24-Apr-2024 07:31 GMT.		 UNDER CONSTRUCTION
Anonymous, there are 80 items in your selection (but only 30 shown due to limitation) [1 - 50] [51 - 80]
[Web] AmigaOS4 Petunia benchmarksANN.lu
Posted on 18-Aug-2003 11:11 GMT by Peter Gordon80 comments
View flat
View list
The petunia website has been updated with benchmarks from a recent version. Also, it is claimed that "This version is already converted to AmigaOS4, and running on the native PowerPC system".

Petunia homepage
AmigaOS4 Petunia benchmarks : Comment 51 of 80ANN.lu
Posted by Anonymous on 18-Aug-2003 19:40 GMT
give them the bait, and the losers, like little brain dead fishes, will bite
AmigaOS4 Petunia benchmarks : Comment 52 of 80ANN.lu
Posted by Phill on 19-Aug-2003 08:02 GMT
In reply to Comment 49 (Anonymous):
> The page states that the emulator runs under 4.0 as an application, but is
> not integrated in exec.

I would imagine thats just a case of patching it into dos.library LoadSeg() or something similar. Like the LoadElf stuff you had to do before.

Phill
AmigaOS4 Petunia benchmarks : Comment 53 of 80ANN.lu
Posted by BrianK on 19-Aug-2003 10:03 GMT
In reply to Comment 48 (Anonymous):
Also, here's some more info for people wanting to compare the G3 and G4 processor. (I hope people realize G3 and G4 is Apple's name for the processor.)

http://www.barefeats.com/g400b.html is the link this is from..

1) Apple B&W G3 400
2) Apple B&W G4 400
3) Apple Yikes G4/400

Photoshop Rotate(sec): 1: 7.1 , 2: 5.2, 3: 5.0
Quake 2(Frames/sec): 1:19.9 , 2: 19.7, 3:20.9
CPU Rating (higher# better): 1300, 1284, 1283
FPU Rating (higher# better): 1309, 1500, 1478

Thus, you can see under these tests the machines are very equivalent. MacOS 8.6 was used (it's a 3 year old page).

Here's what I want to see! ALTIVEC Usage
Photoshop Lighting Effects on a 17MB File: 18.5, 6.7 , 6.3.

Amiga -- Please, please include altivec optimizations on AmigaOS4.1. (Since it appears they won't be doing this for 4.0).
AmigaOS4 Petunia benchmarks : Comment 54 of 80ANN.lu
Posted by BrianK on 19-Aug-2003 10:10 GMT
In reply to Comment 53 (BrianK):
Performance again...

One last pages then I'm ending this. It's just some people wanted to see what the comparisions were but didn't go to a search engine and track them down. So, I'm hopefully being helpful to those people.

http://www.macspeedzone.com/archive/hardware/chipcomparisons/G3vsG4.html

G3 v G4 : SpecINT 21.4 v 21.4, SpecFPU 13.8 v 20.4

Thus, if you're using graphic intensive programs, such as ImageFX, that use lots of FPU calls you should see a performance enhancement on the G4 over the G3. However, for most stuff you won't, assuming you're using the same speed G3 and G4 CPU.
AmigaOS4 Petunia benchmarks : Comment 55 of 80ANN.lu
Posted by Don Cox on 19-Aug-2003 11:44 GMT
In reply to Comment 42 (Lando):
"It also runs WarpUP, PowerUP programs, whereas Amithlon is restricted to 68k."

Are there any such programs that are not available also as 68k versions? I don't know of any.
AmigaOS4 Petunia benchmarks : Comment 56 of 80ANN.lu
Posted by Alfred Schwarz on 19-Aug-2003 13:32 GMT
In reply to Comment 55 (Don Cox):
> Are there any such programs that are not available also as 68k versions? I don't know of any.

Games, of course, WipeOut for example.

