Become a fan of Slashdot on Facebook


Forgot your password?
OS X Businesses Operating Systems Software Windows Apple

Cherry OS Claims Mac OS X Capability For x86 1090

jediboytj writes "According to the MacWorld Article, Cherry OS, does what Virtual PC does for Macs, only the opposite. PC Users are now able to run Mac OSX at G4 Speeds (Company claims 80% of the speed of your PC). It also includes full hardware support: hard drive, CPU, RAM, FireWire, USB, PCI, PCMCIA bus, Ethernet networking and modem. The software is being distributed through electronic download at $49.99 USD..." Note: it does not come with a copy of any Apple OS. Anyone in Windowsland tried it to provide a thumbs up (or down)?
This discussion has been archived. No new comments can be posted.

Cherry OS Claims Mac OS X Capability For x86

Comments Filter:
  • Re:Finally... (Score:2, Informative)

    by lakiolen ( 785856 ) on Tuesday October 12, 2004 @01:55PM (#10505097)
    Not counting the cost of "buying" OS X.
  • Re:Oh Boy! (Score:5, Informative)

    by drinkypoo ( 153816 ) <> on Tuesday October 12, 2004 @01:56PM (#10505116) Homepage Journal
    If you put the laserwriter on an etherprint box of some sort, you can print to it using Linux (or Solaris) with netatalk. Laserwriters speak Postscript so nothing could be easier to support once you get them talking. I did it back in the days when a 486 was a tolerably fast computer and it only took me a few hours to get running including compiling the software and building a new kernel with appletalk support.

    I know you were just being a smartass, a time-honored tradition around here, but I couldn't pass up the opportunity to be informative.

  • MirrorDot is Useful (Score:2, Informative)

    by Didion Sprague ( 615213 ) * on Tuesday October 12, 2004 @01:56PM (#10505117)
    I know it's more fun to bitch and moan about the original site being slashdotted, but if you want to RTFA, then simply go to mirrordot: []

    Enough already.
  • Re:Finally... (Score:5, Informative)

    by over_exposed ( 623791 ) on Tuesday October 12, 2004 @01:56PM (#10505133) Homepage
    You still need to buy a copy of OSX. It's gonna run you a *tad* more than $50...
  • by Sc00ter ( 99550 ) on Tuesday October 12, 2004 @01:59PM (#10505169) Homepage
    They have the base for x86 (Darwin). It's just the UI that's for macs only.
  • by Datasage ( 214357 ) * <`Datasage' `at' `'> on Tuesday October 12, 2004 @01:59PM (#10505175) Homepage Journal
    Not only that, but there is no information about the company on the website.

    Its got that feeling of an overnight company. The whois record was only registered in july.

    It wouldnt supprise me if its some company that took pearPC and is trying to sell it.
  • by drinkypoo ( 153816 ) <> on Tuesday October 12, 2004 @02:00PM (#10505196) Homepage Journal
    Saying it's not possible to emulate a G4 at 80% speed is kind of arrogant, don't you think? It's probable that most Altivec instructions will map to SSE2/3 or 3DNow! without much translation at all. A vector operation is a vector operation is a vector operation, after all. An author of PearPC claims that register starvation isn't a serious problem either, so I'm not sure what law says that you can't emulate a G4 at 80% speed...
  • I have no idea (Score:4, Informative)

    by antifoidulus ( 807088 ) on Tuesday October 12, 2004 @02:00PM (#10505199) Homepage Journal
    if this will work or not, but if it does, Apple legal won't be happy. The EULA states that you have to run OS X on Apple branded hardware(probably to kill clones), now I am willing to bet for the time being anyway, Apple will look the other way on non-commerical projects like Pear PC, but they probably won't be very keen on a commericial product that violates the EULA.
  • by TitusC3v5 ( 608284 ) on Tuesday October 12, 2004 @02:01PM (#10505215) Homepage
    MXS Inc. announces CherryOS 1.0 October, 08 2004

    NEWS RELEASE FOR IMMEADIATE RELEASE Contact: Jim Kartes, 866-661-5699 Media contact same.

    Maui, HI (DATE) MXS today announce the immediate availability of Cherry OS software . Cherry OS is a software translator that allows you to install Apple's Operating System on x86 computer architecture. To put it simply you can now run Apple's award winning Panther OS on your PC! This breakthrough in OS development now gives home users, software developers and web designer's ultimate flexibility in both the operating system and hardware platform you use for your personal computer or testing environment.

    Cherry OS runs Panther as a virtual machine on your Windows PC. This virtual machine has full network capabilities including the ability to share folders and access the web. The virtual machine also has complete access to the computer's hardware resources including, Hard Drive, CPU, RAM, Firewire, USB, PCI, PCMIA BUS and RJ45/Ethernet and Modem.

    Arben Kryeziu, Cherry OS inventor and a software developer, got tired of carrying both a Mac and a PC around with him, so he invented Cherry OS. "Think about it," says Arben. "Now about 600 million PC users can have the MAC advantage. One computer to use all software and if PC users would use MAC software to get email, perhaps they would avoid viruses, Trojans and spy-ware." He went on to say that , "You can build and test applications for a Mac on your development PC, test web site design for Mac web browsers without having to buy the hardware, run OS X, the world's best Operating System, on a less expensive hardware platform and use your favorite Mac apps on a PC."

