Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 


Forgot your password?
Close
typodupeerror
×
AMD

X86-64 Simulator - now available (Linux only) 99

Posted by HeUnique from the this-is-good dept.
Well, as the title says - the AMD X86-64 simulator is shipping (for Linux only - for now). You can go here and read the details. It is called SimNow!. Unfortunately, the kernel 64 bit port is not done yet, but hey, you got a nice simulated machine to play with! It's available as RPM for RedHat 6.2 and SuSE 6.4. NOTE: - you'll need PLENTY of RAM to operate this simulator! (384MB RAM is minimum) and disk space - 4 GB! So think before you download this beast!
This discussion has been archived. No new comments can be posted.

X86-64 Simulator - now available (Linux only) More

X86-64 Simulator - now available (Linux only)

Comments Filter:
  • I guess I won't be installing it on that 486 after all :)

  • Didn't Transmeta take care of that one?

    The REAL jabber has the /. user id: 13196

  • If only they had a 32-bit x86 simulator that ran on 64-bit machines, Micorsoft could finally port Windows to 64-bit platforms!

    Wouldn't they need a 16-bit emulator first?

  • Believe me, although it is impractical speed wise, I dearly wish Linux had a VMWare like product that attempted to EMULATE the microprocessor rather than attempting to run it in native mode. Apart from speed, there are cool things you can do with an emulated system, even if it is slow.


    http://www.bochs.com/ [bochs.com] - x86 CPU emulator
  • AMD has not gotten into the multi-processor arena however. Are their 64bit CPUs going to be capable of running in a multi-processor configuration? Isn't this also the way AMD has been able to keep up with Intel on the desktop? They have not been putting their research energy towards the server end like Intel. Considering this, it wouldn't be a good idea for MS to move onto AMD's x86-64 platform while ditching IA-64. Does anyone know if AMD's x86-64 will be able to be put into a multi-processor system?
  • A LOT of windows relies on the fact that a pointer and a DWORD are the same size..
  • In theory I believe NT separates the architecture specific stuff into HAL (Hardware Abstraction Layer?) and should be easier to port. (I don't use NT, I never have, this is just from what I have read).

    But the real trick is stuff that creeps in, especially sign extending and byte ordering. It's real easy to assume chars are always signed or unsigned, what MAXINT is, and byte order. No matter how hard you try, these assumptions creep into the code. With a monster like NT 2000, any port is going to be a real nightmare. Either you review every line of code with regards to sign extending and byte ordering, or you swat bugs as they show up. Neither is foolproof. Both are a real pain in the posterior.

    Linux has been thru this phase, so it gets easier with additional platforms.

    --

  • HP has an IA-64 emulator for Linux (Score:2)

    by Anonymous Coward
    Just the other day, I was browsing developer.intel.com, and found that HP was giving away an IA-64 development kit (including simulator), also for Linux. The URL is: http://developer.intel.com/design/ia-64/linux.htm [intel.com]. Also, SGI is giving away their Pro64 Linux tool suite, apparently GCC & friends with SGI enahancements for IA-64: http://oss.sgi.com/projects/Pro64/ [sgi.com]. Now I'm no IA-64 enthusiast, but the thought of extending x86 junk to 64-bits is just... chilling.

  • Re:Interesting approach. (Score:4)

    by OlympicSponsor ( 236309 ) on Friday October 06, 2000 @07:11AM (#726250)
    "Just because they don't port to every architecture under the sun, doesn't mean that MS can't do it."

    Take it from someone who knows: If you haven't actually done the port, you haven't caught all the issues.
    --
    An abstained vote is a vote for Bush and Gore.
  • I think that was more of an issue of it wasn't worth their bother, than it was much more difficult.

    You no longer need an alpha for a pretty damned powerful workstation.

  • NOTE: - you'll need PLENTY of RAM to operate this simulator! (384MB RAM is minimum) and disk space - 4 GB! So think before you download this beast!


    Finally, I'll be able to balance my cheque book.


    --
    Chief Frog Inspector
  • HAH HAH, no seriously, I just upgraded to WINme (got it from a friend) and it runs pretty nice, and the fixed that one thing so that when you install it, unlike win98, it doens;t rewrite the MBR
  • Here's an amusing disclaimer from one of AMD's news releases [amd.com] regarding the x86-64 chip. Of course this is just to protect themselves from lawsuits, but is pretty funny anyways:

    Forward looking statements in this document include the risks that developers may not support the x86-64 technology and design tools for the technology in a timely manner or at all; that AMD will not successfully implement the technology in its products on a timely basis; and that AMD may not effectively penetrate the enterprise market.



