How fast is PearPC: Benchmarks

[User]

To Ribtorus

I was read that you ran Sheepshaver on PearPC I woul'd be want to know how it work subjectively, and how it have the speed?

[Ribtorus]

I tried Sheepshaver inside PearPC and it never ran right. I was unable to test it.

However, Basilisk II works, if you want to try a Psedo-classic environment. As far as speed goes, it's [b:9d097e8563]completely[/b:9d097e8563] useless. It's nothing more than a "see what I can do" thing. I'm pretty sure that Sheepshaver fully running in PearPC would be a [b:9d097e8563]slower[/b:9d097e8563] emulation than basilisk II.

[User]

I knew that in the web was appear new virtual machine it have name 'Mac on Mac'.It is derivative from 'Mac on Linux'. It only run's while Mac os 9.

Here that link

http://babelfish.altavista.com/babelfis ... m%2fmom%2f

[Guru3]

Well, doesn't the classic environment take care of most ppl anyway? As fro running basiliskII... Just run it native on your PC. I just set mine up again yesterday, and am now having a blast playing EVO on the RemoteCVS server... Sorry people that want builds while I'm playing, but you're not getting them .

[User]

To RibTorus

There is a new benchmarking naming 'Let1kWindowsBloom '. It's creates thousands windows, and possible to know os x performance

here that link

http://www.vgg.com/rob/WindowCarbon.sit

And some results

http://www.vgg.com/rob/WindowsBloomUsers.html

[User]

To Ribtorus

When I saw your xbench results , I can to believe that you pc can emulate mac how g3 300. the results in xbench site identical with some g3 and g4 computers with 300 mhz freq.

[Ribtorus]

Yes, in a few benchmarks, my performance is about in the 200 to 350MHz G3 range. PearPC reports a G3 1.06GHz and the host is a 2 GHz dual Opteron.

The new benchmark, Let1kWindowsBloom, took 148 seconds to create and dispose of 1000 windows. I was running a lot of stuff in Windows at the time, and I'll re-test later with a more optimised run.


EDIT:

I re-ran the benchmark with nothing running in the background, and the time improved to 134 seconds to create and dispose of 1000 windows.

I guess this is very much a function of the video x driver PearPC uses.

I'm curious about that particular driver, and why the PearPC author couldn't or wouldn't say more about it in the interview he gave. At least, I [i:07a412ec0a]think[/i:07a412ec0a] it was about the video when he said he was under an NDA (Non Disclosure Agreement) with Apple.

[User]

I am noted that OsX GUI sometimes can be unusable, and slowly because it works in thousands video mode, but when it works in millions mode ,it works good

[Ribtorus]

Here's a series of benchmarks, but this time comparing PearPC 0.4 with [b:acad4a2311]256MB[/b:acad4a2311] ram, [b:acad4a2311]896 MB[/b:acad4a2311] ram and [b:acad4a2311]1.25GB[/b:acad4a2311] ram. The host Windows XP has 1GB ram.

[b:acad4a2311]256MB[/b:acad4a2311] : [b:acad4a2311]896MB[/b:acad4a2311] : [b:acad4a2311]1.25GB[/b:acad4a2311]

SpeedRun (overall score, higher is better)

[b:acad4a2311]87[/b:acad4a2311] : [b:acad4a2311]90[/b:acad4a2311] : [b:acad4a2311]87[/b:acad4a2311]

SpeedX (higher is better)

[b:acad4a2311]300.6[/b:acad4a2311] MIPS : [b:acad4a2311]273.8[/b:acad4a2311] MIPS : [b:acad4a2311]300.8[/b:acad4a2311] MIPS

LetWindowsBloom (lower is better)

[b:acad4a2311]144[/b:acad4a2311] seconds : [b:acad4a2311]134[/b:acad4a2311] seconds : [b:acad4a2311]134[/b:acad4a2311] seconds

PiCalculator (lower is better)

[b:acad4a2311]85[/b:acad4a2311] seconds : [b:acad4a2311]76[/b:acad4a2311] seconds : [b:acad4a2311]75[/b:acad4a2311] seconds