    Pricing and availability
    Cherry OS is now available only on line at as a download, for $49.95. (Mac software not included)

    About MXS
    MXS is a software development company specializing in video streaming software. ranked our vx30 encoder as the best in the world.

    The products of Maui X-Stream can be viewed on
  • by Anonymous Coward on Tuesday October 12, 2004 @02:02PM (#10505232)
    unfortunately, mirrordot didn't mirror :P
  • Try this instead: (Score:4, Informative)

    by temojen ( 678985 ) on Tuesday October 12, 2004 @02:04PM (#10505256) Journal
    PearPC [], same thing only open source, free, and runs on Windows and Linux.
  • Re:But why? (Score:5, Informative)

    by vhold ( 175219 ) on Tuesday October 12, 2004 @02:06PM (#10505286)
    The best reason I can think of is that all the things I'd want to use a Mac for, are almost totally not CPU bound, whereas all the things I use my PC for are massively GPU/CPU related (games). So basically, I could have most of the best of both worlds in one box. Mac for everything internet/creativity related, and the PC for games/proprietary-work-apps.

    There are lots of other reasons you could contrive, what if you had Mac friends that visit a lot but constantly lament being unable to use your PC? It fundamentally boils down to you wanting _both_, but you need more performance on the PC side, which I really think is more common of a case, just on games alone.
  • Re:But why? (Score:5, Informative)

    by temojen ( 678985 ) on Tuesday October 12, 2004 @02:13PM (#10505366) Journal
    Is there Mac-only software that people would want to run on their x86 machine?

    Absolutely: Safari, Camino, and ie/Mac. Web developers can see what their site will look like and how it will function on a Mac without needing to get more hardware.

    I used to run Win2k on VMWare on Linux so I could see how my sites would look on a PC.

  • Re:Finally... (Score:5, Informative)

    by Kenja ( 541830 ) on Tuesday October 12, 2004 @02:21PM (#10505456)
    "Yay permissive Apple licensing!"

    According to the license you cant run the OS on an emulator because its not "Apple hardware".

  • Re:Try this instead: (Score:5, Informative)

    by isecore ( 132059 ) <isecore.isecore@net> on Tuesday October 12, 2004 @02:23PM (#10505477) Homepage
    Yes, but PearPC doesn't come close to CherryOS so far unproven claim of speed.

    PearPC does run Mac OS X, but at an absolute snails pace (Yes, I've tried it - Three hours to install, approx 1-2 minutes to open a finder-window).

    If CherryOS indeed runs it at a somewhat decent G4-ish speed I'd almost consider 50 bucks to be worth it.
  • by Corrado ( 64013 ) <> on Tuesday October 12, 2004 @02:23PM (#10505487) Homepage Journal
    Click here [] for some screenshots and a running commentary.
  • Re:Try this instead: (Score:3, Informative)

    by Ford Prefect ( 8777 ) on Tuesday October 12, 2004 @02:23PM (#10505489) Homepage
    Can I expect Pear to run like an 800 MHz PPC? 300 MHz maybe?

    Maybe 3MHz. Ish. ;-)
  • by Anonymous Coward on Tuesday October 12, 2004 @02:25PM (#10505512)
    arrrgh my eyes! that has got to be some of the worst color choices i have seen since the debute of
  • by rsax ( 603351 ) on Tuesday October 12, 2004 @02:27PM (#10505535)
    The request for caching content on foreign websites has been covered before so I won't ask why /. doesn't cache sites locally. But it wouldn't have taken much effort by either the editors or submitter to append to []

    Use Coral CDN []! It works and it's available, no excuses except laziness.

  • by DarkBlackFox ( 643814 ) on Tuesday October 12, 2004 @02:33PM (#10505615)
    I managed to get to the documentation page by refreshing rapidly. The manual is avaliable online, and hosted on a differant server. It's a 1.7 mb download, but includes screenshots and information.

    Manual avaliable here:

    or as a .doc
  • by zach_smith ( 159760 ) on Tuesday October 12, 2004 @02:34PM (#10505626)
    I'll bet they will have amazingly similar speed and functionality. How much do you want to bet that CherryOS is ripping off PearPC source code without giving them credit?
  • by Saeger ( 456549 ) < minus author> on Tuesday October 12, 2004 @02:35PM (#10505640) Homepage
    PearPC is really, really, really slow, though it is making progress. For comparison against CherryOS's claim that it runs the guest 80% as fast as the host, it takes PearPC over 5 minutes just to boot MacOSX 10.3 on my 1.2GHz Athlon running Suse9.1 (kernel 2.6.5).

    PearPC is free/FREE, though, and I only use it for Safari compatibility testing, so its speed isn't a major issue for me.


