A lot of the benchmarking here places my PearPC performance in the realm of a 233 to 300MHz G3 as well.
However, there seems to be certain operations (mostly rendering) where the performance drops to the level of an imaginary 67MHz G3. I don't know if that is coincidental with PearPC's reported bus speed, or not.
How fast is PearPC: Benchmarks
by Ribtorus » Wed Oct 20, 2004 8:57 pm
[quote="User":b84528d2da]
To Rib torus
Can you post the launch time of the applications such how :System Preferences,Quicktime,Safari and other
[/quote:b84528d2da]
Sure.
Launch times are as follows:
Safari: [b:b84528d2da]7[/b:b84528d2da] seconds with another [b:b84528d2da]3[/b:b84528d2da] seconds to render yahoo.com (homepage)
iTunes: [b:b84528d2da]7[/b:b84528d2da] seconds
ical: [b:b84528d2da]9[/b:b84528d2da] seconds
iMovie: [b:b84528d2da]9[/b:b84528d2da] seconds
iPhoto: [b:b84528d2da]15[/b:b84528d2da] seconds
iChat: [b:b84528d2da]6[/b:b84528d2da] seconds, including time to connect
System Preferences: [b:b84528d2da]6[/b:b84528d2da] seconds
Quicktime: [b:b84528d2da]3[/b:b84528d2da] seconds
Appleworks 6: [b:b84528d2da]5[/b:b84528d2da] seconds
DVD Studio Pro 3: [b:b84528d2da]13[/b:b84528d2da] seconds
Grab: [b:b84528d2da]2[/b:b84528d2da] seconds
Terminal: [b:b84528d2da]13[/b:b84528d2da] seconds to a fully enabled cursor
Photoshop CS: [b:b84528d2da]60[/b:b84528d2da] seconds
Boot time from PearPC boot select screen: [b:b84528d2da]1[/b:b84528d2da] min [b:b84528d2da]30[/b:b84528d2da] seconds. Sometimes boot time can be just over two minutes with the same build . I don't know why.
To Rib torus
Can you post the launch time of the applications such how :System Preferences,Quicktime,Safari and other
[/quote:b84528d2da]
Sure.
Launch times are as follows:
Safari: [b:b84528d2da]7[/b:b84528d2da] seconds with another [b:b84528d2da]3[/b:b84528d2da] seconds to render yahoo.com (homepage)
iTunes: [b:b84528d2da]7[/b:b84528d2da] seconds
ical: [b:b84528d2da]9[/b:b84528d2da] seconds
iMovie: [b:b84528d2da]9[/b:b84528d2da] seconds
iPhoto: [b:b84528d2da]15[/b:b84528d2da] seconds
iChat: [b:b84528d2da]6[/b:b84528d2da] seconds, including time to connect
System Preferences: [b:b84528d2da]6[/b:b84528d2da] seconds
Quicktime: [b:b84528d2da]3[/b:b84528d2da] seconds
Appleworks 6: [b:b84528d2da]5[/b:b84528d2da] seconds
DVD Studio Pro 3: [b:b84528d2da]13[/b:b84528d2da] seconds
Grab: [b:b84528d2da]2[/b:b84528d2da] seconds
Terminal: [b:b84528d2da]13[/b:b84528d2da] seconds to a fully enabled cursor
Photoshop CS: [b:b84528d2da]60[/b:b84528d2da] seconds
Boot time from PearPC boot select screen: [b:b84528d2da]1[/b:b84528d2da] min [b:b84528d2da]30[/b:b84528d2da] seconds. Sometimes boot time can be just over two minutes with the same build . I don't know why.
- Ribtorus
- Advanced Member
- Posts: 208
- Joined: Sat Aug 28, 2004 12:24 am
by Ribtorus » Thu Oct 21, 2004 3:42 pm
My next series of benchmarking is more challenging; multi-tasking.
I am so impressed with PearPC that I have decided to let it run the new eDonkey GUI mac OSX P2P app. It works fine. SourceForge has download links for it.
