Click Here
home features news forums classifieds faqs links search
6071 members 
Amiga Q&A /  Free for All /  Emulation /  Gaming / (Latest Posts)
Login

Nickname

Password

Lost Password?

Don't have an account yet?
Register now!

Support Amigaworld.net
Your support is needed and is appreciated as Amigaworld.net is primarily dependent upon the support of its users.
Donate

Menu
Main sections
» Home
» Features
» News
» Forums
» Classifieds
» Links
» Downloads
Extras
» OS4 Zone
» IRC Network
» AmigaWorld Radio
» Newsfeed
» Top Members
» Amiga Dealers
Information
» About Us
» FAQs
» Advertise
» Polls
» Terms of Service
» Search

IRC Channel
Server: irc.amigaworld.net
Ports: 1024,5555, 6665-6669
SSL port: 6697
Channel: #Amigaworld
Channel Policy and Guidelines

Who's Online
15 crawler(s) on-line.
 148 guest(s) on-line.
 1 member(s) on-line.


 DiscreetFX

You are an anonymous user.
Register Now!
 DiscreetFX:  3 mins ago
 Gunnar:  8 mins ago
 saimo:  13 mins ago
 amigakit:  31 mins ago
 OldFart:  32 mins ago
 _ThEcRoW:  51 mins ago
 NutsAboutAmiga:  1 hr 19 mins ago
 retrofaza:  1 hr 19 mins ago
 Rob:  1 hr 20 mins ago
 OlafS25:  1 hr 42 mins ago

/  Forum Index
   /  Amiga OS4.x \ Workbench 4.x
      /  RageMem benchmarks between OS4 machines
Register To Post

Goto page ( Previous Page 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 Next Page )
PosterThread
Seiya 
Re: RageMem benchmarks between OS4 machines
Posted on 19-Mar-2011 23:10:10
#21 ]
Super Member
Joined: 19-Aug-2006
Posts: 1473
From: Italia

i have afforded to make a graph about the first post.


cpu test:


L1 Cache:


RAM:


Video Bus:

_________________

 Status: Offline
Profile     Report this post  
klx300r 
Re: RageMem benchmarks between OS4 machines
Posted on 20-Mar-2011 4:34:43
#22 ]
Elite Member
Joined: 4-Mar-2008
Posts: 3833
From: Toronto, Canada

@Seiya

thanks for the charts

_________________
____________________________
c64-2sids, A1000, A1200T-060@50(finally working!),A4000-CSMKIII
! My Master Miggies- Amiga 1000 & AmigaOne X1000 !
mancave-ramblings
X1000 I BELIEVE

 Status: Offline
Profile     Report this post  
DAX 
Re: RageMem benchmarks between OS4 machines
Posted on 20-Mar-2011 11:30:23
#23 ]
Elite Member
Joined: 30-Sep-2009
Posts: 2790
From: Italy

@Seya
I hope that the numbers in the first graph are not MIPS, as in that case G4 values are utter bullcrap.
A Peg2 so clocked would score 2911,8 according to FreeScale, Blender benchmarks corroborate this (difference in redering times).
Also G4 1400 would score 3220. (2.3 Mips x Mhz).

By the way, since the most common Mhz count found on the Peg2 would be 1000Mhz the number of MIPS would then be 2300 (2.3x1000), even less for A1s which have demonstrated a large tendecy to "fry" at speed over 1Ghz.

Last edited by DAX on 20-Mar-2011 at 11:44 AM.

_________________
SamFlex Complete 800Mhz System + AmigaOS 4.1 Update 4
Amiga 2000 DKB 2MB ChipRam GVP G-Force040 Picasso 2 OS3.9 BB2
AmigaCD 32

 Status: Offline
Profile     Report this post  
amitv 
Re: RageMem benchmarks between OS4 machines
Posted on 20-Mar-2011 12:25:32
#24 ]
Regular Member
Joined: 23-Oct-2006
Posts: 346
From: Unknown

@Seiya

Thx

 Status: Offline
Profile     Report this post  
Antique 
Re: RageMem benchmarks between OS4 machines
Posted on 20-Mar-2011 13:12:52
#25 ]
Cult Member
Joined: 8-Jun-2005
Posts: 887
From: Norway

@Seiya

Nice graph.