Ciao, Alfred
AmigaOS4 Petunia benchmarks : Comment 57 of 80ANN.lu
Posted by tarbos on 19-Aug-2003 23:55 GMT
In reply to Comment 48 (Anonymous):
>500Mhz G4 Apple Cube:100MHz Bus and 1MB L2 cache. >500Mhz G3 Apple Ibook: 66MHz Bus and 256KB L2 cache.No surprise the SPEC numbers are quite different,you are comparing apples and pears here.
AmigaOS4 Petunia benchmarks : Comment 58 of 80ANN.lu
Posted by tarbos on 20-Aug-2003 00:23 GMT
In reply to Comment 44 (ehaines):
You have no e-mail address, but if you like you can get it fromhttp://fachschaft.physik.uni-karlsruhe.de/~frank/packages/unstable/l/lm/lmbenchand docs fromhttp://fachschaft.physik.uni-karlsruhe.de/~frank/packages/unstable/l/lm/lmbench-docas Debian package.
AmigaOS4 Petunia benchmarks : Comment 59 of 80ANN.lu
Posted by BrianK on 20-Aug-2003 11:43 GMT
In reply to Comment 57 (tarbos):
>500Mhz G4 Apple Cube:
100MHz Bus and 1MB L2 cache.

>500Mhz G3 Apple Ibook:

66MHz Bus and 256KB L2 cache.
No surprise the SPEC numbers are quite different,
you are comparing apples and pears here.

*****

I'd say the spec numbers are pretty close even accounting for the Bus and Cache issues. I was trying to do the best to find you info on performance that was around the web. (It's called use a search engine.)

As for L2 Cache that's on the processor and G3 and G3 have different amounts nothing someone out side of Motorola can do about that one.
As for bus speeds -- see later emails appended to this thread such as 53 and 54. Apple B&W G3/400 compared to Apple B&W G4/400. Integer performance is within 3% of each other. FPU performance for the G4 is much improved, by about 30%.

Thus, if you're comparing getting an AmigaOne G3/800 w/ AmigaOne G4/800
you're not going to see much in the way of Interger performance gain
probably close to 0. But, if you're using ImageFX for some heavy
FPU manipulation of your graphics the G4 is going to outperform the
G3 for you.

What the heck you're moving from a 604e @ 233Mhz, best case, to a G3 or 4 at 800Mhz. You're going to see a 3-4 fold performance increase. It appears the systems are about 8-10% difference in cost (for the AmigaOne) decide which one makes the best sense for you and go for it. It'll run circles around your Amiga w/ PPC.
AmigaOS4 Petunia benchmarks : Comment 60 of 80ANN.lu
Posted by tarbos on 20-Aug-2003 14:13 GMT
In reply to Comment 30 (BrianK):
>500Mhz G4 Cube: SpecINT:23.7, SpecFP:22.7>500Mhz G3 Ibook: SpecINT:21.2, SpecFP:12.3 900MHz IBM 750FX G3: 40,4 SPECint95, 21,9 SPECfp951000MHz IBM 750GX G3: 52 SPECint95, 30,2 SPECfp95 Funny thing is, in raytracing benchmarks like Cinebench, G4's "stronger" FPU doesnothing to accelerate it - here G4 shows inferior performance in comparison to G3!
AmigaOS4 Petunia benchmarks : Comment 61 of 80ANN.lu
Posted by BrianK on 20-Aug-2003 15:52 GMT
In reply to Comment 60 (tarbos):
900MHz IBM 750FX G3: 40,4 SPECint95, 21,9 SPECfp95
1000MHz IBM 750GX G3: 52 SPECint95, 30,2 SPECfp95

Funny thing is, in raytracing benchmarks like Cinebench, G4's "stronger" FPU does nothing to accelerate it - here G4 shows inferior performance in comparison to G3!