  • by Jim McCoy ( 3961 ) on Tuesday October 12, 2004 @02:40PM (#10505671) Homepage
    The GUI -- windows, mouse for control, pop-ups, etc. -- was invented by Dr. Douglas Engelbart at SRI in the 60s. It was Xerox who applied the metaphor to the PC, added overlapping windows and the LAN, and then coupled it with a development environment that was more that one-off coding hacks (important to be sure, but not close to "inventing" the GUI.)
  • by 0x0d0a ( 568518 ) on Tuesday October 12, 2004 @02:41PM (#10505696) Journal
    That's because it's not possible to get 80% speed with an emulator as described. You *cannot* do this on a PowerPC emulator hosted on an x86 system. Even ignoring things like the fact that the endianness of all integer values is reversed, the PowerPC has several times more general-purpose registers than the x86. Even if the emulation system has zero overhead for its own code, you're going to have to be pulling registers in and out of main memory, which is going to be vastly slower -- that will immediately cut you down to a small fraction of the performance.

    It *might* be possible to write a compiler that can build x86 binaries with PPC binaries as input. It would be hard and the performance would probably still suck, but this is the route that will give the best performance. There has to be a lot of register usage analysis that needs to be done to get something like this even remotely usable, and you are going to want to do this statically.

    If someone ran out and made a legitimate system like this, several things would be true:

    1) These people would probably be from a compiler company, because the work that needs to be done to do this efficiently is *hard* and requires a lot of techniques that compilers use.

    2) If this is a commercial project (i.e. people are actually serious about making money and not getting hit by lawsuits), they would have gotten an OK from Apple and Apple would have made noise promoting this. Why? The only practical reason to build a modern Mac emulator is to run Mac OS X, which, on non-Apple hardware, is a violation of Apple's EULA.

    3) The ROM problem is still present -- you can't make a Mac emulator legally without the Mac ROMs, which Apple keeps copyrighted. -- see #2.
  • Re:So, you're asking (Score:4, Informative)

    by !Freeky2BGeeky ( 657344 ) on Tuesday October 12, 2004 @02:43PM (#10505708)
    but that's linux/ppc not linux/X86
  • Scam alert (Score:5, Informative)

    by saddino ( 183491 ) on Tuesday October 12, 2004 @02:48PM (#10505774)
    Hmm, their main page states, under "Screenshots":

    Desctop & Task Manager

    and under "What can CherryOS do?":

    Skin enadled GUI

    But beyond the typos, their "Client Showcase" features a testimonial from "Secnet Q&A Services" which Google doesn't have any information on (hmm, a Q&A company without a web presence?).

    My guess either an out-and-out scam, or a an attempt to pawn off a modified copy of PearPC in an attempt to generate some $ and scram. Ballsy.
  • by wulfhound ( 614369 ) on Tuesday October 12, 2004 @02:54PM (#10505833)
    ... except that modern superscalar CPUs (certainly x86, and possibly newer PPCs also) don't work like that - the registers you write to in machine code are virtual, and are mapped on to a larger hidden register file in realtime by the CPU. In any case a sure-fire L1 cache hit has negligible latency compared to, well, pretty much anything else on an Intel cpu.
  • by Ignignot ( 782335 ) on Tuesday October 12, 2004 @02:55PM (#10505839) Journal
    Never mind the shortage of general purpose registers on x86 and the lack of a direct mapping between instruction sets

    I won't, because the x86 line has lots of general purpose registers now. They just pretend to be whatever special purpose ones the programs need (if any). We've come a long way since the 386.
  • by Pius II. ( 525191 ) <{PiusII} {at} {}> on Tuesday October 12, 2004 @02:55PM (#10505841)
    There are no Mac ROMs, and there haven't been any since at least 1998.
    Even the classic Mac OS didn't need the ROMs anymore in its last incarnation.
    The less-than-modern Macs had driver support for booting in its ROM, and loaded the Toolbox from a file in the system folder (it's named "Mac OS ROM", though). Modern Macs use OpenFirmware [], which is, as the name says, open. Moreover, it's easily emulated, allowing for running OS X on arbitrary PPC machines (with MOL []). Yes, that means e.g. Genesis or AmigaOne boards. Or anything with a PPC, really.
  • by civilizedINTENSITY ( 45686 ) on Tuesday October 12, 2004 @03:00PM (#10505902)
    The Darwin team is indebted to a diverse collection of open source projects, including the following:

    - Mach, which was originally developed by Project Mach at Carnegie-Mellon University, and later enhanced by the Open Software Foundation (now The Open Group).:

    - 4.4BSD-Lite2, originated in UC Berkeley's Computer Systems Research Group and developed by a large number of contributors::

    * FreeBSD, the primary reference platform for Darwin's BSD kernel development.:
    * NetBSD, the upstream source for a significant portion of Darwin's user-space commands and tools.
    * OpenBSD, with its focus on robustness and security and its integrated cryptography, provides OpenSSH for secure remote access.

    - Apache HTTPD, the world's most popular web server, is included as part of the Darwin distribution, making Apple the largest distributor of Apache.