_________________
I'm an antique. Don't light my fuse

 Status: Offline
Profile     Report this post  
guruman 
Re: RageMem benchmarks between OS4 machines
Posted on 20-Mar-2011 13:59:17
#26 ]
Regular Member
Joined: 20-Jun-2007
Posts: 133
From: Padova, Italy

@Seiya

Nice graphs.

To add a bit more info, some tests have been performed by Ikir Sector users (italian forum) using MEncoder and encoding the same short video using the same command line.

Sam440ep Flex 800:
Pos: 4.0s 101f (100%) 10.71fps Trem: 0min 0mb A-V:0.000 [911:0]

Sam440ep 667:
Pos: 4.0s 101f (100%) 8.16fps Trem: 0min 0mb A-V:0.000 [911:0]

Sam460ex:
Pos: 4.0s 101f (100%) 14.44fps Trem: 0min 0mb A-V:0.000 [911:0]

AmigaOne G4 1GHz:
Pos: 4.0s 101f (100%) 30.01fps Trem: 0min 0mb A-V:0.000 [911:0]

Mac mini 1.5GHz (MorphOS):
Pos: 4.0s 101f (100%) 47.13fps Trem: 0min 0mb A-V:0.000 [911:0]

PowerMac G4 1.42GHz (MorphOS):
Pos: 4.0s 101f (100%) 48.56fps Trem: 0min 0mb A-V:0.000 [911:0]

Kind regards,
Andrea

 Status: Offline
Profile     Report this post  
Varthall 
Re: RageMem benchmarks between OS4 machines
Posted on 20-Mar-2011 14:20:43
#27 ]
Super Member
Joined: 17-Feb-2004
Posts: 1559
From: Up Rough

@guruman

I'd like the first three benchmarks to be repeated, in case they were run from an JXFS partition. On my machine on the first run the filesystem seems to be caching/reorganising data blocks, as it blocks MEncoder for a couple of seconds before it starts writing to the disk, and the results are 30-40% lower because of this. Also, the same holds true if they have been run from RAM disk, on my system writes to RAM disk are a bit slower than the ones on my partitions.

The benchmark anyway should be more meaningful if run using a bigger test video, with the one I provided the resulting speed was reached only at the last 1-2 seconds of the encoding.

Last but not least, the test between the MOS and OS4 version of MEncoder should be done using the same SVN revision of the code on both OSes: I have seen significant speed differences between two SVN revisions, one released last October, the other one last December.

Varthall

Last edited by Varthall on 20-Mar-2011 at 02:23 PM.

_________________
AmigaOne XE - AmigaOS 4.1 - Freescale 7457 1GHz - 1GB ram

 Status: Offline
Profile     Report this post  
DAX 
Re: RageMem benchmarks between OS4 machines
Posted on 20-Mar-2011 15:23:28
#28 ]
Elite Member
Joined: 30-Sep-2009
Posts: 2790
From: Italy

@Varthall
It would be also wise to take comparisons between fully realized/optimized systems VS beta systems with a grain of salt (many values will change for the better in the 460EX case according to both Max Tretene and Hans De Ruiter).

Still wondering how those "3-Mips per Mhz" results for the G4 were obtained (in the first graph), as that is not the case with those CPUs (2.3 Mips per Mhz is correct).

Last edited by DAX on 20-Mar-2011 at 04:05 PM.
Last edited by DAX on 20-Mar-2011 at 03:25 PM.

_________________
SamFlex Complete 800Mhz System + AmigaOS 4.1 Update 4
Amiga 2000 DKB 2MB ChipRam GVP G-Force040 Picasso 2 OS3.9 BB2
AmigaCD 32

 Status: Offline
Profile     Report this post  
Crisot 
Re: RageMem benchmarks between OS4 machines
Posted on 20-Mar-2011 16:13:01
#29 ]
Regular Member
Joined: 25-Jan-2004
Posts: 163
From: France

@DAX

Hi DAX.

You are true and wrong at the same time.

The G4 is able to do 2.3 DMIPS per Mhz. Not MIPS, but DMIPS. DMIPS means Dhrystone Mips. Dhrystone is an old benchmark from 1984, still "quite official" and used by manufacturers.

So the G4 deliver 2.3 MIPS/MHZ with this benchmark, and only this benchmark.