  • Ultimately Intel obviously dont want Windoze ported to anything else since that's a threat to their dominance.

    And Microsoft probably dont want to develop windows concurrently on 10 different architectures (think of the cost!) and so it suits them fine to stick with x86.

    AMD however have shaken things up a little since their sledgehammer will keep MS on the x86 platform and leave intel somewhere else...
  • Ummm 64 bit solaris runs on the ultrasparc 2 and 3 cpus. It's not going to do very much on x86 at all sadly.
  • You heard something, but it is wrong. Yes, there is (or rather was) a port of NT to Alpha. But it was not, repeat *not*, 64 bit. It did not use the Alpha platform to its full potential, so it was kinda useless. You are correct that DEC ported it to Alpha, not Microsoft, and, when Compaq decided that they will no longer port Microsoft's software for free, Microsoft cut all support for NT on Alpha and made it sound like it was Compaq's fault. Incidentally, same happened with other NT ports. e.g. SGI ported NT to MIPS.

    Now that Microsoft cut all ports, there is only one platform Windows 2000 runs on - x86. And it doesn't look like this will change any time soon.

    ___
  • Ah, somebody is reading too much into this. RPM works on several other distros too, is AMD in bed with all of them too? (The slut!) I just think they released RPM for two reasons
    A) RPM is the most common format for "mainstream" Linux.
    B) RedHat is the most common "mainstream" distro.
    They have no intention (and nobody really expects them to) to support every bloody distro out there. If you are 'leet enough to use a different distro, then you can figure out how to have alien convert it.

  • I am no fan of Intel in particular, but I think their aproach to the problem (just start over) has much better long term potential then "lets play nice and use old stuff."

    And if propreity kernels and software eats GNU/Linux's dust, so much the better. We'll all be running smoothly on 64-bit while Microserfs are still trying to stop the rush of bluescreens :)


    ------

  • Yes, the intent of a simulator such as this is to help developers write/port code for the AMD 64-bit "sledgehammer" platform. I believe Intel has had a Merced (Itanium) simulator out for quite some time.

    Jeff
  • can't answer any of your questions, but thats one cool ass subject.
  • All the windows people would become confused.
  • I like more bits and I cannot lie!
    You other brothers can't deny!
  • [begin injection of real-world...]

    Its because the people who actually pay for software, not download it off crack sites, want to be able to use the new software on their existing x86 machines, and because the people who buy new machines actually still want to use Word Perfect 5.0 for DOS on their new Athalon.

    Don't forget that on any computer that has a single user, most of the CPU time is spent either idle or doing GUI work. The 64-bit arch is needed to overcome the I/O bottleneck, like the I/O to main memory, rather than to do anything for the MIPS ratings.
  • Well, sometimes you maybe want to have more than the 2gb of virtual memory that you get from 32 bit addressing minus OS taxes. I'm told my customers often want that.
  • *ahem*...I suppose the reason that this is only for x86 is because it is a X86-64 emulator. It emulates a 64-bit platform on x86 machines. Therefore, solaris, PPC, et. al., can reasonably be left out on this round.


    Andy
    "1 2 3 4 5 - unbelievable, that is my luggage combination!"
  • Uhm...you must have missed when AMD licensed the x86 core from Intel. When AMD started making the K5 (or was it K6?) Intel sued, saying that AMD was stealing their microcode. AMD wound up "winning", and was allowed to use the microcode. They've been since playing mostly catch-up with Intel in terms of things like MMX, but extending the chipset with stuff like 3D instruction set. This while mess then got turned into the Athlon.

    BTW, how is AMD not compatible? Because they don't support the complete instruction set of the PIII?
  • My Atari Jaguar was doing 64 bit years before any of these fancy chips!
  • You hit the nail right on the head -- Sledgehammer is a no pain-little gain upgrade for most people, where the "feature" is that you can run Windows 9x forever and have nice optimised "64-bit" video drivers and other bits here and there to improve your Quake numbers.

    Meanwhile your nice "64-bit" chip is still cranking 16-bit code part of the time. AFAIK, Microsoft has not promised a native Sledgehammer port of Windows, nor has anyone promised application support. Meaning that "64-bit" here is more of a marketing feature (much like MMX and 3DNow) than anything else. Which makes sense because by-in-large AMDs channel seems to be consumer/home boxes.