    Getting from 4.4BSD-Lite2 to Darwin seems to have had contributions [] from both FreeBSD and NetBSD.
    In 1997, Apple Computers, which had an interest in BSD and Unix after having bought NeXT in December 1996, produced a 4.4BSD-Lite2 derivation named Rhapsody in 1997. This eventually evolved, with help from the FreeBSD and NetBSD projects, into Darwin, a system with a MACH microkernel wrapped with a 4.4BSD-Lite2 kernel API. FreeBSD project cofounder and longtime core team member Jordan Hubbard headed this project. Darwin forms the heart of the Mac OS X line of operating systems.
  • by KH ( 28388 ) on Tuesday October 12, 2004 @03:01PM (#10505915)
    I just Xbenched my installation of Mac OS X on PearPC over WinXP.

    It's an AthlonXP 3000 (oced to 2400MHz or thereabout) box with 1GB RAM. I've assigned 512MB for PearPC.

    The overall score is indeed abysmal 2.89. For comparison, my PB 12" (867MHz) gets something in the range of 80, I think.

    But if I look at the score more closely, I notice that major drag comes from vecLib FFT test (scored 0.15!) and all kinds of graphics test (OpenGL test being the worst).

    For other things, it scores about 30 to 60 scores range. Disk test is pretty impressive. I only have a regular ATA drive on my PC. Got the score better than my PB disk.

    These results are quite understandable considering what PearPC is doing. I would say for some tasks, this might even be usable.

    Very impressive, I must say.
  • by Danathar ( 267989 ) on Tuesday October 12, 2004 @03:05PM (#10505957) Journal
  • Re:Thievery (Score:1, Informative)

    by Anonymous Coward on Tuesday October 12, 2004 @03:21PM (#10506109)
    You are taking something that does not belong to you

    No you're not. You're copying it. Copying is different from taking. Words matter. I understand that you'd like to apply the connotation of "steal" and "take," to copyright infringement, because "infringement" doesn't sound as bad. But tough shit. Use the right words.

  • Re:Thievery (Score:2, Informative)

    by micromoog ( 206608 ) on Tuesday October 12, 2004 @03:39PM (#10506336)
    Sorry d00d, but you're still wrong. It's illegal and arguably immoral, but it's still not the same thing as theft or stealing.
  • by kommakazi ( 610098 ) on Tuesday October 12, 2004 @03:46PM (#10506434)
    No the genesis had a 68000 series processor, not a PPC. Old macs weren't powermacs either, they were also 68000 series processors.
  • by darkstream ( 652288 ) on Tuesday October 12, 2004 @03:48PM (#10506467) Homepage Journal
    I produce fractal art and need to use a PC to do it (existing fractal apps on the Mac don't compare with Fractal eXtreme [] or UltraFractal [] - and yes, I've tried several dozen), so using VirtualPC seemed the best choice when I switched over in 2000. But the best configuration for this process was VirtualPC4 under OS9. VirtualPC just didn't make the transition from OS9 to OSX very well. That meant I had to reboot into OS9 just for fractal exploration. Having migrated entirely to OSX over the years, working in OS9 was difficult. All the apps I used were in OSX! I soon was forced to get a PC just for fractal exploration. The GUI was sluggish in VirtualPC6 under OSX and the rendering times were abominable.

    Honestly, anything that requires heavy calculations is either going to break the emulator or run abysmally slow. Although email and web browsing can be tollerable (I often proof webpages using VirtualPC to get a view from the other side of the pond), I can't see any of the iLife apps being usable under CherryOS. They typically tax my 800MHz iMac. I can't imagine how slowly they would run under emulation...

  • by kuwan ( 443684 ) on Tuesday October 12, 2004 @03:54PM (#10506545) Homepage
    I have not worked with or even looked at either instruction set. Nonetheless your average application will spend only a small percentage of its time using Altivec...

    Maybe you should go and get some experience or at least some knowledge before you start talking about something you know nothing about.

    Altivec from its beginning introduced 162 vector instructions that have not changed from the initial G4 to the current G5. On the other hand, Intel's MMX/SSE/SSE2 instructions have evolved over time - roughly 57 in MMX, 78? in SSE and 144 in SSE2. Altivec has been a well-designed and versatile SIMD engine from its beginning while Intel has sort of hacked together their SIMD engine as they've evolved their processors. Intel's implementation is very troublesome for a programmer because he has to do many different things depending on what is available (MMX/SSE/SSE2). These instructions don't map 1:1 for the most part with Altivec. And while SSE2 is much better than SSE, it was only introduced with the Pentium 4.

    Also, Altivec has 32 128-bit registers to only 8 128-bit registers for SSE/SSE2. I don't care what anyone says, trying to emulate 32 registers (when all you have is 8) in an SIMD engine is going to be a lot slower.

    You say that only a small percentage of time will be spent using Altivec, but that's just not true. Apple has optimized a large part of Mac OS X to use Altivec, especially in Quartz (the windowing and compositing engine). This would result in a major slowdown for any emulator in pretty much every application (except for stuff like background daemons). You'd probably do better just to emulate a G3 so as to not run any Altivec code.
  • by adzoox ( 615327 ) on Tuesday October 12, 2004 @04:35PM (#10507087) Journal
    I'm actually going to reveal something you may not know...

    Both Connectix and Insignia (the two main companies that produced Windows emulation for the Mac) were actually just venture capital firms. This is why Connectix, at the height of every product launched, would just sell it off as an asset.