So I let it run, and while it's downloading stuff I ran XBench, SpeedRun, Pi calculator and SpeedX.
With eDonkey running VS without eDonkey running
Pi calculator: [b:12e288a12c]2 min 55[/b:12e288a12c] sec VS [b:12e288a12c]76[/b:12e288a12c] seconds
Speedrun-
Graphics: [b:12e288a12c]15[/b:12e288a12c] VS [b:12e288a12c]19[/b:12e288a12c]
Hard Drive: [b:12e288a12c]58[/b:12e288a12c] VS [b:12e288a12c]118[/b:12e288a12c]
Processor: [b:12e288a12c]32[/b:12e288a12c] VS [b:12e288a12c]56[/b:12e288a12c]
Ram: [b:12e288a12c]87 [/b:12e288a12c]VS [b:12e288a12c]114[/b:12e288a12c]
Overall: [b:12e288a12c]48[/b:12e288a12c] VS [b:12e288a12c]85[/b:12e288a12c]
XBench-
CPU: [b:12e288a12c].32[/b:12e288a12c] VS [b:12e288a12c].57[/b:12e288a12c]
Thread: [b:12e288a12c]18.27[/b:12e288a12c] VS [b:12e288a12c]27.51[/b:12e288a12c]
Memory: [b:12e288a12c]33.81[/b:12e288a12c] VS [b:12e288a12c]48.29[/b:12e288a12c]
Disk: [b:12e288a12c]22.09[/b:12e288a12c] VS [b:12e288a12c]47.82[/b:12e288a12c]
SpeedX, 10 billion instructions, "typical" setting-
[b:12e288a12c]55.2[/b:12e288a12c] seconds, [b:12e288a12c]181.2[/b:12e288a12c] MIPS VS [b:12e288a12c]33.2[/b:12e288a12c] seconds [b:12e288a12c]300.8[/b:12e288a12c] MIPS
Clearly asking PearPC/OSX to share eDonkey with other apps is asking rather too much, as the slowdown is considerable.
I tried to run the PhotoshopCS benchmark, but after about 35 minutes it appeard to stall out. Without eDonkey running, the benchmark takes between 14 and 15 minutes.
I'll try other background apps as well. Perhaps the eDonkey GUI version for OSX is not particularly well optimised yet.
I am so impressed with PearPC that I have decided to let it run the new eDonkey GUI mac OSX P2P app. It works fine. SourceForge has download links for it.
So I let it run, and while it's downloading stuff I ran XBench, SpeedRun, Pi calculator and SpeedX.
With eDonkey running VS without eDonkey running
Pi calculator: [b:12e288a12c]2 min 55[/b:12e288a12c] sec VS [b:12e288a12c]76[/b:12e288a12c] seconds
Speedrun-
Graphics: [b:12e288a12c]15[/b:12e288a12c] VS [b:12e288a12c]19[/b:12e288a12c]
Hard Drive: [b:12e288a12c]58[/b:12e288a12c] VS [b:12e288a12c]118[/b:12e288a12c]
Processor: [b:12e288a12c]32[/b:12e288a12c] VS [b:12e288a12c]56[/b:12e288a12c]
Ram: [b:12e288a12c]87 [/b:12e288a12c]VS [b:12e288a12c]114[/b:12e288a12c]
Overall: [b:12e288a12c]48[/b:12e288a12c] VS [b:12e288a12c]85[/b:12e288a12c]
XBench-
CPU: [b:12e288a12c].32[/b:12e288a12c] VS [b:12e288a12c].57[/b:12e288a12c]
Thread: [b:12e288a12c]18.27[/b:12e288a12c] VS [b:12e288a12c]27.51[/b:12e288a12c]
Memory: [b:12e288a12c]33.81[/b:12e288a12c] VS [b:12e288a12c]48.29[/b:12e288a12c]
Disk: [b:12e288a12c]22.09[/b:12e288a12c] VS [b:12e288a12c]47.82[/b:12e288a12c]
SpeedX, 10 billion instructions, "typical" setting-
[b:12e288a12c]55.2[/b:12e288a12c] seconds, [b:12e288a12c]181.2[/b:12e288a12c] MIPS VS [b:12e288a12c]33.2[/b:12e288a12c] seconds [b:12e288a12c]300.8[/b:12e288a12c] MIPS
Clearly asking PearPC/OSX to share eDonkey with other apps is asking rather too much, as the slowdown is considerable.