    Intel is taking a different marketing tact -- targetting Merced at people who need both 64-bit and some i386 compatiblity and are willing to pay thorugh the nose for it. Merced users will have the 4GB Ought To Be Enough For Anyone problem solved. The problem is that there have been many better 64-bit solutions out there for years, for those who really needed 64-bit, and meanwhile the 32-bit chips are scaling quite well.
  • Reminds me of running Information in Primos and writing Basic programs which ran in Information.

    Why is this thing so slow?

    Still, I'd rather have the real thing. 64 bit int ray tracing should absolutely smoke on this thing.


    --
    Chief Frog Inspector
  • What you want is already available in open source: bochs [bochs.com]
  • "and while te original Athlon platform was unstable I betcha a PIII in an CC820 motherboard with SDRAM is FAR FAR worse..."

    First off, use ispell! Correct spelling is your best friend, incorrect spelling, your worst enemy.

    Any good CPU in a bad chipset will perform bad. How about a PIII in a 440BX motherboard with SDRAM? That's decent, and there are still thousands of servers out there using that. How many servers are there using the Athlon? 30? 35?

  • This is *really* an odd message, Archvile. I would expect a little better.

    All of AMD's chips have been VERY good for compatibility; they've put a lot of energy into that. There have been flaws, a couple of the older ones being TERRIBLE, but the rarity of those flaws is almost incredible. Intel itself does worse.

    And the AMD processor architecture is different from the P6 and 7 architecture. You're talking out of your hat to claim otherwise. The design is different.

    -Billy
  • NT USED to be architecture neutral. Recently though, they gave up the Alpha port (which was really the only port still production quality) and is now more or less x86 only.

  • I finally have something to do with that old 25MHz 486 in the closet!
    Please allow 2-3 weeks for the kernel to compile.

    Actual build time 2.2.16 kernel on an 80386DX/40 8MB RAM: 4hrs

  • "I'm completely fed up with this [windowing system]. I don't see why we need [windowing systems] with 8 diff modes of operation so they can be [source] compatible with software from the 80's. Why can't people make a move to [modern] arc in the PC industry w/o having to have everything be compatible. Can't the major companies who only distribute binaries afford to dist a [BeOS app_server version], a [Berlin version], and another new arch? XFree86, Xig, MetroX, should just give up on this crappy [windowing system] argh."

    I just answered your own question ;)
  • Isn't the way the story was written a bit hypocritical? It doesn't appear that the code is free as in speech, only free as in beer, and it has been released for RedHat 6.2 and SuSe 6.4 only? While I love Linux as much as any other nerd here, they're denying a large portion of the market to test out this machine, including other free OSes (*BSD). Had they released this code for Windows only, I'm sure they're would be plenty of commentary about AMD trying to suck up to M$. Free software should mean free for *everybody*.

    You are more than the sum of what you consume.

  • Wow. (Score:4)

    by Black Parrot ( 19622 ) on Friday October 06, 2000 @06:25AM (#726278)
    If only they had a 32-bit x86 simulator that ran on 64-bit machines, Micorsoft could finally port Windows to 64-bit platforms!

    --
  • > Wouldn't they need a 16-bit emulator first?

    I think that's built in if you buy Win NT/2K. For Win9x, they bundle in a free one called "DOS".

    --
  • Run a x86 64bit emulator on a 32 bit system, to install a 32 bit operating system ;)
  • You're mistaken. First of all, this is not a hardware issue, it's a software issue. There's a bug in the VIA AGP driver that causes memory corruption. This is neither a hardware problem or an AMD problem. Microsoft incorrectly associated these problems with problems with the processor when in fact it's a seperate (but somewhat related) issue. It's a bug, not a hardware incompatibility. And when they mean AGP video they are talking about using AGP's extra features, like sidebanding and AGP texturing. This only affects 3D programs, and programs have to be specifically written with AGP in mind to use most of these features. As far as the GeForce issues, it's a problem with motherboard compatibility, not the processor. A lot of the first Athlon boards didn't provide enough power to the AGP slot, which GeForce cards need. If you turned the I/O voltage up to 3.4V from 3.3 (or replacedd your motherboard) then it worked fine. And I know a lot of dealers who would recommend AMD over Intel. I'm not a dealer, but I recommend AMD, and I use AMD processors exclusively. Case in point: somebody on my floor has an SMP dual P3-700. Besides the processor (I have a single Athlon 750) and motherboard (I have a KA7-100, he has some BX board, don't remember which) our computers are pretty much identical. His crashes all the time, and he's shocked at how fast and reliable my Athlon rig is. (Yes, we are both using Win2K, so he does get SMP.) And mine cost hundreds less. Need I say more?