    Connectix Quickcam = Logitech Quickcam
    Connectix Virtual Game Station = Sony Buyout
    Connectix Virtual PC ( at an undeniable breakthrough point) = Microsodt VPC

    Insignia was the same:

    Softwindows ... I did an unupublished interview with the head of FWB ... he stated that they simply licensed the code rather than bought it from Insignia. The reason they never released an update after leasing the code was because they didn't see any merit at the time in releasing a new OS X version.

    Insignia is supposedly shopping this around.

    I have found that these two companies were essentially started up by venture capital and paid off their investors, dumped their employees, and the owners got filthy rich.

    Now, as for this software. I find it NEXT to impossible that the software is running a G4 at 80% speed of the CPU. If you were to translate this properly - Apple's CPUs are about 1.2X as fast as the equivalent P4 and P3 (G3 & G4 respectively) - so essentially the claim is saying it will run a 100% equivalent Mhz / speed ratio.

    This means if I had a 3Ghz Pentium 4 with 1 Gig RAM - I would have the equivalent of a 2.4Ghz G4!! There's just NO way!

  • by Ignignot ( 782335 ) on Tuesday October 12, 2004 @04:47PM (#10507231) Journal
    they're usable as either. All of the general purpose registers on the current x86 line can be used as either specialized registers (for example, one to hold a memory offset) or as general purpose registers. You can see the difference in different binaries that you use in linux - 386 binaries are going to have references to specialized registers, while 686 binaries will not. The 686 also has a larger instruction set, but that's neither here nor there.
  • by bedouin ( 248624 ) on Tuesday October 12, 2004 @05:03PM (#10507428)
    Disk test is pretty impressive. I only have a regular ATA drive on my PC. Got the score better than my PB disk.

    Probably because the PowerBook (and most laptops) only have 4200 RPM drives in their default configurations.
  • by Anonymous Coward on Tuesday October 12, 2004 @05:09PM (#10507487)
    Having actually written software for MMX/SSE/SSE2 and AltiVec, I can tell you with authority that it is often more problematic to map things to the Intel instruction set because of the limited register size. With MMX, it was especially bad, often requiring 3-4x as many instructions and running correspondingly slower. There is no way you're going to emulate AltiVec at close to hardware speed.
  • by meme_police ( 645420 ) on Tuesday October 12, 2004 @05:15PM (#10507559)
    Like the parent says, they use OpenFirmware (which is a fully programmable Forth environment) now instead of a closed ROM. Nobody is arguing that Apple is no longer using ROM, they're just saying that they're no longer using the closed Mac ROM.
  • Re:So, you're asking (Score:1, Informative)

    by GuineaPigMan ( 663444 ) on Tuesday October 12, 2004 @05:29PM (#10507682)
    Once upon a time, Apple mad an x86 OS called Rhapsody: []
    And it died ca. 1997: _apples_rhapsody/ []
  • by addaon ( 41825 ) <> on Tuesday October 12, 2004 @05:43PM (#10507832)
    Rename registers cannot be accesssed explicitly. The processor can use them so that a single named register (say, eax) maps to more than one rename register (say, numbers 7, 13, and 22) in different in-flight (that is, currently-being-processed) instructions. This is useful in the case that you have, say (using ppc assembly because I know it better):

    add r3, r3, r4
    ori r5, r3, r5
    xor r3, r3, r3

    (which puts r3+r4|r5 in r5, and 0 in r3; again, this is just an example, and kinda silly). here, r3 is used six times. For the first instruction, it is read in one context, and then written in another (writing always creates a new context). The ori then uses the r3 in the second context, and the xor uses it in the second context and makes a third. So, using tN as temporary (or rename) register N, this is the same as

    add t0, r3, r4
    ori r5, t0, r5
    xor t1, t0, t0

    The same could be done for the other registers, of course. The advantage of this is that, because the registers are used consecutively less often, scheduling is easier.

    If you're interested in more details, check out (google) Tomasulo's algorithm.

    Summary: Renaming is cool. Everyone does it. But it doesn't help you emulate more registers, particularly.

  • by nuggetman ( 242645 ) on Tuesday October 12, 2004 @05:52PM (#10507933) Homepage
    This could work out. The reasons why I would NOT get a mac are that they are slower and cost MUCH more than equivalent PCs, but more importantly, can't run Windows

    This point has been debated over and over, so I'll mention the $799 eMacs (educ discount) and $949 ibooks (also discount) and $1199 iMacs and move on.

    But now that Mac OS X is available on the PC (and is fast), perhaps I can use a fast, cheap PC to run OS X.

    One solution, PearPC, is unbearably slow for more than checking website compatibility in OS X. The other, Cheerios (yes I know), may or may not exist and may or may not work, and may or may not just be a $50 version of PearPC.

    Macs only have a chance vs. PCs because they have very efficient architecture. Apple doesn't have nearly enough money to compete with Intel or AMD, so they use a more efficient architecture.

    Why would Apple compete with AMD or Intel? Apple makes computers, and IBM makes the G5s, and Motorola makes the G4. Intel and AMD do not sell computers.

