Correct.

TrevorD

_________________
No, I don't need no reason, I'm just breezin'

@olegil

Quote:

olegil wrote: @Boot_WB I would have that apophenia checked as well

GFY

_________________
Troll - n., A disenfranchised former potential customer who remains interested enough to stay informed and express critical opinions.
opp., the vast majority who voted silently with their feet.

@NutsAboutAmiga

Quote:

As calculated before the P5020 should be 36% faster (theoretically), and that’s not just for some application is it should be all applications.

PA6T - 2.2 DMIPS/MHz = 3960 DMIPS at 1.8 GHz
P5020 - 3.0 DMIPS MHz = 6000 DMIPS at 2.0 GHz

6000/3960 = 1.51

@pavlor


Also wishing for low end series from Acube .
AMP86290 Green Mamba - 2.00 DMIPS/MHz = 3000 DMIPS at 1.5 GHZ

Quote:

pavlor wrote: @NutsAboutAmiga Quote: As calculated before the P5020 should be 36% faster (theoretically), and that’s not just for some application is it should be all applications. PA6T - 2.2 DMIPS/MHz = 3960 DMIPS at 1.8 GHz P5020 - 3.0 DMIPS MHz = 6000 DMIPS at 2.0 GHz 6000/3960 = 1.51

Last edited by Spectre660 on 10-Jul-2013 at 05:30 PM.
Last edited by Spectre660 on 10-Jul-2013 at 05:29 PM.

_________________
Sam460ex : Radeon Rx550 Single slot Video Card : SIL3112 SATA card

@pavlor

It should e 8,800 DMIPS se column two in table on wikipedia.,

http://en.wikipedia.org/wiki/Instructions_per_second

it's 4.4 dmips per core.

So where do you get your number from?

Last edited by NutsAboutAmiga on 10-Jul-2013 at 07:22 PM.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

@CodeSmith
Quote:

(down to the sample device driver written in 68K assembler - hopefully that's not how Hyperion's writing them these days )

An updated example device driver written in C is already in the works.

Quote:

Ah well, at least it seems like some effort has been made to modernize it. If that's kept up and new OS features get documented as they get added, it should be enough I think.

There has not been just "some effort." There has been a very large effort by many dedicated people to improve the documentation and that will continue.

There are also standards (SANA-II) and registries (Amiga Hardware Manufacturer ID Registry) available on the AmigaOS Documentation Wiki.

It takes a lot of effort to update the AmigaOS documentation from Release 2 to Release 4 but it is happening. More volunteers to help work on the wiki are always welcome.

_________________
ExecSG Team Lead

@ssolie

I think you have fixed some important things in your short time as AOS Dev Team Lead.

Way to go Mr. Solie

Last edited by Hondo on 10-Jul-2013 at 07:51 PM.

_________________
On Planet Boing Trevor is God

@NutsAboutAmiga

Quote:

it's 4.4 dmips per core.

I never believe numebers I didn´t forge myself.

Quote:

So where do you get your number from?

PASemi public documents, eg.:

PWRficient Architecture in Critical Embedded Systems. Breakthrough Performance/Watt Multicore Processor. Peter Bannon VP of Architecture and Verification Bus & Boards, 2007. January 15th, 2007. (page 15)

Downloaded years ago, now I can´t find this document on the net. It says:

Performance
Estimated 4,400 Dhrystone MIPS per core

(I can convert this page to jpg and post it somewhere on net, but you can also simply trust me )


4.4 DMIPS/MHz would be simply too high value for CPU that in our benchmarks delivers lesser performance/MHz than G4 (2.3 DMIPS/MHz according to Freescale).

Edit: Just read wiki page you linked... Source document (article on eetimes says 8800 DMIPS at 2 GHz, not 1.8 as in wiki article), I think they meant combined performance of both core (4400 + 4400 DMIPS), it is common marketing practice to combine DMIPS value of all cores (Freescale does it too), but these numbers are misleading (Dhrystone is single core benchmark...).

Last edited by pavlor on 10-Jul-2013 at 08:26 PM.

@Hondo

Very true and M. Solie is a very active contributor seeing the changes log.

Kamelito

@pavlor

Quote:

PA6T - 2.2 DMIPS/MHz = 3960 DMIPS at 1.8 GHz P5020 - 3.0 DMIPS MHz = 6000 DMIPS at 2.0 GHz 6000/3960 = 1.51

Good. In theory, Cyrus dual core will be 50% faster then PA Semi,
and high end Cyrus up to dual that. Add DDR3 RAM, and that is a nice advancement,
a space yet to be used. Some Altivec optimized apps would help G4 and PA Semi users to catch up a bit

More important, there will be a new, faster and probably more available AmigaOne,
that could be promoted as X1000 deserved. Now OS needs to catch up, or it with will be PPC Linux dream machines

_________________
Future Acube and MOS supporter, fi di good, nothing fi di unprofessionals. Learn it harder way!

@vox

Quote:

Good. In theory, Cyrus dual core will be 50% faster then PA Semi,

I think P5020 could be even faster than that, 2.2 DMIPS value is probably over-estimated in case of PA6T. Compare P5020 to G4:

G4 - 2.3 DMIPS MHz = 3266 DMIPS at 1.42 GHz
P5020 - 3.0 DMIPS MHz = 6000 DMIPS at 2.0 GHz

6000/3266 = 1.83

PA6T was in many benchmarks as fast as G4 1.42 GHz. P5020 will be probably faster than PA6T even in most AltiVec depending tasks. Still only theory, until boards are shipped and tested by real applications.