  • Emulating (Score:5)

    by Ashran ( 107876 ) on Friday October 06, 2000 @06:26AM (#726282) Homepage
    I gotta try this one out.
    run a x86-64 emu, install linux on it
    run VMWare, install windows
    and start a gameboy emulator ;)
    this way i get the exact gameboy speed =)
  • "I got those points by posting what I truly believed, though I only gained points when it was the majority opinion. You should try the same."

    What should I try? Going along with every other misguided soul in the tech world? No, I'm going to stand up for what I believe is right. Haven't you ever read the works of Henry David Thoreau, especially "On the Duty of Civil Disobediance"? Consider this: "...any man more right than his neighbors, constitutes a majority of one already." You've already quietly admitted that you're wrong by trying to focus on something else. Face it, you've failed. I am already a majority of one.

  • Nothing neater than a 64-bit environment emulator running on 32-bit boxes.... Hurrah for progress!!! =)

  • WOW!!! (Score:1)

    by Anonymous Coward
    NOW I CAN BUILD A 64-BIT vi AND EDIT SOME REALLY REALLY BIG FILES!!!
  • I think it's rather funny that you linked to that issue. I have that exact problem. On a Pentium 3/450 with 440BX motherboard. The problem is that a lot of older motherboards were made with off-spec AGP power, which causes GeForce-series cards to corrupt sections of memory and crash.
    ------------------
    A picture is worth 500 DWORDS.

  • Maxis? (Score:4)

    by Tirisfal ( 140294 ) on Friday October 06, 2000 @06:27AM (#726287)
    SimNow!.... sounds like something else by Maxis
  • I think to get regular gameboy speed after all of that you'd need to start out with a Beowulf cluster :) Some Gameboy emulators right now don't even give you full speed on today's mid-range machines.
  • I'm slowly starting to get addicted to diffrent hardware platforms, i've recently gotten some sparcs off of e-bay, And i hope this new amd chips comes off nicely,

  • Re:Interesting approach. (Score:3)

    by DickBreath ( 207180 ) on Friday October 06, 2000 @06:53AM (#726290) Homepage
    Isn't Win NT 5 (ooops, I mean Win 2000) architecture neutral like Linux? Just because they don't port to every architecture under the sun, doesn't mean that MS can't do it.

    What I think this means is that the x86-64 architecture could extend the life of the Win 9X lineage (Win 95, Win 98, Win 98 SE, Win ME). This would enable people who aren't successfully coerced by MS's attempts to migrate to Win 2000 to continue paying for upgrades to the Win 9X codebase -- but recompiled for better hardware. The IA-64 couldn't support the descendants of Win 9x. I suppose this means that x86-64 could breathe another hundred years of life into the Win 9x codebase.

    For Win 2000, I think MS could support IA-64 as well as x86-64.

    If MS were to throw a lot of support behind x86-64 and little support behind IA-64, then it would seem to be in Intel's best interest to really get behind Linux in order for there to be some universe of software that they can sell IA-64 hardware into.
  • If you just want "practice 64 bit code" then get an old alpha and install NetBSD or Linux or whatever on it.

    A used noname/AXPpci33 board with some ram is much cheaper than an extra of 256MB RAM for your computer (and probably faster as well).

    Or: Get IA64sim. Yes it's for Intel's 64bitter (Merced/Itanium or whatever it's its name now) but it's open source, doesn't run on linux only and doesn't need that much RAM.
  • Perhaps you could run Solaris on this? I know that on Sparc64 hardware, a 64-bit kernel is used, but I'm not sure about Solaris x86. Even if they didn't it wouldn't take long to port the kernel over to x86-64.

    Just a thought.
  • First you tell us it's for Linux-only, and I understand from previous slashdot articles that it's my job as a rabid Linux zealot to support all software released for Linux only. But then at the end, you say I should think before downloading it. Make up your mind! Am I supposed to be a zealot, or am I supposed to think?
  • Judging on the same basis intel should be sued by Zilog because of the same argument. As a fact... the only thing AMD and Cyrix haven't been able to perfectly emulate have been the many design flaws (FDIV bug, instability at more than 1 GHz, an MTH that doesn't work, etc. etc.). As a matter of fact, AMD and Cyrix designs are derivative compatible works that the only thing they share in common with Intel is the Instruction Set (And not even so because AMD uses a RISC engine that decodes and translates the IA-32 instructions into RISC86), and trust me it took them a really long time to get the designs to work (K6 while not a bad processor wasn't exactly a champion).

