Imagination Tech. buys MIPS

Discussion in 'CPUs and Overclocking' started by Raghu, Nov 6, 2012.

  1. Raghu

    Raghu Senior member

    Joined:
    Aug 28, 2004
    Messages:
    396
    Likes Received:
    1
    #1 Raghu, Nov 6, 2012
    Last edited: Nov 6, 2012
  2. Loading...

    Similar Threads - Imagination Tech buys Forum Date
    Intel licensing AMD's GPU tech? CPUs and Overclocking May 16, 2017
    Imagination details Warrior P5600 CPU core, promises high speed in a small package CPUs and Overclocking Oct 14, 2013
    AMD announces HSA Foundation: AMD, ARM, Texas Instruments, Imagination and MediaTek CPUs and Overclocking Jun 12, 2012
    I am I imagining things about ivy bridge? CPUs and Overclocking Apr 23, 2012
    Imagine a PowerPC processor for computer. CPUs and Overclocking Nov 17, 2003

  3. meloz

    meloz Senior member

    Joined:
    Jul 8, 2008
    Messages:
    320
    Likes Received:
    0
    They are trying to survive and stay relevant.
     
  4. MisterMac

    MisterMac Senior member

    Joined:
    Sep 16, 2011
    Messages:
    779
    Likes Received:
    0
    Is this like a defensive move by ARM ?

    And why is Intel greenlighting it?


    Trying to make the mobile world into a x86 thing where intel gets control of 90%?.


    I don't get this - some wise heads explain.
     
  5. Dark Shroud

    Dark Shroud Golden Member

    Joined:
    Mar 26, 2010
    Messages:
    1,576
    Likes Received:
    0
    Intel is green lighting it because they're getting access to a portion of the patents.

    I don't like the little end comment of the article that ignores AMD.

    I'll also give VIA an honorable mention because they do sell a lot of chips in Asia for micro-portables running XP.

    Through Sea Micro we'll probably see AMD making ARM SOCs using arm cores wrapped in AMD IP paired with AMD IGPs.
     
  6. MisterMac

    MisterMac Senior member

    Joined:
    Sep 16, 2011
    Messages:
    779
    Likes Received:
    0
    Intel "needs" patents in uARCH\ISA areas?
    (And i mean needs, not "nice to have" or useful - i mean NEED).
     
  7. Idontcare

    Idontcare Elite Member

    Joined:
    Oct 10, 1999
    Messages:
    21,130
    Likes Received:
    8
    Of course they do. It is surprising, I mean I get the reasoning why you would be surprised that this is the reality for a company as large as Intel, but it is true nevertheless.

    For example, a proof if you will, consider that the only reason Cyrix managed to get their x86 license was because they sued Intel over patent violations of Intel's P5/P6 chips which used power saving tricks that Cyrix pioneered. The x86 license for Cyrix was the "out of court settlement" that made all that go away.

    Again, totally surprising that Intel would actually allow itself to be beholden to others like that but the reality is that they did/do.

    AMD64 is probably the most noteworthy example. Intel couldn't make/sell x86 processors that were 64bit compatible were it not for their licensing AMD's patents on making it happen. We were all supposed to be using 64bit Itaniums right now had things worked out to plan for Intel.
     
  8. MisterMac

    MisterMac Senior member

    Joined:
    Sep 16, 2011
    Messages:
    779
    Likes Received:
    0
    I think those are old examples.

    And the AMD64 one is not really comparable - Intel had a different plan.
    The market just rejected it. Even intel can't get away with everything :p

    I was more curious as to what - MIPS offers the ARM'ish LLC group with intel in it?

    It's not like one just designs a ISA\uARCH\CPU - and become a threat to any established market.

    MIPS has promise - but it's still far off.
    The cheap price? - might as well be change for intel at this point.
     
