[Closed due to flaky BB] Next Powermac 970 with up to 2,5 GHZ ?

11819212324

Comments

  • Reply 401 of 476
    boy_analogboy_analog Posts: 315member
    Ho hum. I'm a little bored with all the x86 talk. Let's go crazy and talk about, um, the 970.



    Given the kind of bandwidth that we're expecting with the 970, isn't it about time to dust off the Raycer rumours? Over in the Ars "Future CPU" thread our very own Programmer seems to be hinting at something along these lines.



    In short: 970s are hungry for bandwidth; standard RAM will have trouble keeping these beasts fed, especially when packaged as duals; really really fast RAM is really really expensive; so let's find a way to streamline RAM and bus usage.



    Ergo, Raycer tech.



    Comments?
  • Reply 402 of 476
    matsumatsu Posts: 6,558member
    HOLY CRAP... if I said it once I said it a million times. When you look at QE, you're looking at Raycer. There is no "chip" and there NEVER will be!
  • Reply 403 of 476
    rickagrickag Posts: 1,626member
    I'll try again



    I did notice the Broadcom chips seem to be the BCM5703S. Are these currently being used?





    BCM5703S 10/100/1000BASE-T Controller With Integrated Transceiver

    The BCM5703S 10/100/1000BASE-T Gigabit Ethernet Media Access Control and Serializer/Deserializer (SERDES) is a fully integrated interface solution for high-performance network applications. The BCM5703S is a highly integrated solution combining a triple-speed, IEEE 802.3 compliant Media Access Controller (MAC), PCI and PCI-X bus interfaces, on-chip buffer memory, and an integrated SERDES transceiver in a single device. The BCM5703S is fabricated in a low-voltage .13µ CMOS process, providing a low-power system solution. By itself, the BCM5703S provides a complete single-chip Gigabit Ethernet NIC or LOM solution.



    Support for the following 802.3 functions is featured in the MAC: VLAN tagging, layer 2 priority encoding, link aggregation, and full-duplex flow control.



    The device provides both PCI v2.2 and PCI-X v1.0 bus interfaces. The BCM5703S provides large on-chip buffer memory for stand-alone operation. Dual on-chip highperformance processors enable custom frame processing features, including TCP segmentation.



    <a href="http://www.broadcom.com/"; target="_blank">Broadcom's web page</a>



    As for any one wishing to reply, a simple "Nothing here move along" would satisfy me.



    but this processor(er um dual processor) does have a PCI-X bus interface among other things.
  • Reply 404 of 476
    @homenow@homenow Posts: 998member
    [quote]Originally posted by Whisper:

    <strong>

    YellowBox = Cocoa on Windows

    RedBox = x86/Win32 emulation in OS X

    BlueBox = Classic Mode</strong><hr></blockquote>



    I might be remembering it wrong, but I thought that Yellow Box was the Cocoa layer of Rhapsody, both PowerPC and X86. Red Box was Win32 API's on X86 versions of Rhapsody, not PowerPC. And BlueBox was Classic (OS 8-9 API's) on PowerPC only, and would not be available for X86 versions of Rhapsody. So to break it down:



    Rhapsody PowerPC includes-

    Yellow Box (OS X-Cocoa)

    Blue Box (Classic mode in OS X)



    Rhapsody X86 includes-

    Yellow Box (OS X-Cocoa)

    Red Box (Win 32)



    Apple had a hard time selling this to developers becouse they would have had to translate all their code to Object C to make it work in Yellow Box. Apple came up with the Carbon API set to address this concern, and make migration of code easier for developers. I could check on this info, but my developers release CD's of Rhapsody is at work.



    Now, what would be great is if Apple put together a developers platform like Code Worrior, but added a module that would translate other languages (C+, C++) to object C, and verrify (debug) the Object C code based on the original language. It would need to handle C, C+, C++, Object C, and Java. This would give developers an easy path to get their legacy code into Object C while they were learning the language. Then add into it a way to compile the code for Win 32 as well, for a single "core code" development (I think it would still need to be modified for each platforms API's for optimal performance). Then sell it the program for Windows, Linux, and OS X at a price point under Code Worrior, yet with as much polish and as full of a feature set as Code Worrior. The problem, this is directly competing with the premier development package on the market.
  • Reply 405 of 476
    summasumma Posts: 2member
  • Reply 406 of 476
    tjmtjm Posts: 367member
    [quote]Originally posted by summa:

    <strong>Info on IBM´s new 2 Ghz G4 processor

    <a href="http://www.futurebytes.ch/content/article/1044658801.php"; target="_blank">http://www.futurebytes.ch/content/article/1044658801.php</a>;



    </strong><hr></blockquote>



    Hmmm.... A 2.0 GHz "G4" with 64-bit address space and 64-bit integers? 4.3 GB/s bus bandwidth? Made by IBM? Gee, I wonder what this "Gekko II" chip is derived from?
  • Reply 407 of 476
    ghstmarsghstmars Posts: 140member
    Moki what do you think about this?? sandbaggin 2?
  • Reply 408 of 476
    ghstmarsghstmars Posts: 140member
    IBM G4 2.0?
  • Reply 409 of 476
    davegeedavegee Posts: 2,765member
    [quote]Originally posted by summa:

    <strong>Info on IBM´s new 2 Ghz G4 processor

    <a href="http://www.futurebytes.ch/content/article/1044658801.php"; target="_blank">http://www.futurebytes.ch/content/article/1044658801.php</a>;



    </strong><hr></blockquote>





    Wow...