With other algorythms, the G4 can deliver more or less MIPS, depending of the optimisation level of the code.

RageMem just gives "maximum reachable MIPS". A G4 can deliver 3 MIPS/MHZ, but of course only in some restrictive conditions. This power is possible with some optimisation for some specific programs.

It's not representative of "every day use", but it's in the readme

_________________
AmigaOne XE --- 7455A@1400 --- Radeon 9000 Pro --- RAM 2GB --- SSD 40GB

 Status: Offline
Profile     Report this post  
Seiya 
Re: RageMem benchmarks between OS4 machines
Posted on 20-Mar-2011 16:39:53
#30 ]
Super Member
Joined: 19-Aug-2006
Posts: 1473
From: Italia

@DAX

the value are from marko's benchmark

MAX MIPS: 465 //a4000t
MAX MIPS: 1598 //sam440ep-flex
MAX MIPS: 2332 //sam460ex
MAX MIPS: 3797 //peg2
MAX MIPS: 4194 //a1-xe

_________________

 Status: Offline
Profile     Report this post  
DAX 
Re: RageMem benchmarks between OS4 machines
Posted on 20-Mar-2011 16:46:18
#31 ]
Elite Member
Joined: 30-Sep-2009
Posts: 2790
From: Italy

@Crisot
Thanks for the explanation (they should specify DMips, as it doesn't say so anywhere on FreeScale website).

@Seya
don't kill the messenger (you in this case ).

Still quite misleading, those Mhz values are used by no-one basically (or only a few untill they fried or will fry their systems).
Mentioning a 1.4 Ghz A1 is like mentioning a liquid helium cooled overclocked Sam460, pointless as not widely available (if at all).

Last edited by DAX on 20-Mar-2011 at 04:54 PM.
Last edited by DAX on 20-Mar-2011 at 04:50 PM.

_________________
SamFlex Complete 800Mhz System + AmigaOS 4.1 Update 4
Amiga 2000 DKB 2MB ChipRam GVP G-Force040 Picasso 2 OS3.9 BB2
AmigaCD 32

 Status: Offline
Profile     Report this post  
Seiya 
Re: RageMem benchmarks between OS4 machines
Posted on 20-Mar-2011 18:33:13
#32 ]
Super Member
Joined: 19-Aug-2006
Posts: 1473
From: Italia





_________________

 Status: Offline
Profile     Report this post  
kas1e 
Re: RageMem benchmarks between OS4 machines
Posted on 20-Mar-2011 18:41:32
#33 ]
Elite Member
Joined: 11-Jan-2004
Posts: 3549
From: Russia

@Seiya

Pretty good and interesting graphs. Will be nice if you can add peg2/1ghz and a1/1ghz (as they the most popular, in compare with peg2/1.2 and a1-/1.4.

btw, sam460/1550 imho wrong ? (should be 1150)

Last edited by kas1e on 20-Mar-2011 at 06:42 PM.

_________________
Join us to improve dopus5!
zerohero's mirror of os4/os3 crosscompiler suites

 Status: Offline
Profile     Report this post  
Rose 
Re: RageMem benchmarks between OS4 machines
Posted on 20-Mar-2011 18:55:48
#34 ]
Cult Member
Joined: 5-Nov-2009
Posts: 982
From: Unknown

@Varthall
Quote:
Last but not least, the test between the MOS and OS4 version of MEncoder should be done using the same SVN revision of the code on both OSes: I have seen significant speed differences between two SVN revisions, one released last October, the other one last December.


You could as well demand that test should be done 3rd monday of may on full moon while tester holds three eagle feathers on left hand and a dead squirrel in right one.

Test should be done on what's available in wild. If some operating system doesn't have new version it shouldn't mean that other OS's test result should be cripled. Otherwise there isn't much sense in benchmarkking. Nothing prevents you to make your own benchmark where you use ancient versions on other os to get results of your liking.

Thought you have a good point that test video should be a lot of bigger.

 Status: Offline
Profile     Report this post  
Varthall 
Re: RageMem benchmarks between OS4 machines
Posted on 20-Mar-2011 19:41:24
#35 ]
Super Member
Joined: 17-Feb-2004
Posts: 1559
From: Up Rough

@Seiya

As I said, those values are not guaranteed to be realistic - I don't know how the test on the various Sam boards have been executed.

