether, offering similar features. I'd be disappointed if the undisclosed 40 nm core really is supposed to just be Titan with a shrinkage applied to it to make up for the abandonment of Intrinsity's Fast14 technology.

> It is even rumored that the next PlayStation portable will be Cell based.

That rumour is utter nonsense. The PSP2 will most likely be nVidia Tegra 2xx (ARM Cortex-A9 core) driven.

http://www.semiaccurate.com/2010/07/14/sonys-psp2-powered-nvidias-tegra-line/

> Freescales planned e5500 cored products may be fast (Mips) than that
> processor or Titan.

In terms of DMIPS the e5500 will *definitely* be faster than the sinked Titan or a single PA6T core.

Quote:

Personally, I's settle for a Titan derived processor without Fast14. At 40nm it might scale fairly well. Even a 1.8 Ghz variant could be much more powerful than AMCCs current line.

Until we see first hand the floating point capabilities of Freescales planned e5500 core, the (DMIPS) disadvantage of the Titan (or its successor) may not be as great as we're assuming.

More than any other factor, the comparative price of e5500 cored and Titan cored processors and their availability may affect what products each is incorporated into.

I could see a 1.8Ghz "Titan" based motherboard, if the cost could be kept down, making a nice substitute for Acubes product line or as an alternative to a Pegasos or Mac.

> Personally, I's settle for a Titan derived processor without Fast14. At 40nm
> it might scale fairly well. Even a 1.8 Ghz variant could be much more powerful
> than AMCCs current line.

While the APM83290 was supposed to run up to only 1.5 GHz, the Titan core itself was announced to scale up to 2.0 GHz. I expect a Titan replacement core to not undercut that figure, and the Titan replacement core based processor should really not undercut the 1.5 GHz figure of the APM83290. So yes, 1.8 GHz would be fine for the processor I guess, and you're right that it would definitely be much better performing than anything they have right now, be it officially announced or sampling or in volume production.

> Until we see first hand the floating point capabilities of Freescales planned
> e5500 core, the (DMIPS) disadvantage of the Titan (or its successor) may
> not be as great as we're assuming.

I really don't get your point here. The DMIPS per MHz figures of both the e5500 and the Titan are officially known: 3.0 vs. 2.0. That's 50% more DMIPS per MHz for the e5500. So to get on par the Titan would have to be clocked 50% faster than the e5500, which it is not. Quite to the contrary, the e5500 is announced to scale to a 25% higher clock rate than the Titan (or even 47% if we look at the announced processors based on the cores). That's a 87.5% higher DMIPS performance altogether for the e5500 (and even 120% higher on a processor level).
Of course that's all assuming that the Titan replacement core will deliver the same DMIPS per MHz figure as Titan, which would be the case if it's really just a shrinked Titan without Fast14.

> I could see a 1.8Ghz "Titan" based motherboard [...] as an alternative to a [...] Mac.

An "alternative" to a G4 Mac as in "similar performance", yes.

My only point was that MIPS as a general measurement is useful, but its not a perfect way to compare floating point power. Without Altivec and sporting a floating point engine thar's twice as fast as its predecessor, how willing pointing point operations compare between the 5500 and the Titan?

> My only point was that MIPS as a general measurement is useful, but its not a
> perfect way to compare floating point power.

Dhrystone (DMIPS) is a pure integer benchmark. It doesn't measure floating point performance at all. Floating point performance can be measured for instance by Whetstone (MWIPS), LINPACK (MFLOPS) or LAPACK. See:

http://en.wikipedia.org/wiki/Dhrystone
http://en.wikipedia.org/wiki/Whetstone_(benchmark)
http://en.wikipedia.org/wiki/LINPACK
http://en.wikipedia.org/wiki/LAPACK

> how willing pointing point operations compare between
> the 5500 and the Titan?

That remains to be seen. All processors from Applied Micro which have an FPU (i.e. some of the PPC440 core based ones and all of the PPC464FP core based ones) deliver 2.0 MFLOPS per MHz in both single and double precision. I've yet to find numeric floating point performance figures for the e5500. All I know is that compared to the e500mc the e5500 has twice the single precision performance per clock and up to quadruple the double precision performance per clock. But I don't have any real numbers for the e500mc either.

Edit:

Freescale's e500mc presentation notes that e500mc has "same floating point unit as e300", but which is half clocked in relation to the core clock and thus half performant. So I'd conclude that compared to the e300 the e5500 has the same single precision performance per clock and up to twice the double precision performance per clock. Now we only need the floating point performance figures for the e300 to get the whole picture ;-)

[ Edited by Andreas_Wolf on 2010/8/17 3:35 ]

Update:

> the Titan replacement core based processor should really not
> undercut the 1.5 GHz figure of the APM83290.