XBench (higher is better)

cpu- [b:acad4a2311]0.57[/b:acad4a2311] : [b:acad4a2311]0.58[/b:acad4a2311] : [b:acad4a2311]0.55[/b:acad4a2311]
thread- [b:acad4a2311]24.9[/b:acad4a2311] : [b:acad4a2311]27.5[/b:acad4a2311] : [b:acad4a2311]27.86[/b:acad4a2311]
memory- [b:acad4a2311]49.24[/b:acad4a2311] : [b:acad4a2311]48.70[/b:acad4a2311] : [b:acad4a2311]50.17[/b:acad4a2311]
Quartz- [b:acad4a2311]23.18[/b:acad4a2311] : [b:acad4a2311]24.52[/b:acad4a2311] : [b:acad4a2311]24.21[/b:acad4a2311]
OpenGL- [b:acad4a2311]2.18[/b:acad4a2311] : [b:acad4a2311]2.16[/b:acad4a2311] : [b:acad4a2311]2.06[/b:acad4a2311]
U/I- [b:acad4a2311]23.60[/b:acad4a2311] : [b:acad4a2311]26.81[/b:acad4a2311] : [b:acad4a2311]25.80[/b:acad4a2311]
Disk- [b:acad4a2311]52.88[/b:acad4a2311] : [b:acad4a2311]47.96[/b:acad4a2311] : [b:acad4a2311]56.49[/b:acad4a2311]

Photoshop CS action script (lower is better)

[b:acad4a2311]18[/b:acad4a2311] min [b:acad4a2311]20[/b:acad4a2311] sec : [b:acad4a2311]14[/b:acad4a2311] min [b:acad4a2311]30[/b:acad4a2311] sec : [b:acad4a2311]13[/b:acad4a2311] min [b:acad4a2311]10[/b:acad4a2311] sec

Application startup times (lower is better)

Photoshop CS- [b:acad4a2311]61[/b:acad4a2311] sec : [b:acad4a2311]60[/b:acad4a2311] sec : [b:acad4a2311]55[/b:acad4a2311] sec
Terminal- [b:acad4a2311]15[/b:acad4a2311] sec : [b:acad4a2311]13[/b:acad4a2311] sec : [b:acad4a2311]20 [/b:acad4a2311]sec
Appleworks 6- [b:acad4a2311]6[/b:acad4a2311] sec : [b:acad4a2311]5[/b:acad4a2311] sec : [b:acad4a2311]7.5[/b:acad4a2311] sec
System prefs- [b:acad4a2311]6[/b:acad4a2311] sec : [b:acad4a2311]6[/b:acad4a2311] sec : [b:acad4a2311]9[/b:acad4a2311] sec
iCal- [b:acad4a2311]8[/b:acad4a2311] sec : [b:acad4a2311]9[/b:acad4a2311] sec : [b:acad4a2311]10[/b:acad4a2311] sec
iMovie- [b:acad4a2311]12[/b:acad4a2311] sec : [b:acad4a2311]15[/b:acad4a2311] sec : [b:acad4a2311]12[/b:acad4a2311] sec
iTunes- [b:acad4a2311]9[/b:acad4a2311] sec : [b:acad4a2311]7[/b:acad4a2311] sec : [b:acad4a2311]7[/b:acad4a2311] sec
iChat- [b:acad4a2311]10[/b:acad4a2311] sec : [b:acad4a2311]6[/b:acad4a2311] sec : [b:acad4a2311]12[/b:acad4a2311] sec
Safari- [b:acad4a2311]11[/b:acad4a2311] sec : [b:acad4a2311]10[/b:acad4a2311] sec : [b:acad4a2311]10[/b:acad4a2311] sec

Boot time (lower is better)
[b:acad4a2311]1 min 24 sec [/b:acad4a2311]: [b:acad4a2311]1 min 33 sec[/b:acad4a2311] : [b:acad4a2311]1 min 30 sec[/b:acad4a2311]

More ram showed significant improvements for Photoshop (1.25GB ram gave a [b:acad4a2311]28[/b:acad4a2311]% improvement over 256MB ram) , LetWindowsBloom (1.25GB gave a [b:acad4a2311]7[/b:acad4a2311]% improvement over 256MB ram) and PiCalculator (1.25GB gave an [b:acad4a2311]11[/b:acad4a2311]% improvement over 256MB ram). Other tests showed quite variable results with no clear advantage.

On my system at least, allocating PearPC/OSX with 256MB ram or 1250MB ram yields only a limited speed advantage. From these benchmarks, it is clear that the CPU and other factors are still the major limiting factor in performance.

[Guru3]

good work with those tests, seems to agree with everything else that we've been seeing