I tried to run the PhotoshopCS benchmark, but after about 35 minutes it appeard to stall out. Without eDonkey running, the benchmark takes between 14 and 15 minutes.
I'll try other background apps as well. Perhaps the eDonkey GUI version for OSX is not particularly well optimised yet.
- Ribtorus
- Advanced Member
- Posts: 208
- Joined: Sat Aug 28, 2004 12:24 am
by Ribtorus » Thu Oct 21, 2004 6:02 pm
Another multi-tasking test. This time I ran iTunes and connected to a radio station, opened up iChat, then ran safari. I then ran SpeedX, SpeedRun, Pi Calculator and XBench.
SpeedX- 35.4 seconds. 282.5 MIPS Not much of a slowdown at all
SpeedRun-
Graphics:17
Hard Drive: 112
CPU: 55
RAM: 137
Overall: 81
Again, not much of a slowdown
Pi Calculator- 1 minute 41 seconds. That's a considerable slowdown.
XBench-
CPU: .54
Thread: 24.99
Memory: 47.71
Disk: 47.33
Not much of a slowdown there either.
Then I kept iTunes, iChat and safari running, and fired up eDonkey.
SpeedX- 94.23 seconds, 106.1 MIPS, obviously a big hit on performance.
Speedrun-
Graphics: 12
hard Drive: 39
CPU: 22
RAM: 54
Overall: 32
That's another big hit.
Pi calculator- 5 minutes 12 seconds. A big hit.
Xbench-
CPU: .28
Thread: 19.34
Memory: 25.80
Disk: 35.95
And that's another big hit.
Clearly there is something about eDonkey which makes it quite innefficient as far as multi-tasking goes.
But the good news is, Apps like Safari, iChat, and iTunes, which are typically likely to be left running in the background, don't place too much of a burden on PearPC/OSX.
SpeedX- 35.4 seconds. 282.5 MIPS Not much of a slowdown at all
SpeedRun-
Graphics:17
Hard Drive: 112
CPU: 55
RAM: 137
Overall: 81
Again, not much of a slowdown
Pi Calculator- 1 minute 41 seconds. That's a considerable slowdown.
XBench-
CPU: .54
Thread: 24.99
Memory: 47.71
Disk: 47.33
Not much of a slowdown there either.
Then I kept iTunes, iChat and safari running, and fired up eDonkey.
SpeedX- 94.23 seconds, 106.1 MIPS, obviously a big hit on performance.
Speedrun-
Graphics: 12
hard Drive: 39
CPU: 22
RAM: 54
Overall: 32
That's another big hit.
Pi calculator- 5 minutes 12 seconds. A big hit.
Xbench-
CPU: .28
Thread: 19.34
Memory: 25.80
Disk: 35.95
And that's another big hit.
Clearly there is something about eDonkey which makes it quite innefficient as far as multi-tasking goes.
But the good news is, Apps like Safari, iChat, and iTunes, which are typically likely to be left running in the background, don't place too much of a burden on PearPC/OSX.