    Oh, yeah another thing... AMD doesn't make north+south bridges as Intel do, and while te original Athlon platform was unstable I betcha a PIII in an CC820 motherboard with SDRAM is FAR FAR worse...

    ZoeSch
  • If you're dealing with small virtual drive images (which I imagine you are, having stored them on CD) you can do exactly the same Save..Muck..Restore.. process using Linux on the native hardware and forgetting the virtual machine..

    Create a 640 meg HD partition as /dev/hda1.
    Install Linux on hda2+.
    Install Windows9x on hda1, reboot Linux using a boot disk, and restore LILO.
    dd if=/dev/hda1 of=/diskimages/windows98SE.image bs=1M count=640.
    Wipe hda1 with random data or zero fill it before installing the next OS.
    To reinstall a image, simply boot Linux, dd if=/diskimages/windows98SE.image of=/dev/hda1 bs=1M count=640.

    Presto! Native speed, native hardware, and you can use all of the great Unixy tools on the images to do snapshot diffs, binary diffs, etc..

  • You show me a 120 wpm typist, and I'll show you someone who will BADLY type ahead of anything in Win(blows) with a mythical 6GHz Ultra-Athalon.

    Somethings work better in DOS mode.
  • Ok, so I've 2 256MB registered PC-100 SDRAMs in a system that has a couple of extra UW 9GB drives around... but it's only a 300a@450... hmmm, guess I'll wait for a while.

    Damn.
  • I know, I was being sarcastic. It has two semi-32bit processors, one graphics and one number cruncher. Based on that logic, my SMP system is a 96bit system.
  • i don't think win2k is architecturally neutral....

    else why would it not support anything but x86 at release when nt 4 supports at least mips, x86, alpha and ppc, and up till the later stages in the win2k release process it ran on alphas.

    just my $.03 cents (tax and all)
  • Don't forget that the other reason that x86 will not die is that the dominant OS (Windoze) is not ported to other architectures, not to mention any of the Windoze programs that people want to use. If MS ported their stuff to other platforms, and Windoze software companies did as well, then there would be a mechanism for the obsolescence of x86. But as long as MS is tied to x86, it will live as long as MS does. And that looks like a depressingly long time to me.
  • Nice sig ! first a Queensryche reference yesterday, now Dream Theater (The words to "Ytse Jam" choke me up !):p It caught my eye because I live down the street from Jordan Rudess' house, and have been a fan since '90 or so when they were Majesty (Turn that word backwards !) Nothing really intelligent to say


    - Save The Whales ,Collect the whole set !
  • the thought of extending x86 junk to 64-bits is just... chilling.

    I thought so too at first, but read the documentation. In 64-bit mode, many of the x86's wackiest features and instructions go away. No more segments/selectors, and you get 8 new general purpose registers to play with (in addition to EAX and friends). It actually "looks and feels" more like 64-bit MIPS than x86. I think GCC oughta be able to crank out some nice code with 8 more registers to play with :)
  • i would think that NT lost a lot of it's platform independence when MS moved the video driver stuff into the kernel...
  • since some troll begged for it:

    "Can you imagine a Beowulf cluster of these simulators?"

  • Maybe because it costs real bucks, and nobody wants it.

    A few years back, MS was considering dropping the PowerPC port of future versions of NT unless Apple wanted to contribute to it's development. Apple didn't and thus no PPC for Win 2000.

    Is there even a possible market for Win 2000 on a MIPS? Alpha?

    The answers to these questions probalby explain the lack of other architectures. Not that they can't port it.
  • But they have done the port. NT ran on four different architectures.

    It seems that they're just not maintaining it.
  • While I entirely love AMD, here are a couple of thoughts that entered my mind while reading this thread:

    Intel spends the energy on developing and implementing the instruction, and essentially produces an 'alpha'(no pun intended) processor...enter, FDIV 'n friends.

    AMD maintains compatability, with an instruction set already defined...They never had to worry about developing their own instruction set(aside from RISC86, IIRC) because they knew that where Intel went, Microsoft(read: most of the victimized world.) would follow. (And don't get _anyone_ started on the connections, good or bad, between MS and Intel.)

    Mike
  • Isn't Win NT 5 (ooops, I mean Win 2000) architecture neutral like Linux? Just because they don't port to every architecture under the sun, doesn't mean that MS can't do it.

    Yes, it is. If you know where to look, the first few betas of win2000 are available on alpha as well as x86, they just dropped it half way through, because it wasn't worth their time. But you can bet your ass the hooks are still in there, for just this situation. They may be dicks in redmond, but they're definitrely not idiots.