    Stop talking about of your ass.
  • Re:Good news.. (Score:3, Informative)

    by Graymalkin ( 13732 ) * on Tuesday October 12, 2004 @05:55PM (#10507965)
    The way SSE2 and AltiVec work on a core level is entirely different. On the Mac VecLib is a framework that gives abstracted access to AltiVec functionality so I don't need to write up inline assembly for commonly used mathematical routines. If I took a lot of time to make my program work with the VecLib framework in OSX I'd have to spend even more time making sure the x86 port didn't get futzed up trying to run on the register starved SSE2 unit on someone's P4. The VecLib frame work would have to dance around the P4's 8 vector registers in really grotesque ways to get the same functionality as is provided by the AltiVec unit on G4 and G5 processors.

    There's plenty of other frameworks that are heavily tied to AltiVec now. While it would be possible to gut them and get them working fine with SSE2 it would be a huge undertaking. It's taken years to get the good AltiVec support the exists right now, it would take several more to get an x86 port up to snuff.
  • by renoX ( 11677 ) on Tuesday October 12, 2004 @06:01PM (#10508025)
    > the registers you write to in machine code are virtual, and are mapped on to a larger hidden register file in realtime by the CPU

    Yes, register renaming help but as the compiler don't see those hidden register, it may have to spill some value into the cache to free a register because it needs one and here the register renaming can't help you..

    I think that the x86-64 good performance is partly because of this: going from 8 GPR to 16 is a big win, especially on x86 *ahem* less than orthogonal architecture).
    The difference between 16 and 32 GPRs is much less interesting..
  • PearPC HD files (Score:4, Informative)

    by TravisWatkins ( 746905 ) on Tuesday October 12, 2004 @06:21PM (#10508248) Homepage
    Either this is PearPC with a fancy GUI or they 'borrowed' some code from PearPC. On the video of their installer you can see macosx_3gb.rar being copied. The HD files for PearPC have to be a specific size so only a select few work.

    Also, no one has made a foolproof HD creator that works 100% so obviously CherryOS couldn't steal that. That's why their profile setup only allows 3GB or 6GB HDs. That's what is available for PearPC.
  • EULAs (Score:4, Informative)

    by Garabito ( 720521 ) on Tuesday October 12, 2004 @06:38PM (#10508432)
    The real legality of such statements is only known after it has been tried in a court of law, which AFAIK it has not.

    Maybe it hasn't been tried for Apple software, but at least one EULA was declared enforceable [] in an U.S. court. Sad, isn't it?

  • Re:So, you're asking (Score:3, Informative)

    by mabinogi ( 74033 ) on Tuesday October 12, 2004 @06:39PM (#10508447) Homepage
    Rhapsody is alive and well - it's called OS X now.

    Rhapsody wasn't the name for the x86 port, it was the name for the next generation Mach + NeXT Step based MacOS, which is what became OSX.

    It doesn't suprise me that they had x86 builds early on, but I could hardly say they "Made an x86 OS"
  • by PipsqueakOnAP133 ( 761720 ) on Tuesday October 12, 2004 @07:26PM (#10508816)
    Yeah, but because you can't address these extra registers, they're useless for emulation purposes. All this does is let you have more inflight instructions (google for Tomasulo)

    And as a side note, the G3/G4/G5 PPCs probably have those as well, since they're not a x86 specific thing. I know that the 604 does, and it's a generation 2 PPC.
  • by Autumnmist ( 80543 ) on Tuesday October 12, 2004 @08:09PM (#10509174)
    From what I hear, CherryOS *IS* PearPC (which is GPL). Sounds like they're trying to rip people off with an impossible dream.

    CherryOS = PearPC? []
  • by catwh0re ( 540371 ) on Tuesday October 12, 2004 @08:11PM (#10509197)
    Generally I will doubt any claim that suggests they can run PPC code on x86 hardware at any considerable speed, such as 80%, or even 50% for that matter. The PPC chipset has more general purpose registers than x86, how they map around the instructions to fit on an x86 chipset is usually inadequet and some kind of register emulation must take place. Taking any register functionality off chip is a method of emulation, that works, however it's incredibly slow, by comparison to native speeds. This is why it's trivial to get good speeds out of x86 code on PPC chipsets through emulation, and why the reverse is usually a marketing scam.
  • Re:Finally... (Score:4, Informative)

    by damiam ( 409504 ) on Tuesday October 12, 2004 @08:12PM (#10509206)
    The G4 and especially the G5 have a helluva lot more registers than the P4. Whatever the relative benchmarks may be, it's extremely difficult to emulate a PowerPC on x86 at decent speeds, because there aren't enough registers.
  • by cmholm ( 69081 ) <<cmholm> <at> <>> on Tuesday October 12, 2004 @08:23PM (#10509284) Homepage Journal
    MXS, the company behind CherryOS, Maui X-Stream, and the web hosting service, isn't showing me a brick-and-mortar so far. Hitting the Google caches, I find 'contact' pages that alternately list a Wailuku and a Lahaina office, both of which were copyright 2003. I tried the Wailuku address during lunch. The office was empty, sign removed from door, and no listing on the building directory. I may try the Lahaina address for fun over the weekend.
  • by revjonnylove ( 444471 ) <revjonnylove@yPL ... minus physicist> on Tuesday October 12, 2004 @08:35PM (#10509384)

    Now that the seemingly Pacific sized wave of traffic has rolled over Hawaii based, some more information can be gleamed from it's now visible pages. Their press release [] contact is stated as Jim Kartes.