- Ribtorus
- Advanced Member
- Posts: 208
- Joined: Sat Aug 28, 2004 12:24 am
by HFU » Thu Oct 28, 2004 2:31 am
The latest Altivec 0.4Pre branch from RichardGoodwin.com compare with [url=http://forums.pearpc.net/viewtopic.php?p=6141&highlight=hfu#6141:9264a61368]previous[/url:9264a61368] non-Altivec 0.4Pre result. It actually says G4 emulation now. Here is the result for XBench:
[b:9264a61368]Panther 10.3.5:[/b:9264a61368] 1.36GHz/G4 (Windows) | 1.36GHz/G3 (Windows)
[color=green:9264a61368][b:9264a61368]XBench:[/b:9264a61368][/color:9264a61368]
[color=blue:9264a61368][b:9264a61368]CPU Test:[/b:9264a61368][/color:9264a61368] [b:9264a61368]1.84[/b:9264a61368] | 0.56
- GCD Loop: 40.3 | [b:9264a61368]43.98[/b:9264a61368]
- Floating Point Basic: [b:9264a61368]2.79[/b:9264a61368] | 2.74
- AltiVec Basic: [b:9264a61368]0.85[/b:9264a61368] | N/A
- vecLib FFT: [b:9264a61368]0.92[/b:9264a61368] | 0.15
- Floating Point Library: [b:9264a61368]16.80[/b:9264a61368] | 14.60
[color=blue:9264a61368][b:9264a61368]Thread Test:[/b:9264a61368][/color:9264a61368] 28.77 | [b:9264a61368]29.21[/b:9264a61368]
- Computation: [b:9264a61368]59.72[/b:9264a61368] | 59.62
- Lock Contention: 18.95 | [b:9264a61368]19.34[/b:9264a61368]
[color=blue:9264a61368][b:9264a61368]Memory Test:[/b:9264a61368][/color:9264a61368] [b:9264a61368]66.61[/b:9264a61368] | 44.85
- System: [b:9264a61368]63.46[/b:9264a61368] | 55.85
-- Allocate: [b:9264a61368]65.06[/b:9264a61368] | 57.82
-- Fill: 53.75 | [b:9264a61368]55.95[/b:9264a61368]
-- Copy: [b:9264a61368]75.13[/b:9264a61368] | 53.91
- Stream: [b:9264a61368]70.09[/b:9264a61368] | 37.47
-- Copy: [b:9264a61368]76.61[/b:9264a61368] | 40.43 [altivec]
-- Scale: [b:9264a61368]60.57[/b:9264a61368] | 28.62 [altivec]
-- Add: [b:9264a61368]72.37[/b:9264a61368] | 42.11 [altivec]
-- Triad: [b:9264a61368]73.08[/b:9264a61368] | 42.83 [altivec]
[color=blue:9264a61368][b:9264a61368]Disk Test:[/b:9264a61368][/color:9264a61368] 70.33 | [b:9264a61368]81.70[/b:9264a61368]
- Sequential: 41.56 | [b:9264a61368]49.82[/b:9264a61368]
-- Uncached Write: 20.48 | [b:9264a61368]31.06[/b:9264a61368]
-- Uncached Write: 128.22 | [b:9264a61368]129.05[/b:9264a61368]
-- Uncached Read: [b:9264a61368]29.67[/b:9264a61368] | 28.93
-- Uncached Read: 169.01 | [b:9264a61368]173.0[/b:9264a61368]
- Random: [b:9264a61368]228.50[/b:9264a61368] | 227.0
-- Uncached Write: 199.86 | [b:9264a61368]202.24[/b:9264a61368]
-- Uncached Write: 125.13 | [b:9264a61368]125.96[/b:9264a61368]
-- Uncached Read: [b:9264a61368]590.44[/b:9264a61368] | 545.85
-- Uncached Read: [b:9264a61368]355.02[/b:9264a61368] | 344.13
[color=blue:9264a61368][b:9264a61368]Quartz Graphic Test:[/b:9264a61368][/color:9264a61368] [b:9264a61368]22.49[/b:9264a61368] | N/A
[color=blue:9264a61368][b:9264a61368]OpenGL Test:[/b:9264a61368][/color:9264a61368] [b:9264a61368]2.29[/b:9264a61368] | N/A
[color=blue:9264a61368][b:9264a61368]User Interface Test:[/b:9264a61368][/color:9264a61368] [b:9264a61368]22.15[/b:9264a61368] | N/A
[color=green:9264a61368][b:9264a61368]Cinebench:[/b:9264a61368][/color:9264a61368] [b:9264a61368]7CB-CPU[/b:9264a61368] | 5CB-CPU
[color=green:9264a61368][b:9264a61368]SpeedX:[/b:9264a61368][/color:9264a61368] [b:9264a61368]39.98[/b:9264a61368] | 41.45 seconds, [b:9264a61368]250.1[/b:9264a61368] | 243.2 MIPS @ 10 billion
Ribtorus, I'm amazed you actually seat through Cinebench.