    Custom RISC IBM G4 "Geckko II"

    0.13 micron IBM copper wire technology

    2.07 Ghz

    1757 Dmips(Dhrystone 3.32)

    64-bit Integer & 128-bit floating-point

    4.3GB/second peak bandwidth

    (64-bit address space, 128-bit data bus 533.33 MHz clock)

    L1: instruction 64KB, data 64KB (1 way) L2: 512KB (2 way)

    Custom ATI/Nintendo Radeon "Flipper II"

    Approx. 8MB sustainable latency : 5ns (1T-SRAM)

    Approx. 4MB sustainable latency : 5ns (1T-SRAM)

    25.6GB/second (Peak)

    12.8GB/second (Peak)

    32-bit color, 24-bit Z buffer



    ---------------------------

    Compare w/Todays Specs

    ---------------------------



    Custom IBM Power PC "Gekko"

    0.18 micron IBM copper wire technology

    485 MHz

    1125 Dmips (Dhrystone 2.1)

    32-bit Integer & 64-bit floating-point

    1.3GB/second peak bandwidth (32-bit address space, 64-bit data bus 162 MHz clock)

    L1: instruction 32KB, data 32KB (8 way) L2: 256KB (2 way)

    Custom ATI/Nintendo "Flipper"

    Approx. 2MB sustainable latency : 6.2ns (1T-SRAM)

    Approx. 1MB sustainable latency : 6.2ns (1T-SRAM)

    10.4GB/second (Peak)

    2.6GB/second (Peak)

    24-bit color, 24-bit Z buffer



    Linky: <a href="http://www.nintendo.com/systems/gcn/specifications.jsp"; target="_blank">http://www.nintendo.com/systems/gcn/specifications.jsp</a>;



    Dave
  • Reply 410 of 476
    kidredkidred Posts: 2,402member
    [quote]Originally posted by TJM:

    <strong>



    Hmmm.... A 2.0 GHz "G4" with 64-bit address space and 64-bit integers? 4.3 GB/s bus bandwidth? Made by IBM? Gee, I wonder what this "Gekko II" chip is derived from? </strong><hr></blockquote>



    It said on the spec sheet for the game cube that are confident it will dominate the next decade. Well, i guess they forgot about a little console that has a handful of games, ah, what's it called, uh, ah yes, PLAYSTATION. A 2.0ghz is fine, but the PS3 (which IBM is also making) is equal to 50 1.4ghz G4s (so I read ) as it does a teraflop. Add to that HUNDREDS of games and I don't see how the came cube is going to 'dominate' unless they mean M$'s Xbox defenition of dominate.



    Should be nice either way tho, all I care about is M$ getting left behind in a strorm of dust so I can continue my evil laughter.
  • Reply 411 of 476
    @homenow@homenow Posts: 998member
    Console life spans are pretty long in general. The Game Cube was released less than 18 months ago right? From the 64 to the Cube Nintendo took at least 4 years between models, yet this spec sheet looks to be pretty "polished" for an internal only release. I would say that it is most likely a hoax, too optamistic and too short of a production run on the Game Cube. Nintendo may be pressured to push up development speeds, as Sony is apparently doing with the PS 3, but this seams very early to see even a mock up of a product.
  • Reply 412 of 476
    [quote]Originally posted by summa:

    <strong>Info on IBM´s new 2 Ghz G4 processor

    <a href="http://www.futurebytes.ch/content/article/1044658801.php"; target="_blank">http://www.futurebytes.ch/content/article/1044658801.php</a>;



    </strong><hr></blockquote>



    Is that for real, or just some Maya designer's wet dream?
  • Reply 413 of 476
    tjmtjm Posts: 367member
    [quote]Originally posted by OverToasty:

    <strong>



    Is that for real, or just some Maya designer's wet dream?</strong><hr></blockquote>



    I've spent a few minutes researching this, and it appears to be a fake. A very elaborate and sophisticated one, however. I'm impressed with the quality. Looks like real professional work. Lots of pointless work, it seems to me, unless it is something the artist is using for a portfolio.



    I was suspicious when I saw the "blue-light DVD" drive. Sony is just coming out with those (due in April, IIRC) and they are extremely expensive at the moment. Not something for a mass-production game console for quite a while.