****
I think you copied something wrong. You don't seem to list a G4 here. I believe that 750FX and 750GX are both G3 chips. The G4 series is a 7450 or 7455. I can track down the number if you like.
AmigaOS4 Petunia benchmarks : Comment 62 of 80ANN.lu
Posted by Anonymous on 20-Aug-2003 15:57 GMT
In reply to Comment 61 (BrianK):
Check out
http://www.vesalia.de/

The G3 is Apple's name for the chip or what Motorola calls the 750FX as used in the Amiga One.

The G4 is Apple's name for the chip or what Motorola calls the 7455 as used in the Amiga One.

As for the PegasOS side of things as listed on http://www.vesalia.de/
They use the G3 or what Motorola calls the 750CXe.
They use the G4 or what Motorola calls the 7450.

Thus when you were comparin
AmigaOS4 Petunia benchmarks : Comment 63 of 80ANN.lu
Posted by BrianK on 20-Aug-2003 16:16 GMT
In reply to Comment 62 (Anonymous):
Just to add confusion there are different G3 and G4 chips in the system. Apple doesn't change the name as the chips change.

Apple-Motorola-IBM all have an alliance on PowerPC chips

750 series is what Apple calls the G3.
750FX appears to be in the Apple Ibook and the AmigaOne
750CX appears to be in the Apple IMac and the PegasOS
750GX doesn't appear to be in anything it seems it's sampling but not in production until late 2003. Rumored to be in the next Ibook speed bump.
--- It's in neither AmigaOne nor PegasOS. But rumored to be the next iBook speed up CPU.

745x series is what Apple calls the G4.
7450 appears to be in the Apple G4 Macs, such as the dual 800Mhz, and in the PegasOS.
7455 appears to be in the Apple G4 Macs, such as the dual 1Ghz, and in the AmigaOne.

Apple's Dual 800Mhz 7450 has SpecINT:241 and SpecFPU:146
Apple's Dual 1Ghz 7455 has SpecINT:307 and SpecFPU:188

Thus the 7455 appears to be a minor improvement over the 7450 in FPU processing but in daily use you won't see any pratical difference between an 800Mhz 7450 or 800Mhz 7455.


*****
The overall point really is either G3 or G4 of any lineage is going to kick ass all over my A4000T w/ 604e@233Mhz and 060@50Mhz. But, with a G4 you will see some improvements over the G3 in FPU processing.
AmigaOS4 Petunia benchmarks : Comment 64 of 80ANN.lu
Posted by Anonymous on 20-Aug-2003 16:56 GMT
In reply to Comment 63 (BrianK):
Differences

G3:PegasOS: 750CXe, codename Anaconda, 400-600Mhz, 256K L2 cache, .18 micron process w/ 6 Watts @ 600Mhz
G3:AmigaOne: 750GX, Sahara, 500-900Mhz, 512K L2 cahce, .13 micron process w/ 5.4 Watts @ 800Mhz

G4:PegasOS: 7450, G4+ or G4e, 533-867Mhz, .18 micron process w/ 26 Watts @733Mhz
G4:AmigaOne: 7455, Apollo, 733-1250Mhz, .18 micron process w/ 21.3 Watts @ 1Ghz.

All chips have 64K L1 Cache.
AmigaOS4 Petunia benchmarks : Comment 65 of 80ANN.lu
Posted by David Scheibler on 20-Aug-2003 18:03 GMT
In reply to Comment 64 (Anonymous):
Pegasos G4 is based on 7447.
AmigaOS4 Petunia benchmarks : Comment 66 of 80ANN.lu
Posted by Anonymous on 20-Aug-2003 20:13 GMT
In reply to Comment 65 (David Scheibler):
Interesting.

http://www.magneticsystemsnyc.com/Pegasos.htm backs you up as a 7447.
http://www.pegasosppc.com/tech_specs.php states it can handle up to a 7450. But their image appears to show a 7447. (http://www.pegasosppc.com/images/g4/g4_bill.jpg) and their PDF (http://www.pegasosppc.com/files/g4.pdf) talks about the 7447.

http://www.vesalia.de/ shows 7450.
http://www.compcity.nl/pegasos.htm is showing up to a dual 7450.

http://www.pegasos-czech.com/stranky/specifikace/specifikace1.htm and a few other sites list a G4 but not which model.