With this new Altivec build, PearPC feels a bit more snappier for overall use.
HFU
[b:9264a61368]Panther 10.3.5:[/b:9264a61368] 1.36GHz/G4 (Windows) | 1.36GHz/G3 (Windows)
[color=green:9264a61368][b:9264a61368]XBench:[/b:9264a61368][/color:9264a61368]
[color=blue:9264a61368][b:9264a61368]CPU Test:[/b:9264a61368][/color:9264a61368] [b:9264a61368]1.84[/b:9264a61368] | 0.56
- GCD Loop: 40.3 | [b:9264a61368]43.98[/b:9264a61368]
- Floating Point Basic: [b:9264a61368]2.79[/b:9264a61368] | 2.74
- AltiVec Basic: [b:9264a61368]0.85[/b:9264a61368] | N/A
- vecLib FFT: [b:9264a61368]0.92[/b:9264a61368] | 0.15
- Floating Point Library: [b:9264a61368]16.80[/b:9264a61368] | 14.60
[color=blue:9264a61368][b:9264a61368]Thread Test:[/b:9264a61368][/color:9264a61368] 28.77 | [b:9264a61368]29.21[/b:9264a61368]
- Computation: [b:9264a61368]59.72[/b:9264a61368] | 59.62
- Lock Contention: 18.95 | [b:9264a61368]19.34[/b:9264a61368]
[color=blue:9264a61368][b:9264a61368]Memory Test:[/b:9264a61368][/color:9264a61368] [b:9264a61368]66.61[/b:9264a61368] | 44.85
- System: [b:9264a61368]63.46[/b:9264a61368] | 55.85
-- Allocate: [b:9264a61368]65.06[/b:9264a61368] | 57.82
-- Fill: 53.75 | [b:9264a61368]55.95[/b:9264a61368]
-- Copy: [b:9264a61368]75.13[/b:9264a61368] | 53.91
- Stream: [b:9264a61368]70.09[/b:9264a61368] | 37.47
-- Copy: [b:9264a61368]76.61[/b:9264a61368] | 40.43 [altivec]
-- Scale: [b:9264a61368]60.57[/b:9264a61368] | 28.62 [altivec]
-- Add: [b:9264a61368]72.37[/b:9264a61368] | 42.11 [altivec]
-- Triad: [b:9264a61368]73.08[/b:9264a61368] | 42.83 [altivec]
[color=blue:9264a61368][b:9264a61368]Disk Test:[/b:9264a61368][/color:9264a61368] 70.33 | [b:9264a61368]81.70[/b:9264a61368]
- Sequential: 41.56 | [b:9264a61368]49.82[/b:9264a61368]
-- Uncached Write: 20.48 | [b:9264a61368]31.06[/b:9264a61368]
-- Uncached Write: 128.22 | [b:9264a61368]129.05[/b:9264a61368]
-- Uncached Read: [b:9264a61368]29.67[/b:9264a61368] | 28.93
-- Uncached Read: 169.01 | [b:9264a61368]173.0[/b:9264a61368]
- Random: [b:9264a61368]228.50[/b:9264a61368] | 227.0
-- Uncached Write: 199.86 | [b:9264a61368]202.24[/b:9264a61368]
-- Uncached Write: 125.13 | [b:9264a61368]125.96[/b:9264a61368]
-- Uncached Read: [b:9264a61368]590.44[/b:9264a61368] | 545.85
-- Uncached Read: [b:9264a61368]355.02[/b:9264a61368] | 344.13
[color=blue:9264a61368][b:9264a61368]Quartz Graphic Test:[/b:9264a61368][/color:9264a61368] [b:9264a61368]22.49[/b:9264a61368] | N/A
[color=blue:9264a61368][b:9264a61368]OpenGL Test:[/b:9264a61368][/color:9264a61368] [b:9264a61368]2.29[/b:9264a61368] | N/A
[color=blue:9264a61368][b:9264a61368]User Interface Test:[/b:9264a61368][/color:9264a61368] [b:9264a61368]22.15[/b:9264a61368] | N/A
[color=green:9264a61368][b:9264a61368]Cinebench:[/b:9264a61368][/color:9264a61368] [b:9264a61368]7CB-CPU[/b:9264a61368] | 5CB-CPU
[color=green:9264a61368][b:9264a61368]SpeedX:[/b:9264a61368][/color:9264a61368] [b:9264a61368]39.98[/b:9264a61368] | 41.45 seconds, [b:9264a61368]250.1[/b:9264a61368] | 243.2 MIPS @ 10 billion
Ribtorus, I'm amazed you actually seat through Cinebench.