    Gfunk007
  • Requires 384MB of RAM and 4 gigs of HDD space?

    Hey... you gus SURE that isn't the next windows release?

    ~Reave
  • *ahem*...I suppose the reason that this is only for x86 is because it is a X86-64 emulator. It emulates a 64-bit platform on x86 machines. Therefore, solaris, PPC, et. al., can reasonably be left out on this round.

    No, it doesn't emulate a 64-bit platform on x86 machines. It emulates an X86-64, a 64-bit x86 platform. There's no reason it couldn't run on some other machine. Nintendo/SNES/Gameboy emulators work just fine on x86 PCs, despite the lack of 6502/65816/Z80s in 'em.

    And FYI, here [umd.edu] is an public domain IA64 simulator, which you could compile on a SPARC, Alpha, PPC, whatever... it only simulates the CPU though, not a full machine.

  • Mmm, troll :)

    I've never seen an AMD web server able to sustain a ping under 100ms.

    yerfable ~> ping -c5 dahan
    PING dahan.metonymy.com (10.1.1.66): 48 data bytes
    64 bytes from 10.1.1.66: icmp_seq=0 ttl=255 time=0.007 ms
    64 bytes from 10.1.1.66: icmp_seq=1 ttl=255 time=0.008 ms
    64 bytes from 10.1.1.66: icmp_seq=2 ttl=255 time=0.008 ms
    64 bytes from 10.1.1.66: icmp_seq=3 ttl=255 time=0.008 ms
    64 bytes from 10.1.1.66: icmp_seq=4 ttl=255 time=0.008 ms

    ----dahan.metonymy.com PING Statistics----
    5 packets transmitted, 5 packets received, 0.0% packet loss
    round-trip min/avg/max/stddev = 0.007/0.008/0.008/0.000 ms

    dahan ~> dmesg | head -4
    NetBSD 1.5_ALPHA2 (SPIFF) #340: Tue Sep 26 19:33:23 CDT 2000
    khym@dahan.metonymy.com:/usr/src.local/sys/arch/i3 86/compile/SPIFF
    cpu0: AMD K6-2 (586-class)

    There ya go, 0.008 milliseconds, i.e., 8 microseconds. Way under 100ms.

  • I was under the impression that MIPS was the original development platform for NT? I think it was ported to i386 afterwards.

    I'll never forgive MS and the Q for killing the alpha port, but it did give me a good excuse to put FreeBSD on my AlphaStation :-)
  • Most of the people cannot even try it! it requires 384MB RAM (minimum!), 4GB Disk space and a good 700+Mhz Processor

    So think, before you download this beast..
  • I'd always thought that that was a deliberate feature to make it impossible for people to dual boot... I wouldn't put it past MS.

  • pissing in the cheerios (Score:3)

    by gtx ( 204552 ) on Friday October 06, 2000 @06:33AM (#726316) Homepage
    while this is a neat idea and all, does this really have any true value of application? I guess what I'm asking is, what are most people going to get out of this, besides a huge kernel and a sudden hunger for RAM? Is this supposed to be for people that want to code for X86-64 to get a head start? If so, is it really wise to base coding for a 64 bit chip on what happens in a 64 bit simulator on a 32 bit processor?
  • AMD have confirmed that 'sledgehammer' (x86-64) will already have two cpu's on the die - so yes, you could definitely say it supports multiprocessing.

    Tomshardware Article [tomshardware.com]

    You'd be tempted to think that multiprocessing versions of the chip will come out which are a single die with 4, 8, 16 etc cpu's on it.

    Quake seven here we come.

  • How about after the Windows, instead of a Gameboy, we do a C=64 :-)

    8-bit STILL rules!
  • Where can I send my campaign contributions?
  • Most x86 platforms (read: all) are based on a 32 bit architecture. That means the processor can work with at most 32 bits at one time. A 64 bit architecture means the processor can work wih 64 bits at a time. This emulator is a program that runs on a 32 bit machine which simulates a 64 bit processor. Cool stuff, eh? :)
  • Actually, that's the reason behind simulators, Intel did that with the IA-64 but only to some parties. AMD actually wants to do three things, a) test the architecture and develop compile optimizations, b)gauge the response of the community towards the processor, c)Piss off Intel :)

    Imagine if it works :)

    ZoeSch

  • Why can't x86 just DIE (Score:1)

    by Anonymous Coward
    I'm completly fed up with this arch. I don't see why we need processors with 8 diff modes of operation so that they can be compatable with software from the 80's. Why can't people make a move to 64-bit arch in the PC industry w/o having to have everything be compatable? Can't the major companies who only distribute binaries afford to dist a Mac Version, x86 Version, and another new arch? Intel, AMD, etc, should just give up on this crappy proccessor family. argh.
  • I finally have something to do with that old 25MHz 486 in the closet!
    Please allow 2-3 weeks for the kernel to compile.