Seems we know at least the name of the Titan replacement core based processor: APM86xxx.

https://morph.zone/modules/newbb_plus/viewtopic.php?topic_id=6196&forum=11&post_id=75913#75913

These figures worry me.

>(i.e. some of the PPC440 core based ones and all of the PPC464FP core based ones) deliver 2.0 MFLOPS per MHz in both single and double precision.>

If the Freescale 5500 core produces higher DMIPS but lower floating point performance, how will Titan's descendants look in comparison?


[ Edited by Jim on 2010/8/26 4:56 ]

> These figures for the e5500 worry me.

Why? We (as in "you and me") only have relative figures to other Freescale cores, no absolute ones yet, haven't we?

> If the Freescale core produces [...] lower pointing point performance

We don't know this yet, do we?

No, I guess we get no firm answer until both are tested.

> the XCGPU you mentioned, Andreas, may be two separate dies
> under a common heatspreader.

Seems it is not:

http://arstechnica.com/gaming/news/2010/08/microsoft-beats-intel-amd-to-market-with-cpugpu-combo-chip.ars
http://www.eetimes.com/electronics-news/4206424/Microsoft-IBM-Xbox-360-250G-chip

Edit: added link to another article

[ Edited by Andreas_Wolf on 2010/8/25 6:36 ]

Quote:

Unreal! Thanks for that reference Andreas. I've been dying to know whether this was the case or not. If IBM, Global Foundries and ATI (owned by AMD) all cooperated with Microsoft to accomplish this its really impressive.
The only real limit I can see to the design is the FSB connection between GPU and CPU. Its one chip, but the design is still basically the same as the older models.
Not that I expect them to truly integrate the CPU and GPU, but I'm hoping that at some point that's AMD or Intel's intention. Otherwise the whole design looks like an early dual core Pentium (where the processors have to share and communicate through a bus).
I know its a small complaint, but future CPU/GPU combos could use a better interconnects than the system bus.

Still, a neat piece of work. Pity they can't be purchased. I still think it would make a great base for a motherboard (as would the Cell B.E.).

I know there are plenty of people who think I may be a little too optimistic about the PPC's future, but there are some really nice designs out there (and the XBOX360 and PS3's designs remind me more of an Amiga than a standard PC does).

Again, thanks for keeping the question in mind (if your memory was any sharper I'd think you were digital).

Update:

> Seems we know at least the name of the Titan
> replacement core based processor: APM86xxx.

"Applied Micro will announce a new family of multicore SoC processors that use SMP and AMP architectures to address the demanding needs of applications within the Enterprise, Data Center, Industrial, Gateways / Access Point, and more."
http://www.appliedmicro.com/events/linley2010.php
http://www.linleygroup.com/Seminars/conference_fall_program_day1.html
http://www.linleygroup.com/Seminars/conference_fall_talks.html

Edit:
Not only "announce" but also "introducing and demonstrating".
Source

[ Edited by Andreas_Wolf on 2010/9/24 23:18 ]

Yes not bad specs, if they can get the price reasonable and the CPU MHz up a bit, like they did with the original SAM, it may be a nice little box.

MorphOS Port anyone? :)

Addendum:

> compared to the e500mc the e5500 has twice the single precision
> performance per clock and up to quadruple the double precision
> performance per clock.

Nothing about FPU performance, but an interesting comparison between e500mc and e5500 regarding integer related performance:

http://www.power.org/events/POWERWEBINAR082010/PowerOrgWebinarMilAeroGieske-1.pdf (page 13)

According to the rightmost chart the e5500 delivers only about 76% of the DMIPS per Watt delivered by the e500mc.

Performance per watt figures don't scare me as much as that next comment "e500mc & e5500 have less single-thread performance than some cores".

> Performance per watt figures don't scare me
> as much as that next comment

Really? That comment is in no way surprising to me. The "e5500" bar in the leftmost chart shows the e5500 at 2.2 GHz, i.e. 6600 DMIPS. The "MaxPerf" bar to the right is twice as high as the "e5500" bar, i.e. 13200 DMIPS. While I don't know which is the best performing core in the whole wide world in terms of DMIPS, I'm sure that 13200 DMIPS is no unmatched figure.
For lists of figures for various processors refer to:

http://www.roylongbottom.org.uk/dhrystone%20results.htm
("Dhry2", Opt" column is the one to look at)

http://www.anime.net/~goemon/benchmarks.html
(values in the "RESULT" column must be divided by 1757 to get DMIPS figures)

Yes, I guess if the statement could be so obtuse (not even mentioning what "cores" they're talking about) that it doesn't carry much weight.
But does the DMIPS per watt figure make any real statement about the relative power of each processor, or is it just a measure of efficiency?