With this new Altivec build, PearPC feels a bit more snappier for overall use.HFU
- HFU
- New Member
- Posts: 3
- Joined: Sat Sep 18, 2004 7:11 pm
by Ribtorus » Thu Oct 28, 2004 1:40 pm
Cool.
Thanks for posting more results.
How about more Cinebench and Strata3D CX? You can run them at night and see the results in the morning. But I'm very curious to see if others run up against that 50X slowdown barrier for rendering.
Thanks for posting more results.
How about more Cinebench and Strata3D CX? You can run them at night and see the results in the morning. But I'm very curious to see if others run up against that 50X slowdown barrier for rendering.
- Ribtorus
- Advanced Member
- Posts: 208
- Joined: Sat Aug 28, 2004 12:24 am
by Ribtorus » Thu Oct 28, 2004 4:06 pm
[quote="HFU":435843f176]
The latest Altivec 0.4Pre branch from RichardGoodwin.com compare with [url=http://forums.pearpc.net/viewtopic.php?p=6141&highlight=hfu#6141:435843f176]previous[/url:435843f176] non-Altivec 0.4Pre result. It actually says G4 emulation now. Here is the result for XBench:
[b:435843f176]Panther 10.3.5:[/b:435843f176] 1.36GHz/G4 (Windows) | 1.36GHz/G3 (Windows)
[color=green:435843f176][b:435843f176]XBench:[/b:435843f176][/color:435843f176]
[color=blue:435843f176][b:435843f176]CPU Test:[/b:435843f176][/color:435843f176] [b:435843f176]1.84[/b:435843f176] | 0.56
- GCD Loop: 40.3 | [b:435843f176]43.98[/b:435843f176]
- Floating Point Basic: [b:435843f176]2.79[/b:435843f176] | 2.74
- AltiVec Basic: [b:435843f176]0.85[/b:435843f176] | N/A
- vecLib FFT: [b:435843f176]0.92[/b:435843f176] | 0.15
- Floating Point Library: [b:435843f176]16.80[/b:435843f176] | 14.60
[color=blue:435843f176][b:435843f176]Thread Test:[/b:435843f176][/color:435843f176] 28.77 | [b:435843f176]29.21[/b:435843f176]
- Computation: [b:435843f176]59.72[/b:435843f176] | 59.62
- Lock Contention: 18.95 | [b:435843f176]19.34[/b:435843f176]
[color=blue:435843f176][b:435843f176]Memory Test:[/b:435843f176][/color:435843f176] [b:435843f176]66.61[/b:435843f176] | 44.85
- System: [b:435843f176]63.46[/b:435843f176] | 55.85
-- Allocate: [b:435843f176]65.06[/b:435843f176] | 57.82
-- Fill: 53.75 | [b:435843f176]55.95[/b:435843f176]
