Re: Run OS X at 1/500th speed on your PC!

[Taro Toporific]

PearPC - PowerPC Architecture Emulator

With the price of a used notebook computer nowadays, it's a wonder the anybody bothers with emulators. In case I get bored with my Macs, I snagged an old ThinkPad out of the sodaigomi to run that occasional Windows program I need and I've got a couple of TurboLinux doorstops laying around too.

[Socratesabroad]

Hey, wait a minute?? My Turbolinux paperweight is still on my desk.

[Big Booger]

http://s87840517.onlinehome.us/pearpc.html
There's the install instructions.



http://techzonez.com/forums/showthread.php?t=10087

More information at the link above.

This is excellent. Imagine, great OS, with tons of hardware options Though in this stage it is quite slow.. the OS install alone is going to take 3 hours on a P4 2.8ghz, 1GB DDR PC3200 400Mhz RAM, 160GB SATA 8mb cache drive, ATI Radeon 9600XT,

It's no performer but it works. Pretty amazing coding from the creators of PearPC.

View the PearPC website:
http://pearpc.sourceforge.net/about.html#features

[Big Booger]

Non-pearpc install. Runs like bitch.

XBench was 18.31 result. Kind of low but it's on my spare PC.

[tonikoro]

Hrmmnn.
I guess nobody saw this?
WIRED-News "Mac Hacks Allow OS X on PCs"

[Big Booger]

Hrmmnn.
I guess nobody saw this?
WIRED-News "Mac Hacks Allow OS X on PCs"

http://developer.apple.com/transitionkit.html

mmm...

[tonikoro]

well as you know the transition kit was there first, then some hackers tweeked it to run on other PC cores. In the Wired article, they explain what was hacked to make it run on other processors. Though it is based on software that was obtained (not with leagal standing).

[Big Booger]

well as you know the transition kit was there first, then some hackers tweeked it to run on other PC cores. In the Wired article, they explain what was hacked to make it run on other processors. Though it is based on software that was obtained (not with leagal standing).

Either way it sure works well. I think this will make Apple loads of money and will take market share away from Microsoft if Apple go through and allow it to be run on any Intel Chip.

The thing is with TPM, they are trying to lock it down onto certain hardware... I think that is BS...

Apple should have done this years ago... and now we'd have two excellent OSes all mainstream. Instead they went their own way and represented the fringe of computing users... *sorry to all macaddicts.

[mr. sparkle]

jus' get a mini for cryin' out loud....

[Big Booger]

jus' get a mini for cryin' out loud....

Why should we? When we have intel machines that will run OSX without fail? LOL

I have to say it runs well, and is reported to run even better than some mac machines.

[oyajikun]

I'm running it nativly on my Thinkpad R40. Runs much faster than my friends G5.. he isn't pleased at all. :P

[mr. sparkle]

Oh my god. You guys aren't joking are ya? Sounds like Steve and Co. should go for it if it's already being leaked on the web. Can't fight City Hall.

[Big Booger]

Driver support is nil, but there are coders working on it. Sound and video support are the two main obstacles... but considering this is a developers hack or whatever you want to call it, it's amazing.

It shows the Power of OSX when on an intel. Even with my 4 year old P4 1.6ghz machine, it runs smooth. I get a 18.00+ score with xbench. And almost everything besides sound works. Fucking SBlive cards... (they are not even supported natively on OSX)

[oyajikun]

I get 100% Hardware Acceleration with my ATI Driver, sound works, LAN, Bluetooth keyboard and mouse.. the only thing that isn't functional is my Wireless LAN.

I can run most Power PC applications using Rosetta but there are a few apps that won't run such as Toast Titanium and Finalcut 5.

[Caustic Saint]

I can run most Power PC applications using Rosetta but there are a few apps that won't run such as Toast Titanium and Finalcut 5.

Are we to assume you've bought an Apple developer license, as well as those two apps?

If not, this topic has gone deep into warez territory and isn't long for the boards.

The original discussion of PearPC is one thing, as it's loading legit copies of OSX using a PowerPC emulator. What you guys are discussing is an entirely different subject.[/img]

[Charles]

..Even with my 4 year old P4 1.6ghz machine, it runs smooth. I get a 18.00+ score with xbench...

Yeah, that falls right in line with the reports that an old Stinkpad beats a G5.

NOT.

G5 xBench scores go up to 158.

Actually, an 18 on xBench is about as fast as an old Powerbook G3/500 "Pismo." You should be proud, you beat a 500mhz laptop with a chip from 2 generations ago!

