A-EON Technology CVBA and Varisys Ltd Announce Partnership

[unusedunused]

I'm sure google will bring up something.


The PA6T:

Are fabbed at 65nm process
Dual core, 32/64-bit
Have altivec
Run beyond 2GHz (haven't found a direct source for the ceiling, but 2.5GHz is quoted a lot).
TDP of ~7W @ 2GHz
Upto 8MB L2 cache (configurations from 512K upwards) with "MOESI" cache coherency (see AMD64 docs for more on that).
Virtualisation/Hypervisor features
Dual integrated DDR2 memory controllers at ~1066MHz

can you post links to this ?
there was TDP of 25 Watt written here.and more of 7 Watt for average.

http://www.realworldtech.com/page.cfm?ArticleID=RWT032607001221&p=2

The pa6 is a slow embedded CPU(there is no wonder possible.less power bring less performance /MHZ but for sure better as AMCC) and i doint think it give performance /MHZ of a G4

the only benchmark i know of pa6 is this.

http://www.coremark.org/benchmark/index.php?pg=benchmark

here pa6 is not so good(around, as AMD. and PA6 have lots bigger caches.
but thats the only benchmark i know of pa6 so i only write i think pa6 is little slower as G4, but to say it is a little slower need more benchmarks of pa6.

The Intel Atom is called as a slow CPU.this are the values of 2 threads

P.A. Semi PA6T-1682M 1.7GHz GCC 4.3.2(Debian 4.3.2-1.1) ____ coremark/MHZ= 3.234___ 5497.80 2 PThreads

Intel Atom N270 1.6GHz GCC4.4.1____ coremark/MHZ= 2.855 __ 4567.253 2:PThreads vote 02/18/10

this Atom N270 is no real 2 core CPU.

http://de.wikipedia.org/wiki/Intel_Atom

It have only hyperthreading.But you can see in Core mark /MHZ its only 13% slower as a Pa6 in single core value.

modern X86 are lots faster as Atom CPU, and i think a G4 is also much more faster as 16% as a Atom CPU.

also the ATOm CPU have very small 1.Level caches.lots smaller as pa6.also the 2. level cache of ATOM is smaller

when i see the heatsink of the X1000 i see not that this is able to cool a pa6t.

The heatsink look realy small on the CPU, if it should be a pa6 CPU it have a TDP of 17 Watt.

When look at passive Heatsink specs, a 53 mm * 53 mm *16 mm Heatsink have a rate of 7,3 Kelvin /Watt.

http://www.conrad.de/ce/de/product/188549/PGA-CPU-KUeHLKOeRPER-53-X-53-MM/SHOP_AREA_27805&promotionareaSearchDetail=005

Its a german page, but you can look at the value R th .... 7.3 K /W

This mean a CPU that use 1 Watt increase the temperature in the Heat sink of 7,3 Kelvin.

A CPU with a TDP of 10 Watt increase temperature of 73 Kelvin of the heatsink.To get real temperature in the CPU, you need add the Kelvin of the surrounding temperature too.
Airflow give better values.so there are use Fans on the Heatsink, when TDP get over 10 Watt.

I think the X1000 CPU can not have more TDP as 6 Watt CPU with this HeatSink and a pa6t as X1000 CPU is maybe then not possible.

maybe somebody that have see the X1000 can give more precise info about the HeatSink size of the X1000.

But because the X1000 Heatsink is not so high and have not so much free space, i think it have higher Kelvin /Watt as 7.3

the better cooler is of course this with lower Kelvin /Watt.

[unusedunused]

CPU: Intel Core 2 Duo 1596MHz
Compiler/OS: GCC4.3.0 20080428 (Red Hat 4.3.0-8)
Coremark/MHz: 3.421
Coremark: 5459.87
Threads: 2 pthreads
Benchmarked: 06/09/09

Of course, there are many more powerful intel processors than that, but compared to existing PPC processors found in OS4 boxes, that should be a big step up. Alas when comparing to freescale CPU's you can't find any G4 processors on that benchmark site. Or at least I didn't see any.

Usual benchmark caveats apply but the conditions used don't seem too different.

but look on this page there is another 1,6 GHZ Core 2 Duo and other that reach 5000 or 6000 Coremark  /MHZ.the value of 3400 this core 2 duo get you post is too low.maybe the test run only single thread


http://www.eembc.org/benchmark/reports/benchreport.php

Edit:

I look on the link you post and search for TDP, but i do not find this word.

when you search in google pa6t tdp you find the 25 Watt often.
here is also a link

http://www.linuxfordevices.com/c/a/News/Apple-acquires-Linuxfriendly-chip-vendor/

so because of the small heatsink of the X1000, it can be possible that there is no pa6t in.but what CPU is it then

maybe amcc design a new CPU with altivec and 64 bit.....
but thats really strange too, because normaly they announce such things years before series production start, so Hardware /Software developers have enough time to bring OS to this CPU.