Thanks for the clarification on that.


Note to PegasOS... You have some vendors of your product advertising the wrong CPU model number. Also, you may have a couple areas of your own website to clean up.
AmigaOS4 Petunia benchmarks : Comment 67 of 80ANN.lu
Posted by David Scheibler on 20-Aug-2003 21:06 GMT
In reply to Comment 66 (Anonymous):
See this screenshot of MorphOS using the G4 7447 1GHz CPU card:

http://batman.jypoly.fi/~saku/yhdistys/tapahtumat/assembly03/kuvat/Assembly03_Pegasos_G4_Specs.jpg

But it's true that some websites have incorrect information. But I think this
will be fixed.
AmigaOS4 Petunia benchmarks : Comment 68 of 80ANN.lu
Posted by tarbos on 20-Aug-2003 22:55 GMT
In reply to Comment 61 (BrianK):
>The G4 series is a 7450 or 7455. There is a number of different variants out there such as 7400, 7410 (originalG4 core, quite comparable to G3) and 7450, 7440, 7451, 7441, 7455, 7445, 7457and 7447. >I can track down the number if you like. This would be nice since I only found SPEC2000 results for the more recent G4.
AmigaOS4 Petunia benchmarks : Comment 69 of 80ANN.lu
Posted by Anonymous on 20-Aug-2003 22:59 GMT
In reply to Comment 62 (Anonymous):
>They use the G4 or what Motorola calls the 7450. They use the 7450 _family_ - sometimes called G4+ or G4e. >Thus when you were comparing 750GX and FX you were comparing 2>different versions of what Apple calls the G3. Yes, it's all confusing. My point is that G3 improved processing power per MHz in its differentincarnations and therefore the comparison between G3 and G4 should be madewith more recent CPUs than the old 500MHz ones.
AmigaOS4 Petunia benchmarks : Comment 70 of 80ANN.lu
Posted by tarbos on 20-Aug-2003 23:02 GMT
In reply to Comment 63 (BrianK):
>Apple's Dual 1Ghz 7455 has SpecINT:307 SPECmark only tests one CPU - latest numbers from Motorola for a 1.25GHz 7455is (est.) 500 SPECint2000.
AmigaOS4 Petunia benchmarks : Comment 71 of 80ANN.lu
Posted by Anonymous on 20-Aug-2003 23:05 GMT
In reply to Comment 63 (BrianK):
>Thus the 7455 appears to be a minor improvement over the 7450 in FPU processing Again I must ask - are the L3 caches the same size and type (SDR/DDR) and the FSB the same?Else you cannot make this statement about the actual CPU itself.
AmigaOS4 Petunia benchmarks : Comment 72 of 80ANN.lu
Posted by tarbos on 20-Aug-2003 23:10 GMT
In reply to Comment 67 (David Scheibler):
>But it's true that some websites have incorrect information. Just look at the master website http://www.bplan-gmbh.de/news/pegasos_e.html >But I think this will be fixed. I don't think so as I informed them a long time ago about it...
AmigaOS4 Petunia benchmarks : Comment 73 of 80ANN.lu
Posted by BrianK on 21-Aug-2003 09:40 GMT
In reply to Comment 67 (David Scheibler):
Definitely see the posting you responded to where 7447 was shown on the PegasOS site.
AmigaOS4 Petunia benchmarks : Comment 74 of 80ANN.lu
Posted by BrianK on 21-Aug-2003 09:44 GMT
In reply to Comment 69 (Anonymous):
Actually ideally the comparison would be made with the actual model, speed, motherboard, etc. used for the AmigaOne and PegasOS machine. This helps to eliminate confusion and make the tests more realistic to what we're trying to accomplish.