[Big Booger]

..Even with my 4 year old P4 1.6ghz machine, it runs smooth. I get a 18.00+ score with xbench...

Yeah, that falls right in line with the reports that an old Stinkpad beats a G5.

NOT.

G5 xBench scores go up to 158.

Actually, an 18 on xBench is about as fast as an old Powerbook G3/500 "Pismo." You should be proud, you beat a 500mhz laptop with a chip from 2 generations ago!

Charles you fail to realise that this developer pack has not been optimised for Intel machines.. nor AMD. They are using a Developer package that has been hacked and using patched SSE3 capabilities.

If you want to see much higher Xbench results on SSE3 compatible processors:

http://www.360hacker.net/forums/viewtopic.php?t=379

The highest I saw so far is a 60.28, which nearly blew away a G5. ANd then you consider the price you paid for your "Pismo" and what I paid for my 4 year old P4. ]The following are results for my PowerMac G5 1.6 Ghz.

NOTE: I paid $3200 for the system when it first came out, then another $900 on RAM, another $150 for the Harddrive, and $1300 for my Studio Display (New) (I'm getting the feeling it was a little bit wasted Sad )

Results 66.26
System Info
Xbench Version 1.2
System Version 10.4 (8A428)
Physical RAM 2048 MB
Model PowerMac7,2
Processor PowerPC 970 @ 1.60 GHz
L1 Cache 64K (instruction), 32K (data)
L2 Cache 512K @ 1.60 GHz
Bus Frequency 800 MHz
Video Card GeForce FX 5200
Drive Type WDC WD1600JD-00HBB0
CPU Test 78.84
GCD Loop 76.42 4.03 Mops/sec
Floating Point Basic 80.13 1.90 Gflop/sec
vecLib FFT 80.81 2.67 Gflop/sec
Floating Point Library 78.14 13.61 Mops/sec
Thread Test 45.61
Computation 41.30 836.62 Kops/sec, 4 threads
Lock Contention 50.93 2.19 Mlocks/sec, 4 threads
Memory Test 75.94
System 76.96
Allocate 94.98 348.80 Kalloc/sec
Fill 83.80 4074.63 MB/sec
Copy 60.54 1250.37 MB/sec
Stream 74.94
Copy 75.45 1558.39 MB/sec [G5]
Scale 75.35 1556.65 MB/sec [G5]
Add 74.76 1592.56 MB/sec [G5]
Triad 74.22 1587.72 MB/sec [G5]
Quartz Graphics Test 71.06
Line 70.86 4.72 Klines/sec [50% alpha]
Rectangle 73.37 21.91 Krects/sec [50% alpha]
Circle 73.35 5.98 Kcircles/sec [50% alpha]
Bezier 77.82 1.96 Kbeziers/sec [50% alpha]
Text 61.95 3.88 Kchars/sec
OpenGL Graphics Test 103.30
Spinning Squares 103.30 131.04 frames/sec
User Interface Test 54.83
Elements 54.83 251.64 refresh/sec
Disk Test 62.98
Sequential 110.42
Uncached Write 114.78 70.47 MB/sec [4K blocks]
Uncached Write 105.25 59.55 MB/sec [256K blocks]
Uncached Read 110.53 32.35 MB/sec [4K blocks]
Uncached Read 111.55 56.06 MB/sec [256K blocks]
Random 44.05
Uncached Write 15.73 1.67 MB/sec [4K blocks]
Uncached Write 109.22 34.97 MB/sec [256K blocks]
Uncached Read 94.33 0.67 MB/sec [4K blocks]
Uncached Read 133.70 24.81 MB/sec [256K blocks]
------------------------------------------------------------------------------------

The base benchmark for Xbench is the following on a PowerMac G5 Dual 2.0 Ghz....