  • Re:To be 3l33t try this (Score:5)

    by Mindwarp ( 15738 ) on Friday October 06, 2000 @06:58AM (#726324) Homepage Journal
    Most of the people cannot even try it! it requires 384MB RAM (minimum!), 4GB Disk space and a good 700+Mhz Processor

    Wow, what a coincidence! The exact same minimum requirements as Windows ME!

    --

  • Speed is not the primary goal of emulation (Score:4)

    by DickBreath ( 207180 ) on Friday October 06, 2000 @07:01AM (#726325) Homepage
    I use emulators for certian kinds of work, and believe me, speed is about the furthest thing from my mind.

    The easiest to understand example is Virtual PC for Power Macintosh. It emulates the hardware of a PC. That is, the microprocessor, an emulated S3 Trio 64 video card, an emulated DEC 21041 ethernet controller on IRQ 11, etc. This is all emulated. Macintosh has no idea of what an IRQ is. It's all just a software simulation. Like simulating the Enterprise Bridge on the holodeck of a Romulan ship.

    Execution speed is not great. But that is not my primary interest. The things you can do with, or to, an emulated system are amazing. Since the entire hard disk of the emulated PC is just a file on the Macintosh, I can make a before and after snapshot of the entire hard drive (including unused sectors) from before and after the installation of some program. Since the Mac is capable of mounting the file as a virtual drive (sorta loopback filesystem) I can then analyze exactly what files were altered on the hard drive. I can make before and after registry dumps. I keep CD-ROM's with pre-burned images of virgin installs of all my un-favorite OS's, such as Win 95, Win 95 SP1, Win 95 OSR2, Win 98, Win 98 SE, Win NT 4, etc., etc., etc. I can restore a virgin install of a Virtual PC hard drive from a CD-ROM in under 4 minutes. I can tinker with the DLL's and registry with impunity. After all, I'm not screwing up a real machine that will take hours to reinstall -- it's just a simulation that can be reset back to virgin status in under 4 minutes.

    Believe me, although it is impractical speed wise, I dearly wish Linux had a VMWare like product that attempted to EMULATE the microprocessor rather than attempting to run it in native mode. Apart from speed, there are cool things you can do with an emulated system, even if it is slow.
  • The IA64 prototype Dell showed off last week was running an early Windows port. Couldn't tell you how well it works though.
  • Yes. I have high hopes.

    Virtual PC is just very mature. It has other nice things like virtual floppies in disk files -- that are perfectly compatible with Apple Disk Copy utility.
  • I use linux, on Alpha, and x86. That being said:
    Windows NT (which was originally written for Alphas if what people have told me is correct) did run on a 64-bit platform -Alpha. It ran well, and only crashed once in about a year of running it. It was not written by Microsoft however, and when DEC was bought by Compaq, they saw lots and lots of money going to M$ so that DEC could write their code. Compaq stoped that. I have heard that the betas of NT 5 (win 2000) were more stable than the final release on x86. NT 4 on alpha is still stable, but after the lack of support, it likely has some security holes that haven't been patched, and costs lots of money.
  • lol does that count punctuation, more or less a sustained 120 wpm typist should be ablet o look at anything and touch type at 120wpm.
  • Really man what is your trip?

    You run around spreading stuff like this and calling me an idiot because I said "Comparing two sites over the internet wherein you have next to no data about the hardware or site configurations and saying that servlets are slower because the site x loads slower than site z" is totally unscientific and you tripped out asking me if I knew what traceroute is.

    Now you are sitting here trying to spread more crap.

    I mean, really I can understand some things but here you go again, "an amd webserver cant sustain a ping under 100 ms".

    That is such garbage and you and I both know it.

    lets see there is no such thing as an AMD web server. There are webservers with AMD chips in them but your just spouting random bullshit and I dont mind saying so.

    Lets see were you pinging over the net?, were you pinging over a LAN? Was the server you were pinging loaded, what kind of NIC were you using, were you using the most recent drivers, what operating system were you using, what kind of memory was in the machines, what kind of motherboards were in use, what is the connection speeds on both ends.