So, any idea what the lowest range in the price for this new line might be? And the uppermost?

Last edited by NomadOfNorad on 10-Jul-2013 at 11:10 PM.

_________________
"I love peacenicks, they're so easy to conquer." --Ivan J Ironfist, the Dictator

@thread,

Lets not get too worked up regarding how fast the Cyrus boards will be and wait for real world performance testing "AFTER" they are released. It is enough to know that new hardware appears to be on it's way, plus the CPU used is not already discontinued and from initial reports, the new systems will be both faster and cheaper than the existing X1000's.

Remember how many users touted the X1000 as a "G5 Killer", which caused too much renewed in-fighting and then those making the claims had the real performance of the X1000 thrown back in their faces when the testing results did not live up to their claims. I would not want to see a repeat of that again (because you know there are some members who hang around here just to cause trouble and humiliation, when ever the opportunity arises).

Just be happy that Trevor is going forward with new hardware development (if PPC hardware is your cup of tea), and wait to see actual benchmark tests to determine just how fast these new boards will be.

My 2 Cents.

_________________
Amiga! The computer that inspired so many, to accomplish so much, but has ended up in the hands of . . . . . . . . . .

No AltiVec ?!

_________________

José

@pavlor

The G4 comes in many models from 7400 to 7448.

one onces used in the AmigaONE was the 7445 and 7455.
at 800 Mhz and 933Mhz.

http://www.ghs.com/news/220722e.html

2583 MIPS / 1Ghz = 2.5 Mips/Mhz.

some who had to fix there AmigaONE's use 7447 like I do.
It was locked at 1Ghz (and 1.2 Ghz I think)

(2300 DMIPS)

that last mac's used the 7448.

can be clocked up to 1.7 Ghz according to Freescale.

2.3 Mips/Mhz.

1.7Ghz = 3910 DMIPS.

http://www.freescale.com/webapp/sps/site/prod_summary.jsp?code=MPC7448

It funny but 7455 looks to be faster then newer 7448 cpus

Last edited by NutsAboutAmiga on 11-Jul-2013 at 05:30 PM.
Last edited by NutsAboutAmiga on 11-Jul-2013 at 01:41 AM.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

@ssolie

Of course it's an ongoing process and all involved are dedicated to it, but the current focus seems to be on industry partners (standards and registries) rather than 3rd party development (beginner's guides). I imagine that it's inevitable; if there's no good hardware support in the OS there won't be any beginners to write guides for. Surely something could be done though - there must be talented prose writers in the community who would be willing to work with the devs to produce a good set of manuals while the devs get on with dev work.

Incidentally, I must agree with what others have said - OS4.x owes a lot to you. Thank you.

@ssolie

Quote:

An updated example device driver written in C is already in the works.

Awesome!

I'm very interested in PCI (and PCI-Express) bridges, which are a problem with XE uboot, I cannot see devices on the other end of my PCI-Express bridge slot adaptor.

_________________
All glory to the Hypnotoad!

@amigadave

Quote:

Remember how many users touted the X1000 as a "G5 Killer", which caused too much renewed in-fighting and then those making the claims had the real performance of the X1000 thrown back in their faces when the testing results did not live up to their claims. I would not want to see a repeat of that again (because you know there are some members who hang around here just to cause trouble and humiliation, when ever the opportunity arises). Just be happy that Trevor is going forward with new hardware development (if PPC hardware is your cup of tea), and wait to see actual benchmark tests to determine just how fast these new boards will be.

Well, PA Semi has some newer technologies and faster RAM then G5, but at the same time never went to so high frequencies and was ment to be "power efficient"

Surely, speed hype is relative, especially when we know that its up to OS to properly use multi cores, yet to be seen.

Surely, there is a speedy future ahead, demanding even more money.
But that is way better then having old hardware or no hardware at all.

X1000 for example is something at least comparable to what PC users have, and remains yet to be used at full speed. Now is understandable why OS 4.2 has such high end goals as breaking RAM barrier and introducing multi core, which will be more of benefit of X1000 users and beyond. Without such high goals, those new machines would be useless.

_________________
Future Acube and MOS supporter, fi di good, nothing fi di unprofessionals. Learn it harder way!

@amigadave

Well, some of the problem is that very few people realized just how low performance a "G5" has when compared to a "G4", clock for clock. Sort of like Pentium 4 vs Pentium 3, there's a reason why Core and Pentium M were P3 based rather than P4

_________________
This weeks pet peeve:
Using "voltage" instead of "potential", which leads to inventing new words like "amperage" instead of "current" (I, measured in A) or possible "charge" (amperehours, Ah or Coulomb, C). Sometimes I don't even know what people mean.

@NutsAboutAmiga

Quote:

It funny but 7455 looks to be faster then newer 7448 cpus

If my memory serves me right, Freescale mentions 2.3 DMIPS/MHz for all G4 CPUs (at least since 7450).

eg.:
http://cache.freescale.com/files/32bit/doc/fact_sheet/PPCSALESFACT.pdf

Note: Value for 7410 seems to be strange, but who knows.


7448 has bigger L2 cache than 7455, but 7455 has also L3 cache. Performance will be probably application depending.