Results 100.00
System Info
Xbench Version 1.2
System Version 10.4.2 (8C46)
Physical RAM 1536 MB
Model PowerMac7,3
Processor PowerPC G5x2 @ 2.00 GHz
L1 Cache 64K (instruction), 32K (data)
L2 Cache 512K @ 2.00 GHz
Bus Frequency 1 GHz
Video Card GeForce 6800 Ultra
Drive Type Maxtor 6B300S0
CPU Test 100.00
GCD Loop 100.00 5.27 Mops/sec
Floating Point Basic 100.00 2.38 Gflop/sec
vecLib FFT 100.00 3.30 Gflop/sec
Floating Point Library 100.00 17.41 Mops/sec
Thread Test 100.00
Computation 100.00 2.03 Mops/sec, 4 threads
Lock Contention 100.00 4.30 Mlocks/sec, 4 threads
Memory Test 100.00
System 100.00
Allocate 100.00 367.23 Kalloc/sec
Fill 100.00 4862.23 MB/sec
Copy 100.00 1571.73 MB/sec
Stream 100.00
Copy 100.00 2065.46 MB/sec [G5]
Scale 100.00 2065.97 MB/sec [G5]
Add 100.00 2130.21 MB/sec [G5]
Triad 100.00 2139.25 MB/sec [G5]
Quartz Graphics Test 100.00
Line 100.00 6.66 Klines/sec [50% alpha]
Rectangle 100.00 29.86 Krects/sec [50% alpha]
Circle 100.00 8.15 Kcircles/sec [50% alpha]
Bezier 100.00 2.52 Kbeziers/sec [50% alpha]
Text 100.00 6.26 Kchars/sec
OpenGL Graphics Test 100.00
Spinning Squares 100.00 126.86 frames/sec
User Interface Test 100.00
Elements 100.00 458.95 refresh/sec
Disk Test 100.00
Sequential 100.00
Uncached Write 100.00 61.40 MB/sec [4K blocks]
Uncached Write 100.00 56.58 MB/sec [256K blocks]
Uncached Read 100.00 29.27 MB/sec [4K blocks]
Uncached Read 100.00 50.26 MB/sec [256K blocks]
Random 100.00
Uncached Write 100.00 10.59 MB/sec [4K blocks]
Uncached Write 100.00 32.01 MB/sec [256K blocks]
Uncached Read 100.00 0.71 MB/sec [4K blocks]
Uncached Read 100.00 18.56 MB/sec [256K blocks]
------------------------------------------------------------------------------------[/quote]
ANd now let's compare it to an AMD SS3 compatible processor that costs far less:

Results 60.28
System Info
Xbench Version 1.2
System Version 10.4.1 (8B1025)
Physical RAM 1024 MB
Model ACPI
Drive Type Maxtor 6E040L0
CPU Test 35.49
GCD Loop 118.23 6.23 Mops/sec
Floating Point Basic 78.05 1.85 Gflop/sec
vecLib FFT 12.72 419.78 Mflop/sec
Floating Point Library 77.87 13.56 Mops/sec
Thread Test 77.44
Computation 54.97 1.11 Mops/sec, 4 threads
Lock Contention 131.00 5.64 Mlocks/sec, 4 threads
Memory Test 87.12
System 66.40
Allocate 236.47 868.39 Kalloc/sec
Fill 49.32 2398.17 MB/sec
Copy 48.36 998.79 MB/sec
Stream 126.63
Copy 120.33 2485.42 MB/sec
Scale 118.06 2439.04 MB/sec
Add 135.77 2892.25 MB/sec
Triad 134.38 2874.78 MB/sec
Quartz Graphics Test 81.38
Line 65.86 4.38 Klines/sec [50% alpha]
Rectangle 78.35 23.39 Krects/sec [50% alpha]
Circle 78.66 6.41 Kcircles/sec [50% alpha]
Bezier 101.79 2.57 Kbeziers/sec [50% alpha]
Text 91.25 5.71 Kchars/sec
OpenGL Graphics Test 43.80
Spinning Squares 43.80 55.56 frames/sec
User Interface Test 96.10
Elements 96.10 441.03 refresh/sec
Disk Test 55.45
Sequential 109.18
Uncached Write 91.51 56.18 MB/sec [4K blocks]
Uncached Write 96.95 54.85 MB/sec [256K blocks]
Uncached Read 161.22 47.18 MB/sec [4K blocks]
Uncached Read 108.81 54.68 MB/sec [256K blocks]
Random 37.16
Uncached Write 13.39 1.42 MB/sec [4K blocks]
Uncached Write 81.59 26.12 MB/sec [256K blocks]
Uncached Read 80.56 0.57 MB/sec [4K blocks]
Uncached Read 120.38 22.34 MB/sec [256K blocks]
_________________
Ciao,
Alejandra.

Mac OS X X86 working in:
VMWare,
Native on AMD64 3500+ (ClawHammer SSE2 only) on Nforce 4 MoBo and
Celeron D (SSE3 supported) On a VIA Chipset Mobo.

[Big Booger]

I can run most Power PC applications using Rosetta but there are a few apps that won't run such as Toast Titanium and Finalcut 5.

Are we to assume you've bought an Apple developer license, as well as those two apps?

If not, this topic has gone deep into warez territory and isn't long for the boards.