The original question was about the performance of these machines. The 800Mhz G3 AmigaOne is going to perform the same as the 800Mhz G4 AmigaOne. However, the G4 should be noticably quicker on FPU related tasks.

If Amiga and Pegas want they can send me boards I'll gladly review, gladly run some benchmarks on Linux on both boards, post the results to the Web and send the boards back.
AmigaOS4 Petunia benchmarks : Comment 75 of 80ANN.lu
Posted by BrianK on 21-Aug-2003 09:54 GMT
In reply to Comment 71 (Anonymous):
Who cares SpecINT and SpecFPU aren't neccessarily the best tests to convey what exact performance is. They're easy measure to relate the performance of the chips to one another. Unfortunately no one's done a PegasOS to AmigaOne performance evaluation yet. I'd say the same tests within Linux and perhaps within MoL should cover it. Thus, using Apple's machines and their posting, which are oh so more prevalent on the web, was to get a relative expectation of the performance for the 2 new machines and their 2 processors. Please don't try to make this exact.

Simply put what's your butt reaction? How are these machines going to feel when you're using them. Here's your performance the G3 and G4 will kick ass all over the 604e and the 060 machines, at least 3x the speed, and even more so against the 040/030 and 020 set. For most applications G3 and G4 are both at 800Mhz and will perform about the same, seat of pants remember. For those users that are graphically intesive users using programs such as ImageFX for image processing the performance relies on FPU calls more frequently. As such the G4's higher FPU rates should process your images faster and you'll get done with your project quicker.
AmigaOS4 Petunia benchmarks : Comment 76 of 80ANN.lu
Posted by tarbos on 21-Aug-2003 11:33 GMT
In reply to Comment 74 (BrianK):
>However, the G4 should be noticably quicker on FPU related tasks. With certain double-precision floating point operations, yes.But can you explain why Cinebench runs faster on G3?
AmigaOS4 Petunia benchmarks : Comment 77 of 80ANN.lu
Posted by BrianK on 21-Aug-2003 12:45 GMT
In reply to Comment 76 (tarbos):
I think your best course of action is work with Cinebench and MorphOS. Perhaps they are utilizing less efficent FPU calls or the JIT Emulation isn't as strong in that area or there aren't as many FPU calls as you think.

Alternatively, I'll be glad to provide a shipping address where you can ship me a PegasOS G3 and G4 system along with a copy of Cinebench and I'll try to track it down for you.
AmigaOS4 Petunia benchmarks : Comment 78 of 80ANN.lu
Posted by Anonymous on 21-Aug-2003 19:48 GMT
In reply to Comment 77 (BrianK):
Sorry, it's a native Mac application - see benchmark results athttp://www.macspeedzone.com/html/hardware/machine/comparison/all/master_list_9.html
AmigaOS4 Petunia benchmarks : Comment 79 of 80ANN.lu
Posted by BrianK on 21-Aug-2003 20:42 GMT
In reply to Comment 78 (Anonymous):
I have a Mac G4 Tower and G3 iMac if you'd like me to test the application for you.

If you want to see how/what it does on the MorphOS I'll get Mac on Linux running and test for you. I have a purchased version of OS9 and OS8 sitting here. (I of course have OSX but those licenses are in use so I can't use them)

But, really it's probably more to do with the applications programmers not being properly optimizied or having some other problem then the G3/G4 speed. Bad apps can make good processors look bad.
AmigaOS4 Petunia benchmarks : Comment 80 of 80ANN.lu
Posted by Anonymous on 21-Aug-2003 20:45 GMT
In reply to Comment 79 (BrianK):
Cinebench -- I've seen various postings with people complaining of the poor quality of this benchmarking tool, unrealistic numbers, and it not relating to any other benchmarking tool. Maybe it's fixed but there's more then a few complainers out there about it's quality.
Anonymous, there are 80 items in your selection (but only 30 shown due to limitation) [1 - 50] [51 - 80]
Back to Top