Hades60-PMacG4-PS3-EeePC - comp.sys.atari.st

Message from discussion Hades60-PMacG4-PS3-EeePC

More options May 9, 7:27 am

Newsgroups: comp.sys.atari.st

From: OL <o...@lutece.net>

Date: Fri, 9 May 2008 04:27:24 -0700 (PDT)

Local: Fri, May 9 2008 7:27 am

Subject: Re: Hades60-PMacG4-PS3-EeePC

  Reply
  |
  Reply to author
  |
  Forward
  | Print
  | View thread
  | Show original
  | Report this message
  | Find messages by this author
  

Hello

> Hi,

> OL writes :

> > Francois LE COAT wrote :
> >> Reading the motherboard perfs result page (cleaned short version)

> > I just clean it look at <http://kronos.lutece.net/>

> Good.

> You may use an ATARI spreadsheet for charts like Texel, GemGraph,
> 3D-Calc, LDW Power, VIP, ATARI Works etc.

????!!!!! This software have an HTML output?!

> >> (a) PlayStation 3 Cell 3.2GHz ARAnyM 0.9.5 (no JIT) 259
> >> (b) Dual-G4 1.25GHz ARAnyM 0.9.5 Mac OS X (no JIT) 304
> >> (c) Hades60 60MHz freeMiNT 1.15.12 + N.AES 2.0 670
> >> (d) CT63 100MHz 16MHz bus 1194
> >> (e) Coldfire Evaluation Board V4e 200MHz 1375
> >> (f) EeePC Celeron 630MHz - WinXP - ARAnyM 0.9.6 JIT 1845
> >> (g) EeePC 701 Celeron 630MHz Debian ARAnyM 0.9.6 JIT 2523
> >> (h) EeePC Celeron 900MHz - WinXP - ARAnyM 0.9.6 JIT 2897
> >> (i) MacBook Pro 2.4GHz ARAnyM 0.9.6 Mac OS X JIT 10735

> >> Comments

> >> (a) and (b) will there be a JIT for ARAnyM / PPC one day ?

> > Perhaps yes, but you forget put an other interesting result
> > Athlon 2600+ Linux Aranym 0.9.5 (No JIT + MMU) 485
> > or better
> > MacBook 2.2 2.2GHz OSX Aranym 0.9.6 (No JIT) 734
> > that is far better than under Playstation or EeePC without JIT

> Of course yes. But the principle of a benchmark is not to compare
> perfs of systems in the worst case. One tries to take the best we
> can obtain from a computing configuration. Otherwise you don't mix
> results, and you speak about "best case", "worst case" or even in
> the "mean case" etc. You should compare what can be compared.

?????
Why? for example Athlon 2600+ Linux Aranym 0.9.5 (No JIT + MMU) 485
is comparable with Playstation 3 values, in this 2 case there is no
JIT, that is the only differences of this 2 computers.

> >> (c) and (d) we are not speaking of the same display perfs !

> > Yes this why there is 2 other tables for this look in web site.

> With ET4000 1Mb V-RAM graphic card Hades60 displays 800x608 64k colors.

And??? Notice the only difference with 32 bit and 16 bit for the
second table is only the transform from 24 bit to this display and in
fact transform from 24 to 16 bits is a little bit faster than
transform 24 to 32 bits.

> >> (e) partially or fully ATARI compatible ?

> > I not known any computer fully compatible with ATARI computer, Hades
> > was not, Milan too like Aranym when use fVDI

> 68k family is *binary* compatible, Coldfire is not, not 68k compatible.

Like Playstation nor PC nor Mac Computer is not!
Now big diffrence beetween a specific hardware computer is it is very
easy to have native software under it that is not easy at all under
Aranym. I wan't compile for V4e Tiny_gl.ldg I just change compil
target and thats all as I do for 68060 in past, it run, I'm able to do
this under Aranym but I should add code in Aranym itself and create a
new ldg that call native one and have a translation of address in
Atari world and host world, near 3 weeks of work to do this and
probably Aranym team will not agree to add all sort of source code in
there software!