-- Copy: [b:435843f176]75.13[/b:435843f176] | 53.91
- Stream: [b:435843f176]70.09[/b:435843f176] | 37.47
-- Copy: [b:435843f176]76.61[/b:435843f176] | 40.43 [altivec]
-- Scale: [b:435843f176]60.57[/b:435843f176] | 28.62 [altivec]
-- Add: [b:435843f176]72.37[/b:435843f176] | 42.11 [altivec]
-- Triad: [b:435843f176]73.08[/b:435843f176] | 42.83 [altivec]
[color=blue:435843f176][b:435843f176]Disk Test:[/b:435843f176][/color:435843f176] 70.33 | [b:435843f176]81.70[/b:435843f176]
- Sequential: 41.56 | [b:435843f176]49.82[/b:435843f176]
-- Uncached Write: 20.48 | [b:435843f176]31.06[/b:435843f176]
-- Uncached Write: 128.22 | [b:435843f176]129.05[/b:435843f176]
-- Uncached Read: [b:435843f176]29.67[/b:435843f176] | 28.93
-- Uncached Read: 169.01 | [b:435843f176]173.0[/b:435843f176]
- Random: [b:435843f176]228.50[/b:435843f176] | 227.0
-- Uncached Write: 199.86 | [b:435843f176]202.24[/b:435843f176]
-- Uncached Write: 125.13 | [b:435843f176]125.96[/b:435843f176]
-- Uncached Read: [b:435843f176]590.44[/b:435843f176] | 545.85
-- Uncached Read: [b:435843f176]355.02[/b:435843f176] | 344.13
[color=blue:435843f176][b:435843f176]Quartz Graphic Test:[/b:435843f176][/color:435843f176] [b:435843f176]22.49[/b:435843f176] | N/A
[color=blue:435843f176][b:435843f176]OpenGL Test:[/b:435843f176][/color:435843f176] [b:435843f176]2.29[/b:435843f176] | N/A
[color=blue:435843f176][b:435843f176]User Interface Test:[/b:435843f176][/color:435843f176] [b:435843f176]22.15[/b:435843f176] | N/A
[color=green:435843f176][b:435843f176]Cinebench:[/b:435843f176][/color:435843f176] [b:435843f176]7CB-CPU[/b:435843f176] | 5CB-CPU
[color=green:435843f176][b:435843f176]SpeedX:[/b:435843f176][/color:435843f176] [b:435843f176]39.98[/b:435843f176] | 41.45 seconds, [b:435843f176]250.1[/b:435843f176] | 243.2 MIPS @ 10 billion
Ribtorus, I'm amazed you actually seat through Cinebench. With this new Altivec build, PearPC feels a bit more snappier for overall use.
HFU
[/quote:435843f176]
I still find no benefit (apart from XBench scores) to running the Altivec builds. I tried the latest from RichardGoodwin with no significant change to XBench or SpeedX. I also get screen degradation and tiny black dots showing up all over the place with Altivec builds.