  9. Idontcare

    Idontcare Elite Member

    Joined:
    Oct 10, 1999
    Messages:
    21,130
    Likes Received:
    8
    US Patents are good for 15 yrs (use to be 17), old examples are still impacting the market today.

    IP is what keeps Intel from fielding a GPU or an iGPU that rivals AMD or Nvidia.

    Who knows what MIPs had in the field of low-power IP that Intel wants to retain access to or take advantage of going forward? I don't have a clue what Mips may have, but I know to never assume they don't have something because in this industry even the guys in their garages have something...that is why the patent trolls have an industry to work in the first place.
     
  10. Raghu

    Raghu Senior member

    Joined:
    Aug 28, 2004
    Messages:
    396
    Likes Received:
    1
    Intel has a CPU cross-licensing agreement with AMD. If AMD makes 128bit, Intel gets the license too. If Intel makes SSE3, AMD gets to make it too.

    Intel got GPU licenses from NVIDIA in exchange for chipset license earlier. After recent court battle Intel is paying money to NVIDIA for GPU license, since NVIDIA cant make chipsets anymore.

    I think NVIDIA has GPU cross-licensing agreement with ATI/AMD.

    Wonder what Apple gains out of this?
     
  11. Ferzerp

    Ferzerp Diamond Member

    Joined:
    Oct 12, 1999
    Messages:
    6,205
    Likes Received:
    14
    Not even Intel was able to kill x86...
     
  12. ViRGE

    ViRGE Elite Member, Moderator Emeritus

    Joined:
    Oct 9, 1999
    Messages:
    31,164
    Likes Received:
    29
    The short answer is that this is a defensive move on behalf of the CPU industry as a whole. MIPS has been in poor shape for quite some time now - ARM has pretty much booted them out from high-margin consumer products using SoCs - and as such MIPS was putting themselves up for sale. MIPS has always had a polite relationship with with the rest of the industry, but there was significant concern that a patent troll could buy MIPS solely to sue Intel, AMD, IBM, and the ARM Consortium.

    This isn't a new risk for the the CPU industry, and as a result they have already been buying up patents for quite some time through the Allied Security Trust. So they are using Allied Security Trust in conjunction with IMG to scoop up MIPS before a patent troll can get them.

    Allied Security Trust gets the bulk of the MIPS patents (for around $350mil), and will give all of its members access to those patents. Meanwhile IMG gets the remaining patents and MIPS' operations (for $60mil). Someone needed to take MIPS operations because of their outstanding customer commitments, and that someone will be IMG. However it's not clear whether IMG intends to continue with the MIPS architecture, or if they'll roll the engineers and technology into their other products. The latter seems more likely, but IMG isn't saying a whole lot.
     
  13. Vesku

    Vesku Diamond Member

    Joined:
    Aug 25, 2005
    Messages:
    3,739
    Likes Received:
    24
    China's state sponsored CPUs are based on MIPS, probably was a factor in Imagination deciding they were worth buying. Companies are still optimistic about the growth potential of the Chinese market.
     
  14. Roland00Address

    Roland00Address Golden Member

    Joined:
    Dec 17, 2008
    Messages:
    1,797
    Likes Received:
    8
    On another note, can ARM make a high performing cpu with all the patents that amd and intel have without significantly raising the prices on said cpus to pay for all the licensing fees? ARM gets royalties of 6.7 cents for each cpu sold (not this is the mean average of all cpus they design from low performing to high performing.)

    Couldn't in the end intel raise the cpu prices of all the cpus in the industry by charging higher royalties to ARM (as well as other companies.) Either you buy an intel cpu and intel makes money directly, or you buy a non intel cpu and intel makes money indirectly?
     
  15. ShintaiDK

    ShintaiDK Lifer

    Joined:
    Apr 22, 2012
    Messages:
    20,395
    Likes Received:
    128
    ARM doesnt scale up well. Also why the improvements are so marginal.

    But like with AMD, ARM suffers the same problem. R&D budgets. Patents are all secondary.
     
