@wawa

Dual core PA6T is best on laptops.

_________________
Amiga Developer
Amiga 500, Efika, Mac Mini and PowerBook

@G4 vs G5


Barefeats G4(dual)vsG5MP.
Similar differences than what is seen under MOS.

G4vsG5, both 2Ghz


Old PPC vs x86 bench.
See the UT2004 test with different builds, optimization can help G5 a lot.
PCIe tests...

PowerMacG5 info.
Power Mac G5 FAQ...what models have PCIe...

Last edited by KimmoK on 04-Jan-2013 at 08:24 PM.
Last edited by KimmoK on 04-Jan-2013 at 08:22 PM.
Last edited by KimmoK on 04-Jan-2013 at 08:21 PM.
Last edited by KimmoK on 04-Jan-2013 at 08:15 PM.
Last edited by KimmoK on 04-Jan-2013 at 08:10 PM.
Last edited by KimmoK on 04-Jan-2013 at 08:07 PM.

_________________
- KimmoK
// For freedom, for honor, for AMIGA
//
// Thing that I should find more time for: CC64 - 64bit Community Computer?

68k emulation PowerMac G5 vs Pegasos II G4 vs PowerBook G4 (source: Amigamccc.org)
PowerMac G5 2.3 GHz seems to be as fast as 1.75-2.0 GHz G4 (Pegasos II G4 baseline) in these tests. Not bad, but certainly not impressive.


3D benchmarks (Quake 3; source: Morphzone.org; post 24, Jambalah)

PowerMac G4 1.8 GHz ATI Radeon 9250 128M:
1260 frames 18.7 sec. 67.3 fps

PowerMAc G5 2.7 GHz ATI Radeon 9650SE 256M:
1260 frames 30.9 sec. 40.7 fps

PowerMac G5 2.7 GHz ATI Radeon 9800XT 256M:
1260 frames 50.5 sec. 24.9 fps

Well, R300 drivers still need some speed up.

Last edited by pavlor on 05-Oct-2013 at 02:26 PM.

@pavlor

Quote:

68k emulation PowerMac G5 vs Pegasos II G4 vs PowerBook G4 (source: Amigamccc.org) PowerMac G5 2.3 GHz seems to be as fast as 1.75-2.0 GHz G4 (Pegasos II G4 baseline) in these tests. Not bad, but certainly not impressive.

I am not Mac expert but this appears to be in line with benchmarks from OSX world. In some cases G5 specific optimizations can give significant performance boost but G4 systems are the best buy. IMO of course :)

_________________
Amiga Developer
Amiga 500, Efika, Mac Mini and PowerBook

@itix

Quote:

In some cases G5 specific optimizations can give significant performance boost but G4 systems are the best buy. IMO of course :)

Yes, some benchmarks look really nice (lame, MPlayer, Blender), some not so (68k emulation, OGR, DosBox).

It would be interesting to know, if newer designs from Freescale (eg. e5500 core) are faster than good old G4/G5. Coremark score of e6500 does indeed look impressive, but thas means nothing in terms of real life performance. Once e5500 based "X2000" is benchmarked, we will know more.

@pavlor

G5 has two FPU units per core vs. one FPU unit on G4 per core. On emulation (68k, x86) it is not so much useful but could explain why e.g. Blender is so much faster on G5. G5 has better memory throuhput and larger L1 cache which helps on some applications but its longer L1 cache lines render some G4 memory optimizations ineffective.

_________________
Amiga Developer
Amiga 500, Efika, Mac Mini and PowerBook

@pavlor

Quote:

Yes, some benchmarks look really nice (lame, MPlayer, Blender), some not so (68k emulation, OGR, DosBox).

Well in some emulators like MAME or UAE, a G5 completely crushes a G4, so be careful with your comparisons, really.

As for the tests involving 3d cards, it's of course not related to the G5 but to the 3d driver maturity.

@Fab

Quote:

Well in some emulators like MAME or UAE, a G5 completely crushes a G4

So G5 aware UAE was released? Nice!

Quote:

so be careful with your comparisons, really.

That was exactly my point.

Quote:

As for the tests involving 3d cards, it's of course not related to the G5 but to the 3d driver maturity.

Support for more modern GFX cards is nice step forward, but these drivers certainly need some work - eg. Radeon 9800XT much slower than Radeon 9600?

@pavlor

Yes, it was available before MorphOS for G5 was released, actually. :)