    Jim also happens to be the admin and tech contact for A quick Googling of his name [] brings up several links, including the website for MauiGiclee, a Maui based printing company [] which lists one Jim Kartes as it's president. How many Jim Kartes can their be in Hawaii? lists only 1 []. Finding info online is fun.

    Further Googling and whois searches show that Jim has a hand in many things Maui.

    Lets list a few of em: [] (Paradise Television Network Inc) [] (Video Steaming Tech) [] (Print Production) [] (Emulation Software)
    I'm sure the list goes on. Jim's a busy man, you see.

    Predictably, all these websites sport the same type of Java Applet video found on Seems like VX30 (aka MXS Inc.) has been busy supplying Java based video steaming tech to a lot of Jim's other businesses.

    At any rate, these businesses (excluding, by nature of this thread, the cherry in question) seem to have been operating for some time, the oldest site being registered in 1996. They also seem quite legitimate in their desire to provide services and products, bothering to list themselves with superpages, register 1-800 numbers, etc. These are not signs of scam artists looking to make a quick get-away, so that possibility can be put to rest.

    The following options still remain:

    1. CerryOs is a ripoff of PearPC (though the company has reportedly denied these accusations by phone)

    2. The product is real and unique, though the performance promises are exagerated.

    3. This is legit and we should all stop wasting time with such nonsense : )

    I hope it's the latter.

  • by Krach42 ( 227798 ) on Tuesday October 12, 2004 @08:43PM (#10509454) Homepage Journal
    Maybe you should go and get some experience or at least some knowledge before you start talking about something you know nothing about. If _ANYONE_ says this to my post, I'm going to hurt them. And hurt them hard. First of all, it's not relatively all that hard to mimic the AltiVec instructions using SSE/SSE2/SSE3. Secondly, the majority of operations done by many AltiVec instructions use a limited number of AltiVec registers. The ones that do use the significant number of registers are things like matrix operations, complex mathematics, and bioinformatics type things. Next up, AltiVec is very well behaved. It does not perform any accesses on anything but 16-bit alligned addresses. If you access some value (x4)+y it will access the 128-bit vector starting at address (x4). This means that one can avoid all the horrible nastiness of emulating cross-page accesses. They're all well defined. Also, AltiVec produces little side-effects (the only exception that the non-load AltiVec instructions effect is if the MSR_VEC bit is not set) and no condition code changes (except for the vcmp* instructions) All of this adds up to that there are numerous benifits to implementing the AltiVec instructions in SSE, or even in just normal "scalar" code. This is because you can make the individual instructions much more efficient than those that require the overhead of the regular engine. Overall, it's hard to really say that there is much of a benifit. There are a lot of things that fall into the "overhead" section, of just matching the instructions to x86 code that does the exact same thing. But there's also a lot of benifit that comes from it. PearPC is making good progress in this regard. :P
  • by applextrent ( 821630 ) on Tuesday October 12, 2004 @09:11PM (#10509644) Homepage
    For more on this software, and issue, you can visit my site CherryOS: Interview With Creator, Plus Screenshots []
  • Re:Not likely (Score:3, Informative)

    by Bastian ( 66383 ) on Tuesday October 12, 2004 @09:49PM (#10509866)
    OS X is heavily optimized with processor-specific functions, be it AltiVec optimizations for G4s or just a lot of black magic to make the Altivec-oriented code still run well on a G3. The next version is going to include features that offload a lot of processing to the graphics card. I'm sure they put little to no effort into making sure that any of their code above the Darwin will run properly on a little-endian machine.

    That's potentially a whole lot of rewriting (and potentially creating a need to mantain yet another code branch for various portions of the OS) in order to get an OS that is still going to only work on a very small portion of the PC hardware out there. And I'm not talking "you won't be able to burn DVDs" not working, I'm talking "the OS won't run, period, because Core Image doesn't support your graphics card."

    Which means that they will have a target market consisting of people like you who are willing to buy the one and only one OS X Approved PC. Of course, to make that available as something other than a homebuild, Apple will have to make it themselves. Which will probably make it end up costing not much less than any other Apple computer because it will end up being a solid magnesium pyramid with no visible apertures or seams or something like that because that's what Apple does.

    At which point Apple has gone through a ridiculous wad of cash in order to make your Mac work less smoothly than other Macs. But at least it cost you $100 less.

    Methinks Apple would be much wiser to spend that money on continuing to improve the value of their PPC hardware. Maybe that way they can save you $150 on a better computer, instead.
  • by Anonymous Coward on Tuesday October 12, 2004 @09:58PM (#10509938)
    I was the one who posted previously... the special deals section quickly sells out of the $529 deals. Although two weeks ago they had them for $499 with no modem.