  16. Exophase

    Exophase Diamond Member

    Joined:
    Apr 19, 2012
    Messages:
    4,440
    Likes Received:
    5
    This comment is as bad as the claim that x86 doesn't scale down well. Can you give a single technical reason why the architecture inhibits scaling up? I doubt it. High performance CPUs are just not a good market for an ARM licensee to push into.

    What you probably meant to say is that current ARM CPUs can't scale up as high as high performance x86 CPUs. Neither can low power x86 CPUs like Atom or Bobcat. And those high performance x86 CPUs can't scale low as well.
     
  17. Idontcare

    Idontcare Elite Member

    Joined:
    Oct 10, 1999
    Messages:
    21,130
    Likes Received:
    8
    Don't make the mistake of conflating microarchitecture with the ISA.

    x86 is an ISA, not a microarchitecture. Microarchitectures scale in power and performance, not the ISA. ISA merely defines capability.

    ARM gets used as both a moniker for the ISA as well as the microarchitecture which only clouds the discussion when comparing to x86 (the ISA).

    ARM, the ISA, of course can scale in performance. But right now ARM's microarchitectures are not engineered to scale up in performance to the heights of which x86, the ISA, ends up being scaled with its respective microarchitectures (Ivy Bridge, Piledriver, etc).

    One thing holding back the microarchitecture is the memory subsystem (bandwidth) and the cache hierarchy. It is currently designed for low-power and low-production-cost needs, not designed at all to scale up to high performance.

    Custom cores and platforms from companies like Calxeda could change that.
     
  18. Ferzerp

    Ferzerp Diamond Member

    Joined:
    Oct 12, 1999
    Messages:
    6,205
    Likes Received:
    14
    Of course it could change that. The question is, does anyone have the spare funds and time required to do so? I get the feeling that some people have the impression that fast microarchitectures grow on trees and don't understand that it has taken decades of development and billions of dollars to get something like IB, or Power7, or the like (as they all build on their predecessors). You don't just go out and build something that can compete with these in performance by saying "well guys, I guess it's time to make the fastest CPU ever. Let's go license ARM IP and do it in 4 years".
     
  19. Exophase

    Exophase Diamond Member

    Joined:
    Apr 19, 2012
    Messages:
    4,440
    Likes Received:
    5
    How am I the one who is conflating microarchitecture and architecture? ShintaiDK was doing that and that's why I posted what I did.

    From this context ARM could have only referred to architecture. Because he was responding to someone who asked "can ARM make a high performing CPU."

    Please tell me exactly what it is that made you think I needed this response, I'm genuinely curious as to what could have made you think I didn't know any of this.
     
  20. Idontcare

    Idontcare Elite Member

    Joined:
    Oct 10, 1999
    Messages:
    21,130
    Likes Received:
    8

    Maybe I misunderstood your motivation for posting the question as you did, but I took your invocation of the term "architecture" to imply microarchitecture and not that of the instruction set as one generally refers to the Instruction Set Architecture as simply the ISA (reserving the "architecture" moniker for use as shorthand reference to the microarchitecture).

    I can see now that if I substitute the acronym ISA for the word architecture in your post that your post does indeed capture the essence of the situation sans any and all conflationary concerns I had when I originally read it.
     
  21. Exophase

    Exophase Diamond Member

    Joined:
    Apr 19, 2012
    Messages:
    4,440
    Likes Received:
    5
    Okay, I could see why that would be confusing. I remember having that problem years ago. I'm used to people generally using "architecture" as shorthand for ISA and microarchitecture for a particular hardware implementation/family of implementations. I guess I took that convention for granted >_>

    As for what others are saying re design capability..

    I agree with the claims that Intel (and to a lesser extent AMD) have a substantial design advantage due to already having targeted high performance, and having less R&D money. But the question is just "can ARM make a high performance processor" (or maybe it's extended to "can someone make a high performance ARM processor"?), not if they can provide fierce competition with Intel, AMD, or IBM in single threaded performance. And until someone actually tries to allow a single ARM core to use a ~100W TDP envelope you can't say what the result will be.