    [ 03-05-2003: Message edited by: TJM ]</p>
  • Reply 414 of 476
    outsideroutsider Posts: 6,008member
  • Reply 415 of 476
    summasumma Posts: 2member
    Unfortunately this must be a hoax



    The bluray drives are really expensive

    <a href="http://www.avland.co.uk/sony/bdzs77/index.htm"; target="_blank">http://www.avland.co.uk/sony/bdzs77/index.htm</a>;
  • Reply 416 of 476
    tjmtjm Posts: 367member
    What gives me pause about this, though, is the highly detailed specs. They're reasonable - not like someone just pulled them out of their butt - as though they are based on real hardware. What if Nintendo sees itself slowly fading under the onslaught of the PS2 and is looking for a way to grab the spotlight again and jump-start sales? This thing is fully compatible with the GC - no one would be out anything who had bought software.



    This "Gekko II' chip would obviously be derived from the 970. Suppose that "The Age of the 970" is a carefully planned media event involving IBM, Apple, and Nintendo sometime this summer? IBM unveils its Blades, Apple unveils its new PowerMacs, and Nintendo unveils the Nexus, all at the same time. Big splash, lots of free publicity, etc., etc.



    I have two problems with this scenario yet (I'm sure others of you have more):



    1) The Sony Bluray is indeed very expensive - ~$4000. However, looking at its specs, Sony has thrown everything but the kitchen sink into this thing. I wonder how much of the cost is due to all the "bells and whistles" and price gouging of early-adopters, vs. actual cost of the basic drive.



    The ONLY way I see this could work, then, is if Nintendo had some stripped-down, bare minimum custom drives built and were willing to take gas on the costs for a while to build their marketshare.



    2) The image still bugs me - it just doesn't look real. The something about the textures, or surface sheen, or something that tells me that this is not a photograph of a real object.



    And please, flamers, don't take any of this too seriously - I'm about 99.9% certain this is fake. I'm just doing a bit of thinking out loud.



    Edit:

    Another possibility is that the specs are for real, just the picture is fake (or a mock-up based on a verbal description). I would not be too shocked to see something with these specs from Nintendo before Christmas this year. Of course, I'd be a lot less shocked if nothing at all appeared before next Christmas.



    [ 03-05-2003: Message edited by: TJM ]</p>
  • Reply 417 of 476
    @homenow@homenow Posts: 998member
    Hoax:



    look at the last link on the specs, notice that the controller ports have a blue light on them, and in the center, just below the controller ports there is a green light. Now look in the reflection of the unit, these lights are not reflected.



    Also, I would assume that any production console would have clear or smoked plastic window covering the DVD. If this were a real product, then there would be some highlights and reflections on this plastic interfearing with the view of the DVD.



    [ 03-05-2003: Message edited by: @homenow ]</p>
  • Reply 418 of 476
    programmerprogrammer Posts: 3,467member
    [quote]Originally posted by Matsu:

    <strong>HOLY CRAP... if I said it once I said it a million times. When you look at QE, you're looking at Raycer. There is no "chip" and there NEVER will be!</strong><hr></blockquote>



    Easy Matsu, easy. No need to blow a blood vessel on this.



    What I was repeating over on Ars was simply the persistent rumour (from several independent sources now) that Apple's next chipset will include a few vector signal processors. There is no other information about them, no indication about which team designed them (although the Raycer guys might be obvious candidates), and no information about what they'd be used for. Given what is generally known about VSPs in other hardware, however, it is likely that these would be useful for 3D graphics, 2D graphics, video compression/decompression, audio processing, and potentially some scientific applications (a la Apple's vDSP library). In Apple's current machines these things would make a lot of sense since the G4s can only use 1.3 GB/sec of bandwidth but the chipset has access to closer to 2.7 GB/sec thanks to the DDR333. To put such a thing into a 970-based machine implies to me that they must be fairly fast VSPs otherwise it wouldn't be worth the effort and cost.
  • Reply 419 of 476
    matsumatsu Posts: 6,558member
    Hoax, nintendo will have to seriously think about their viability for yet another console. Sony and M$ can afford to lose a lot more per console than nintendo can. That means Nintendo has to wait longer before it can release a new console and probably still won't get the most powerful hardware for their troubles. They might make a Dreamcast, try to be first and have great games at a cheaper price, but even that didn't work for Sega. Though I tell you, the game world would look a lot different right now if the Dreamcast had shipped with a DVD playing drive.



    In any event, nintendo is faced with the choice of changing their business now with money still in their pockets, or making a valiant last stand, and then changing their business model without any money in their pockets.
  • Reply 420 of 476
    outsideroutsider Posts: 6,008member
    [quote]Originally posted by Matsu:

    <strong>Hoax, nintendo will have to seriously think about their viability for yet another console. Sony and M$ can afford to lose a lot more per console than nintendo can. That means Nintendo has to wait longer before it can release a new console and probably still won't get the most powerful hardware for their troubles. They might make a Dreamcast, try to be first and have great games at a cheaper price, but even that didn't work for Sega. Though I tell you, the game world would look a lot different right now if the Dreamcast had shipped with a DVD playing drive.



    In any event, nintendo is faced with the choice of changing their business now with money still in their pockets, or making a valiant last stand, and then changing their business model without any money in their pockets.</strong><hr></blockquote>

    Nintendo has already committed to another console after the GameCube for 2004.
Sign In or Register to comment.