also new design CPU use always DDR 3 i think

[unusedunused]

@Karlos
>it's just as possible the same applies to the PA6T benchmark. As there's only one such >benchmark for that, you've simply no way of knowing for sure.

thats possible too, but i think the guys that sell pa6 systems have an eye on it.
This benchmark is the one and only benchmark i find for pa6.and when potential buyers find it and its worser than real, then its bad for the pa6t sellers.

a better argument is when you tell, that in real world things may be diffrent, because its no real world bencmark.

but anyway i have no hope that the pa6t is a fast CPU so a price of 1800 Eur is something in reality and a revival is possible.

its only bad, that the most expensivest AOS solution make the most future hope and so most money flow to this projects.

I think the amiga have a better rivival when Hyperion accept that Haage and Partner change from PPC to X86 and support that too...

[unusedunused]

>Fixed it so it makes sense, but I find that highly unlikely. I have composed a sort of chart >outlining the most likely future of the three Amiga-Like OSes:

there is a easy rule.

Invest Money in a comparable in price and power Hardware and support it with Software and OS.

but this work not on amiga land.

lots of money and support is flow to bring AOS to high price PPC Hardware, but this much money help at least to get more apps.But the users must pay a high price for the Hardware and as can see not many do that.

On the other side there is AROS on X86.It have lots cheaper and much faster Hardware, but much much less money flow into it and support of it is few, and thats the Reason that AROS have not so much Apps as MOS or OS4.

So the ideal AOS for a revival have the Hardware as AROS and the support with money as MOS or OS4.

And when the guys with the money really think that it bring better future to support AOS that run only on the expensive HW, and there is no way insight to bring with PPC a more powerful and cheaper HW, then i really have no hope that this solution grow over 800-900 Fan Users.Thats way to few for a revival

and thats really optimistic, when look on the download count of the firefox alpha(a must have i think, because not many software come) i see only 551 downloads, and on aminet 77 and thats for 6 Years that OS4 can use.I am sure, in the past there are more OS4 users.

There was a user count in 2005 or maybe 2006, at amigaworld.net, there was over 850 users with aone microaone etc hw.

maybe a amiga Hardware can 20% more expensive with the same performance you get form current Hardware, but thats whats here on new hardware have such a high price that i dont think the user increase to reach at least the level of OS4 users in 2005.

I think a revival is only possible, when MOS or OS4 can produce new Hardware that is not higher as 20-30% in price by same performance as current systems.

If they can not i think all investet money does not help that more users ever use it.

[unusedunused]

>AFAIK PPC has never really been ahead and only very briefly on a few separate occasions >on a level playing field with x86 in terms of performance.

1997 with the 604e technoligy the cyberstorm PPC have, the PPC was much ahead 486 and Pentium 1/2 CPU really.I measure that out, and the Cyberstorm PPC have a lower price.

The wind of change come only in year 1999 2000 when Motorola do not the new G5 they have announce.Intel do the P3 and AMD do the Athlon.The Athlon performs the PPC 604 y out in performance /MHZ

Now we have 2010 and PPC still have as the fastest performance /MHZ technoligy the 604 technoligy that is called later as G4 and get altivec.

but now 2010 intel and AMD do more and what happen on PPC land except the 604 -> G4 technolgy was only some low performance /MHZ but low Power Chips as qori AMCC 4xx and pa6t.

PPC is cheap today only in low clockrates.but all over 1 GHZ is today expensive on PPC.

but still its no excuse that a AMCC 1 GHZ board is so expensive.See the efika, its sell last for 99$.

sure efika its lot underpowered, but today for 199$ should be a OS4/MOS system with a 1 GHZ PPC CPU and 1 gigabyte RAM board possible.

thats not too high price so maybe more users buy a MOS or OS4 system as a secondary system.

[unusedunused]

>IBM Xenon 3.2Ghz CPU which is compatible with PowerPC.

You mean the X-Box CPU.This is simular to Cell.the PPC is high clocked, but give less performance /MHZ because its a in order design.benchmarks say the 3,2 GHZ give simular performance as a 1,6 GHZ Athlon.

So should faster as PA6t on single thread, have three full cores, so i think you are right, this CPU is the most powerfull PPC.

I see in wikipedia there is a 45 nm version(so its also modernest chip develop for low power), but there are no TDP Values see and more bad it cant buy in low quantities.

>Also when comparing 68060 to x86 remember Motorola was the first to make a clock >doubled CPU for desktop computers. A 50mhz 060 runs internally at 100mhz.

the 68060 was also a great design.A Intel CPU need on 486 or these days Pentium 1 need 10-20 clockcycles for multiplication the 68060 have a hardware mutliplicator that can multiply 32*32 -> 32 bit result in 2 clock cycles.