The original discussion of PearPC is one thing, as it's loading legit copies of OSX using a PowerPC emulator. What you guys are discussing is an entirely different subject.[/img]

I think if you own tiger, then there shouldn't be a problem.

[Caustic Saint]

I can run most Power PC applications using Rosetta but there are a few apps that won't run such as Toast Titanium and Finalcut 5.

Are we to assume you've bought an Apple developer license, as well as those two apps?

If not, this topic has gone deep into warez territory and isn't long for the boards.

The original discussion of PearPC is one thing, as it's loading legit copies of OSX using a PowerPC emulator. What you guys are discussing is an entirely different subject.[/img]

I think if you own tiger, then there shouldn't be a problem.

There's a big difference between a Tiger install disc you bought (or that came with a Mac) and a Tiger for Intel ISO "found" on the net.

[Charles]

Charles you fail to realise that this developer pack has not been optimised for Intel machines.. nor AMD. They are using a Developer package that has been hacked and using patched SSE3 capabilities.

If you want to see much higher Xbench results on SSE3 compatible processors:

http://www.360hacker.net/forums/viewtopic.php?t=379

The highest I saw so far is a 60.28, which nearly blew away a G5.

Why actually, the Developer Kit IS optimised for Intel, that's the WHOLE POINT of the developer kit, to enable you to develop optimized Intel MacOS X code.

Somehow, I missed the bit where you proved that 60 is a larger number than 158.

[Big Booger]

Charles you fail to realise that this developer pack has not been optimised for Intel machines.. nor AMD. They are using a Developer package that has been hacked and using patched SSE3 capabilities.

If you want to see much higher Xbench results on SSE3 compatible processors:

http://www.360hacker.net/forums/viewtopic.php?t=379

The highest I saw so far is a 60.28, which nearly blew away a G5.

Why actually, the Developer Kit IS optimised for Intel, that's the WHOLE POINT of the developer kit, to enable you to develop optimized Intel MacOS X code.

Somehow, I missed the bit where you proved that 60 is a larger number than 158.

Again you're missing a crucial point Charles.... The developer KIT is for development... ie. not optimised, not even fully supported. Optimised is different from compatible Charles... The developer kit is a kit to allow you to develop the code to optimize it.. The Intel and AMD scores reflect a base code install, without optimization.

And the top scores I saw were even higher than what you posted. I think I saw one clocked at 164 on slashdot:

http://macslash.org/article.pl?sid=03/06/24/2219201&mode=thread

And now for the reality:

http://ladd.dyndns.org/xbench/csi.xhtml?machineTypeID=22
Average score for shown machines: 78.55



So in essence, Intel's not that far behind.. And that's with a developmental OS....

Who the hell runs machines on an everyday G5 Powermac that scores at the 158 Xbench range? I am sure with a Dual Xeon, or Dual Opteron, the scores will increase.

The machines you are quoting the 158 bench from run this kind of hardware:

Xbench Version

1.2
System Version

10.4.2 (8C46)
Physical RAM

4096 MB
Model

PowerMac7,3
Processor

PowerPC G5x2 @ 2.50 GHz

It's like comparing a 4 banger to a V10.... A 5ghz PC with 4GB of ram should pounce a 3ghz machine with 1GB of ram...

[Big Booger]

I can run most Power PC applications using Rosetta but there are a few apps that won't run such as Toast Titanium and Finalcut 5.

Are we to assume you've bought an Apple developer license, as well as those two apps?

If not, this topic has gone deep into warez territory and isn't long for the boards.

The original discussion of PearPC is one thing, as it's loading legit copies of OSX using a PowerPC emulator. What you guys are discussing is an entirely different subject.[/img]

I think if you own tiger, then there shouldn't be a problem.

There's a big difference between a Tiger install disc you bought (or that came with a Mac) and a Tiger for Intel ISO "found" on the net.

I agree totally. BUt a tiger is a tiger regardless of the stripes it has.

[FG Lurker]

Actually the developer kit is probably fairly well optimized for the hardware it is running on. Intel chips have high clock speeds without a huge amount of performance, and comparing GHz to GHz across architectures is meaningless. For example, an Athlon 64 3500+ actually runs at 2.2GHz but performance-wise is very similar to a 3.6GHz P4.

So, if you allow for the fact that a single 3.6GHz P4 is probably somewhere close in performance to a single 2 ~ 2.2GHz G5 and then run numbers...they will probably come out about the same.

What I would like to see is OSX running on a dual CPU dual-core Opteron setup...