@Rose
Quote:

You could as well demand that test should be done 3rd monday of may on full moon while tester holds three eagle feathers on left hand and a dead squirrel in right one.

Test should be done on what's available in wild. If some operating system doesn't have new version it shouldn't mean that other OS's test result should be cripled. Otherwise there isn't much sense in benchmarkking. Nothing prevents you to make your own benchmark where you use ancient versions on other os to get results of your liking.

IMHO, tests should be done correctly in order to be meaningful. If a test is done on two different system, the same benchmark should be used.
Anyway, to accomplish this it's much simpler than you think. All is needed is to get what's the SVN version of the MEncoder MOS port, grab the same version from the MPlayer SVN and compile it on OS4.

Varthall

_________________
AmigaOne XE - AmigaOS 4.1 - Freescale 7457 1GHz - 1GB ram

 Status: Offline
Profile     Report this post  
Seiya 
Re: RageMem benchmarks between OS4 machines
Posted on 20-Mar-2011 20:30:23
#36 ]
Super Member
Joined: 19-Aug-2006
Posts: 1473
From: Italia

@Varthall

ok ok
it's not official charts :)
it should be more easy to read tests.

_________________

 Status: Offline
Profile     Report this post  
marko 
Re: RageMem benchmarks between OS4 machines
Posted on 21-Mar-2011 1:34:22
#37 ]
Super Member
Joined: 17-Dec-2007
Posts: 1816
From: Gothenburg, THE front side of Sweden ;), (via Finland), EU

@Seiya

Nice graphs!

_________________
AmigaOS 4.1 FEu2 on Sam440ep-flex 800MHz 1GB RAM
C128, A500+, A1200, A1200/40, AmigaForever 2008+09+16, 5 x86/x64 boxes
Still waiting (or dreaming) for the Amiga revolution...
m4rko.com/AMIGA

 Status: Offline
Profile     Report this post  
Fab 
Re: RageMem benchmarks between OS4 machines
Posted on 21-Mar-2011 1:57:58
#38 ]
Super Member
Joined: 17-Mar-2004
Posts: 1178
From: Unknown

@Varthall

my latest public build of mplayer and mencoder is based on r32620. I have also a more recent build, but didn't notice any meaningful difference in speed, anyway.

 Status: Offline
Profile     Report this post  
Rose 
Re: RageMem benchmarks between OS4 machines
Posted on 21-Mar-2011 9:09:48
#39 ]
Cult Member
Joined: 5-Nov-2009
Posts: 982
From: Unknown

@Fab

Quote:
my latest public build of mplayer and mencoder is based on r32620. I have also a more recent build, but didn't notice any meaningful difference in speed, anyway.


There's allways something wrong with test untill "right platform" wins.

 Status: Offline
Profile     Report this post  
Varthall 
Re: RageMem benchmarks between OS4 machines
Posted on 21-Mar-2011 9:48:17
#40 ]
Super Member
Joined: 17-Feb-2004
Posts: 1559
From: Up Rough

@Fab
Check the results from this thread. That was on a MEncoder version built from sources downloaded on 1st November. The "MPEG-4 encoder - libavcodec" is the same as the test above. The results on the same machine, using the Altivec version:

Old MEncoder:
Pos: 4.0s 101f (100%) 9.31fps Trem: 0min 0mb A-V:0.000 [911:0]

Latest MEncoder:
Pos: 4.0s 101f (100%) 30.01fps Trem: 0min 0mb A-V:0.000 [911:0]

It might still be the "first launch with filesystem journalling" problem, although I remember I have seen significant differences at first launch with both versions.

@Rose
Quote:
There's allways something wrong with test untill "right platform" wins.

No, I can assure you that that was not my intention. I'm not interested in following any flame war between OSes, I have better ways to spend my free time.

Varthall

_________________
AmigaOne XE - AmigaOS 4.1 - Freescale 7457 1GHz - 1GB ram

 Status: Offline
Profile     Report this post  
Goto page ( Previous Page 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 Next Page )

[ home ][ about us ][ privacy ] [ forums ][ classifieds ] [ links ][ news archive ] [ link to us ][ user account ]
Copyright (C) 2000 - 2019 Amigaworld.net.
Amigaworld.net was originally founded by David Doyle