> > but kronos work without any change on it. And this is the faster
> > computer for the VDI! (I have to put results)

> What kind of display system for the Coldfire Evaluation Board ?

This is an Radeon 7000 (128bit bus) on PCI with a lot of native
primitive call, so it's really fast, this sort of graphic card could
be put on your hades, result will be better than on the coldfire card
I think because fVDI kernel was not compil natively and memory access
on this evaluation board is slow in fact, this is not a real computer
but interesting for us.

> >> (f) and (g) GNU / Linux is much better than WinXP ?

> > Linux looks really better than WinXP for run Aranym, this 2 systems
> > can't be compare. Notice probably one was compil (linux version) for
> > EeePC processor while WinXP Aranym version was generic version (for
> > 386!) perhaps the problem is here for a part.

> Figures are there. We speak about what is published, not as it might be.

Oh no, you alway's transform results, even if it was not better on
Linux than for this case(I don't think so but I'm going to ask to
David to have this result for you), can you explain me why most user
will not use Aranym JIT when they can?!!!!! And result is quite
different:
EeePC Celeron 630MHz - WinXP - Aranym 0.9.6 (JIT) 1845
or
EeePC 701 Celeron 630MHz - Linux Debian - Aranym 0.9.6 (JIT) 2523
or
EeePC Celeron 900MHz - WinXP - Aranym 0.9.6 (JIT) 2897

Good reason? Compatibility is very good even with JIT (never be able
to find a software running without JIT that not run under JIT this is
possible probably but wich one?), speed is far better. I know only one
reason not use JIT, it's for developper's only that wan't fix bugs
using memory protection (like using NULL address under JIT not crash).
Do you use your Aranym with MMU and full system memory protection? Me
yes, but you not. For user there is no reason and if speed for you is
the most important, so EeePC is better than Playstation 3 but I think
we don't buy playstation or EeePC for Aranym only!

> >> (h) no GNU / Linux comparison ? you must use EeePC at 900MHz !

> > Why???? There is a test under EeePC at 630MHz with Linux and WinXP
> > thats enough

> We can't compare with EeePC at 900MHz / Linux. Maybe GNU / Linux
> can't control the frequency of the CPU at 900MHz, eventually ?

> >> (i) impressive and expensive ! the most powerful hardware / Linux

> > Yes

> We speak about an Intel Mac (Mac Book Pro) with a dual-boot x86-Linux
> Ubuntu & Mac OS X Leopard, don't we ?

> >> These result figures are cruels, but that's objective information.

> > Not cruels, in fact I use today Aranym without JIT and with MMU (I
> > could use JIT but I have no really interest to go faster), this is not
> > very very fast but it's quite enough to run Atari applications in good
> > conditions, thanks to video speed under fVDI, and probably this is the
> > most important. In my opinion Playstation 3 or EeePC are enough for
> > run Aranym. Have fun.

> Well, as you can see PlayStation 3 and Power Mac G4 are at the tail
> of the performance list. I wouldn't mind if it was a little quicker ;-)

No
EeePC Celeron 630MHz - WinXP - Aranym 0.9.6 (JIT) 1845
or
EeePC 701 Celeron 630MHz - Linux Debian - Aranym 0.9.6 (JIT) 2523
or
EeePC Celeron 900MHz - WinXP - Aranym 0.9.6 (JIT) 2897
That is the true, and we wait for JIT PPC in Aranym that possible
change this reality, and I will be very happy to add this results on
my web page, if you have this send me!

Now I am going to be a bit malicious, have you try to report results
to processor frequency?

At the same speed frequency EeePC under WinXP, Aranym for 386! should
have a value of 477 And Playstation 3 have only 259. And I wait for
results with Linux compil for EeePC!

> I do agree, it's great pleasure and gives great satisfaction indeed.

Yes and that is most important, and EeePC user have the same
satisfaction, strange isn't it!

And now look at this:
http://leeenux.blogspot.com/2008/02/eeepc-et-mac-os-x.html

Good video! I don't know if you can do this on Playstation 3 and if
your Apple computer can boot as fast as this. For an obsolete computer
this not so bad!

Reply Reply to author Forward