    Look every week - and for a whole week look twice a day - you'll find good deals like $529 shipped!

  • by King_TJ ( 85913 ) on Wednesday October 13, 2004 @12:29AM (#10510857) Journal
    Yeah.... I was suspicious of the same thing. At best, I figured this was sort of a "fork" of the PearPC project. Maybe they added some of their own code to handle some G4 specific functions and bundled it up with a cleaner installer/setup program. But I bet it's still just PearPC at the core.

    Their screenshots I saw this morning on their web site were only depicting OS X's main desktop and finder screens. Never once did they show it running a single app! (That was the deal with PearPC too, wasn't it? At first, people could run OS X itself, view the finder, and the prefs panes - but that was about all it could do without crashing.)

    Now, it looks like they're claiming people are "trying to hack the site" and so on, and they only have some video movie available to download/watch. I was getting horribly slow connections to them, but the first 50% or so of the video I watched only showed the program being installed on an XP box. (Big whoop! It has an installer program that can actually copy files over to the PC.)
  • Re:So, you're asking (Score:5, Informative)

    by steeviant ( 677315 ) on Wednesday October 13, 2004 @12:47AM (#10510932)
    Nah, you're thinking of something else. There have been numerous aborted attempts at creating a next generation Mac OS under a variety of strange code names like Pink, Taligent and Copland.

    Rhapsody was the name of the OS [strategy] developed under the leadership of Gil Amelio, it was heavily based on OpenStep (moreso than OS X), hence it's cross platform capabilities. Apple also had a version of the Rhapsody frameworks that ran in NT, which they inherited from NeXT. At that stage, the name for Cocoa was YellowBox, and the Classic environment was called BlueBox IIRC. There was no equivalent to the Carbon frameworks in those early days, which was the subject of much debate.

    Steve Jobs became Interim CEO after Amelio's departure in 1997 and killed the cross platform versions of Rhapsody along with the Mac 'clone' industry. About a year later Apple announced the name change from Rhapsody to Mac OS X. They released Mac OS X Server in 1999, followed a year later by the almost unrecognisable OS X Public Beta.

    Check out these screenshots, which (in order from top to bottom) show the gradual progression from NeXTstep's multi-column Browser to Mac OS X 10.3's Finder*.

    NeXTstep []
    Rhapsody []
    Mac OS X server 1.x []
    Panther []

    *yes, I skipped the aqua Finder.
  • Screenshots (Score:2, Informative)

    by icekillis ( 777986 ) on Wednesday October 13, 2004 @12:54AM (#10510962)
    Here are some screenshots: G [] G [] G [] G [] . I read they were taken by gene O'neil??? (not sure who he is)
  • Re:So, you're asking (Score:3, Informative)

    by grahamlee ( 522375 ) <iamleeg@gmail . c om> on Wednesday October 13, 2004 @05:55PM (#10518027) Homepage Journal
    OPENSTEP-based...NeXTSTEP was an earlier OS that had a very different object library, BSD base etc.
  • by steeviant ( 677315 ) on Thursday October 14, 2004 @03:03AM (#10521434)
    Are you blind? Maybe you didn't read my post? I said that it uses QE to SPEED IT UP, and if QE wasn't there, it would DECREASE PERFORMANCE.

    Actually, that's not what you said. You said:

    "Mac OS X uses Quartz Extreme to render all the windows in 3d with shadows and fancy coloring. No graphics card = horrid windowing performance."

    Implying that without Quartz Extreme windowing performance would be unbearable, that shadows wouldn't work, and that the colours would somehow be affected. All bullshit.

    This is an emulator we're talking about, even if it doesn't support Quartz Extreme it can still achieve high performance.

    Mac on Linux doesn't support Quartz Extreme yet performs admirably. Though PearPC's graphics speed is not very impressive, it's hardly the limiting factor there either.

    I'd contend that the lack of/support for QE has approximately nothing whatsoever to do with performance in an emulator (as anyone whose used a PPC emulator/VM can attest), and that your previous post appeared to say that without QE support the emulator would not be able to render shadows or draw colours correctly. This, as you are obviously aware, is blatantly false.

    Which is why I called you a troll.

    Had you made the assertions you made in this post, I would have supported some of what you say, but I think without some kind of native graphics card translator, QE would be worthless anyway, and in fact would almost certainly be slower.

    As you may already be aware, native graphics card support is not just a matter of 1:1 mapping between the PC and Mac graphics card, because of fundamental architectural differences between x86 and PPC. There would need to be some interception and modification of QE's graphics instructions into the correct form for the PC graphics chipset, which could easily negate any speed benefit.

    In short, you're correct about it being slower, but the Chicken Little-esque, sky-is-falling way you went about stating it in your previous post made it sound like the emulator would be useless simply because it didn't support QE, which is far from truthful.
  • by Anonymous Coward on Friday October 15, 2004 @10:08PM (#10542145)
    Yup, and they managed to get a phat amount of personal information from people that filled the form to be updated of a trial version.

"I don't believe in sweeping social change being manifested by one person, unless he has an atomic weapon." -- Howard Chaykin