> does the DMIPS per watt figure make any real statement
> about the relative power of each processor, or is it just
> a measure of efficiency?

Just imagine the inverse, i.e. Watts per DMIPS (where lower is better of course), which means how much Watts are needed to compute 1 DMIPS. This way it is more easy to understand what this measure of efficiency (there you go) in the shape of a combination of DMIPS and Watts means.
The relative DMIPS performance of the cores, on the other hand, is shown by the leftmost chart, as I explained already.

In case that with "power" you mean rather electrical power, not performance:

(1) 2.2 GHz e5500 has 1.76 times the DMIPS performance of 1.5 GHz e500mc.
(2) 2.2 GHz e5500 needs 1.32 times the Watts to compute 1 DMIPS compared to 1.5 GHz e500mc.

Results:
A) 2.2 GHz e5500 draws 1.76 * 1.32 = 2.3 times the power of the 1.5 GHz e500mc.
B) At the same clock, while being only 20% more performant, the e5500 draws 58% more power than the e500mc.

No Andreas, you had it right the first time, although the second set of figures is also interesting.
What would you think creates these inefficiency issues? The e5500 is clearly a superior processor when compared to the e500, but from both of those points of view it the e5500 suffers in this comparison.

> What would you think creates these inefficiency issues?

Good question. But I've to admit that I really don't know. Freescale presented the e5500 as being basically a 64 bit capable version of the e500mc (hence the 'e500mc64' moniker) with an enhanced FPU (but which of course is irrelevant in terms of integer performance). With that said it's really strange that in terms of DMIPS the e5500 is 24% less efficient than the e500mc.

> for all we know, someone could very well have been working on a non-PPC version
> of MorphOS for a long time already...

If that's the case then it's probably not an x86 version at least. From Commodore USA's CTO:

"When I contacted Team MorphOS, Ralph S told me they had no intention of porting to x86 any time soon."
http://www.amiga.org/forums/showpost.php?p=579267

Addendum:

>> how willing pointing point operations compare between
>> the 5500 and the Titan?

> All processors from Applied Micro which have an FPU [...] deliver 2.0 MFLOPS per MHz in
> both single and double precision. [...] compared to the e500mc the e5500 has twice the
> single precision performance per clock and up to quadruple the double precision
> performance per clock. [...] Freescale's e500mc presentation notes that e500mc has
> "same floating point unit as e300", but which is half clocked in relation to the core clock
> and thus half performant. So I'd conclude that compared to the e300 the e5500 has the
> same single precision performance per clock and up to twice the double precision
> performance per clock. Now we only need the floating point performance figures for the
> e300 to get the whole picture ;-)

I found that 603e/G2_LE/"e300c0" does only 0.4 MFLOPS per MHz according to a posting on powerdeveloper.org:

http://www.powerdeveloper.org/forums/viewtopic.php?p=13781#13781

I don't know if that's single or double precision (or even both) but either way if both my interpretation of that figure and my assumptions above are correct, then it doesn't bode too well for the floating point performance of the e5500 in general as well as compared to the Titan (or probably rather its replacement) specifically.

Update:

>> Seems we know at least the name of the Titan
>> replacement core based processor: APM86xxx.

> "Applied Micro will announce a new family of multicore SoC processors that use
> SMP and AMP architectures to address the demanding needs of applications
> within the Enterprise, Data Center, Industrial, Gateways / Access Point, and more."
> http://www.appliedmicro.com/events/linley2010.php
> http://www.linleygroup.com/Seminars/conference_fall_program_day1.html
> http://www.linleygroup.com/Seminars/conference_fall_talks.html

Seems like the APM86xxx's Power Architecture core(s) may be complemented by a nice co-processor core:

http://www.tensilica.com/news/335/330/AppliedMicro-Picks-Tensilica-s-Dataplane-Processors-for-High-Throughput-Communications-Chip-Design.htm
http://www.tensilica.com/products/xtensa-customizable/xtensa-lx.htm

Edit:

I found an interesting PDF presentation file dating August 2010 over there.
The most interesting pages are #16, which shows a product roadmap mentioning "Future Viper SoCs", and #18, which reads "Viper Technology -> Breakthrough Performance at Disruptive Cost Basis".
To my surprise, the roadmap on page 16 still contains the Titan core based "Gemini" (APM83290), which had been semi-officially declared dead one month before. Strange.

Further insightful (but older) documents are this one (August 2009) and that one (October/November 2009). Both documents's pages 11 show a product roadmap including not only Titan (still vivid at that time) but also dummy-named products "PPC-A" ("Avail 1H 2010"), "PPC-B" ("Avail 2H 2010") and "PPC-C" ("Avail 1H 2012").

[ Edited by Andreas_Wolf on 2010/9/24 23:49 ]