    Lets come back to reality and speak like adults and present each other with useful facts not just silly facts like "slashdot loads in 2.5 seconds winamp's site loads in 10, therefore perl is faster than servlets"

    Jeremy
  • an issue with an Intel spec (AGP)?
    Wonder why that happened.

    There isn't a single dealer I know that would recommend AMD over Intel. I've never seen an AMD web server able to sustain a ping under 100ms. All this points to AMD's contempt for pure performance, as opposed to their lust for a slapdash way of getting profit

    Umm...Where have you been for the last year?
    Athlons have a 10-20% (older spec) preformance increase over an equivelent P3. (Of course, Alphas still cream them 150%(21164s) to about 500%(21264s) that is with gcc the optimised x86 compiler that works on most other things)
    The price of a 1GHz Athlon $550, 1GHz P3 >$1000
    I have dealers who recommend AMD-K6s over P2s, Durons or Athlons over P3s. Not the kind that stick their heads in the sand, and believe Intel's marketing. Plus, Intel's chips are not reliable, the 1.13 GHz recall and I have a friend who got a new (retail-boxed) P3-600 which burned out.

  • Windows Me is a 32 bit simulator. Get your byte alignment straight!
  • You both forgot how to spell "Athlon".

    Do you also pronounce it Ath-a-lon? Eesh. I have a teacher who calls it "Anth-a-lon"

    *shudder*
  • x86 (8086) was a crap processor when it came out. And now its been extended to hell, compare the speed of a G4 and x86 and you see a 400Mhz G4 taking on top-end intel chips.
    Its time to move on and start with a clean implementation, and IA-64 is it.
    I think IA-64 will run IA-32 as a emulation mode while stuff gets ported anyhow.
    Lets not continue with 60/70s based designs just because we can keep using Microsoft Windows. How long will Microsoft continue to hold back progress in the computer industry?
  • thats what I always thought too, but I guess they finally got enough complaints and threatening letters to take it out
  • You're still left carrying all the baggage of 20 years of the x86 line, however. Regardless of if those wacky instructions vanish or not in 64-bit mode, I have a feeling it's not going to be as simple as it would seem from the surface.

    And as for GCC generating better code? I've seen GCC create code consisting of movl %eax, %eax, so I'm not going to hold my breath too much.

  • >Believe me, although it is impractical speed >wise, I dearly wish Linux had a VMWare like >product that attempted to EMULATE the >microprocessor rather than attempting to run it >in native mode. Apart from speed, there are cool >things you can do with an emulated system, even >if it is slow. There is one, check out http://simics.com It also emulates Sparc and Alpha.
  • Are you serious?

    A while ago. I started reading Slashdot when it was pretty young, probably one or two months old. I would have a lower user number but I didn't bother getting a user ID for the first week or two that they were available. But alot of people could say that as well.

    I was searching for the Linux Web Watcher (couldn't remember the name of the site) when I stumbled upon Slashdot.

  • Re:pissing in the cheerios (Score:3)

    by DFDumont ( 19326 ) on Friday October 06, 2000 @06:39AM (#726339)
    Yes, most emulators are used to help the developers work with the target system in a controlled, software-only environment. Nearly all, read anyone with a brain, embedded systems are developed this way, especially since the hardware used is typically one-shot ASIC derived.

    As far as is it valid, that depends on how well the simulator works. There's no fundamental reason it could not be absolutely, 100% correct; just slower.

  • Interesting approach. (Score:5)

    by IGnatius T Foobar ( 4328 ) on Friday October 06, 2000 @06:39AM (#726340) Homepage Journal
    AMD's choice to add 64-bit instructions to the existing 32-bit set opens up some possibilities. While we in the Linux universe don't particularly consider an architecture jump to be something radical (since we have a hardware-agnostic compiler and kernel, as well as source code to most of the important apps), consider what this could mean to Microsoft. All that legacy 32-bit code, all that downloadable Active-X content that is so architecture dependent ... it'll probably run on AMD's x86-64 architecture without having to jump the chip into an emulation mode. They can just retarget their compiler for x86-64, do a build of the existing system, and optimize in a few places. All existing apps run natively because the 32-bit instructions aren't running emulated.

    Put it all together and you've got a good reason for Microsoft to suddenly declare that IA-64 is a train wreck waiting to happen, and x86-64 is going to be the 64-bit architecture supported by Windows. AMD instantly becomes the king of the CPU market while Intel spends another five years retooling.
    --

Slashdot Top Deals

Top Ten Things Overheard At The ANSI C Draft Committee Meetings: (1) Gee, I wish we hadn't backed down on 'noalias'.

Close