The latest Altivec 0.4Pre branch from RichardGoodwin.com compare with [url=http://forums.pearpc.net/viewtopic.php?p=6141&highlight=hfu#6141:435843f176]previous[/url:435843f176] non-Altivec 0.4Pre result. It actually says G4 emulation now. Here is the result for XBench:
[b:435843f176]Panther 10.3.5:[/b:435843f176] 1.36GHz/G4 (Windows) | 1.36GHz/G3 (Windows)
[color=green:435843f176][b:435843f176]XBench:[/b:435843f176][/color:435843f176]
[color=blue:435843f176][b:435843f176]CPU Test:[/b:435843f176][/color:435843f176] [b:435843f176]1.84[/b:435843f176] | 0.56
- GCD Loop: 40.3 | [b:435843f176]43.98[/b:435843f176]
- Floating Point Basic: [b:435843f176]2.79[/b:435843f176] | 2.74
- AltiVec Basic: [b:435843f176]0.85[/b:435843f176] | N/A
- vecLib FFT: [b:435843f176]0.92[/b:435843f176] | 0.15
- Floating Point Library: [b:435843f176]16.80[/b:435843f176] | 14.60
[color=blue:435843f176][b:435843f176]Thread Test:[/b:435843f176][/color:435843f176] 28.77 | [b:435843f176]29.21[/b:435843f176]
- Computation: [b:435843f176]59.72[/b:435843f176] | 59.62
- Lock Contention: 18.95 | [b:435843f176]19.34[/b:435843f176]
[color=blue:435843f176][b:435843f176]Memory Test:[/b:435843f176][/color:435843f176] [b:435843f176]66.61[/b:435843f176] | 44.85
- System: [b:435843f176]63.46[/b:435843f176] | 55.85
-- Allocate: [b:435843f176]65.06[/b:435843f176] | 57.82
-- Fill: 53.75 | [b:435843f176]55.95[/b:435843f176]
-- Copy: [b:435843f176]75.13[/b:435843f176] | 53.91
- Stream: [b:435843f176]70.09[/b:435843f176] | 37.47
-- Copy: [b:435843f176]76.61[/b:435843f176] | 40.43 [altivec]
-- Scale: [b:435843f176]60.57[/b:435843f176] | 28.62 [altivec]
-- Add: [b:435843f176]72.37[/b:435843f176] | 42.11 [altivec]
-- Triad: [b:435843f176]73.08[/b:435843f176] | 42.83 [altivec]
[color=blue:435843f176][b:435843f176]Disk Test:[/b:435843f176][/color:435843f176] 70.33 | [b:435843f176]81.70[/b:435843f176]
- Sequential: 41.56 | [b:435843f176]49.82[/b:435843f176]
-- Uncached Write: 20.48 | [b:435843f176]31.06[/b:435843f176]
-- Uncached Write: 128.22 | [b:435843f176]129.05[/b:435843f176]
-- Uncached Read: [b:435843f176]29.67[/b:435843f176] | 28.93
-- Uncached Read: 169.01 | [b:435843f176]173.0[/b:435843f176]
- Random: [b:435843f176]228.50[/b:435843f176] | 227.0
-- Uncached Write: 199.86 | [b:435843f176]202.24[/b:435843f176]
-- Uncached Write: 125.13 | [b:435843f176]125.96[/b:435843f176]
-- Uncached Read: [b:435843f176]590.44[/b:435843f176] | 545.85
-- Uncached Read: [b:435843f176]355.02[/b:435843f176] | 344.13
[color=blue:435843f176][b:435843f176]Quartz Graphic Test:[/b:435843f176][/color:435843f176] [b:435843f176]22.49[/b:435843f176] | N/A
[color=blue:435843f176][b:435843f176]OpenGL Test:[/b:435843f176][/color:435843f176] [b:435843f176]2.29[/b:435843f176] | N/A
[color=blue:435843f176][b:435843f176]User Interface Test:[/b:435843f176][/color:435843f176] [b:435843f176]22.15[/b:435843f176] | N/A
[color=green:435843f176][b:435843f176]Cinebench:[/b:435843f176][/color:435843f176] [b:435843f176]7CB-CPU[/b:435843f176] | 5CB-CPU
[color=green:435843f176][b:435843f176]SpeedX:[/b:435843f176][/color:435843f176] [b:435843f176]39.98[/b:435843f176] | 41.45 seconds, [b:435843f176]250.1[/b:435843f176] | 243.2 MIPS @ 10 billion
Ribtorus, I'm amazed you actually seat through Cinebench.
With this new Altivec build, PearPC feels a bit more snappier for overall use.HFU
[/quote:435843f176]
I still find no benefit (apart from XBench scores) to running the Altivec builds. I tried the latest from RichardGoodwin with no significant change to XBench or SpeedX. I also get screen degradation and tiny black dots showing up all over the place with Altivec builds.
- Ribtorus
- Advanced Member
- Posts: 208
- Joined: Sat Aug 28, 2004 12:24 am
Who is online
Users browsing this forum: No registered users and 3 guests