    You probably wouldn't see such an attempt in server markets because it's a losing battle to go for high single threaded perf at the expense of perf/W for high throughput wimpier core loads. ie, it's better to go for a niche and win some ground than go for mainstream where you can't compete at all. And you get better throughput perf/W results designing a CPU that can't clock past 2.5GHz than you do designing one that can clock past 4GHz. AMD is especially feeling this disadvantage.

    Where we could theoretically see single threaded performance pushed higher for a custom design is in a console. Since they're on closed ecosystems the competitive pressure is different and the console companies still have an advantage in licensing or even owning processor IP, where ARM is potentially as good a choice as any. And power isn't constrained like in tablets and phones.

    But I'm not really placing bets on that happening either.

    Nonetheless, I wouldn't underestimate the ability of teams who have created decent low power designs to transition somewhat decently to high performance, just like Intel has transitioned somewhat decently to low power. Because a lot of big design advantages benefit both, and it's not like ARM cores today aren't employing several modern technologies, they're not just 1999 era processors. Although it's less obvious than it is with Intel's tick-tock progression ARM's designs have followed a reasonable evolution, with Cortex-A9 being a successor to ARM11 and Cortex-A15 a successor to Cortex-A8, and Cortex-A57 is clearly an improvement over Cortex-A57. No one's starting from scratch here.
     
  22. Cerb

    Cerb Elite Member

    Joined:
    Aug 26, 2000
    Messages:
    17,409
    Likes Received:
    0
    Also, MIPS has a real legacy, and good processors (not fast ones, these days, but that could change). I guarantee you that you own several--but maybe dozens--of them.

    Imagination has started to get into CPUs, too. I wonder if they saw this as an opportunity to make use of a long-supported ISA, so that they wouldn't have to prove the software support side of Meta all over the place?
     
  23. Exophase

    Exophase Diamond Member

    Joined:
    Apr 19, 2012
    Messages:
    4,440
    Likes Received:
    5
    In what devices?

    http://www.mips.com/everywhere/mips-based-products/

    You might have one in your TV or cable box and you might have one in your cable modem but I'm not sure it's a strong guarantee. They're not the ubiquitous microcontrollers that you're sure to find in a wifi module or somewhere in your car.

    IMG has had Meta for a really long time. I don't know why people are only now picking up on it. I guess they made a press announcement about some Meta product recently?

    Moving to MIPS over their established ISA wouldn't win them very much at all. MIPS is an easy starting point but the advantage it gives you in legacy/toolchain is close to nothing. They're probably more interested in the engineers and patents than the ISA.
     
  24. ElFenix

    ElFenix Super Moderator and Elite Member
    Super Moderator

    Joined:
    Mar 20, 2000
    Messages:
    96,493
    Likes Received:
    314
    MIPS powered SGI as well as the N64, amongst other things.


    as for ARM, there's no reason they couldn't develop a cpu uarch that attempts to compete with intel. it's probably not worth it for them because intel has such a huge head start in every area necessary to compete in that market from a uarch standpoint. not to mention intel's 2 year manufacturing lead on the rest of the industry.
     
  25. beginner99

    beginner99 Diamond Member

    Joined:
    Jun 2, 2009
    Messages:
    3,171
    Likes Received:
    94
    exactly and besides that if I had those billion $ I would invest ti otherwise. I mean if the can pull it of and get something competitive out (at least in performance/watt) you will still only make like 1-2k $ per CPU. A CPU which will probably run applications worth of 100k or more. (unless it's used for LAMP ;) ).
     
  26. Idontcare

    Idontcare Elite Member

    Joined:
    Oct 10, 1999
    Messages:
    21,130
    Likes Received:
    8
    Looking at the billions of dollars AMD invested in an attempt to accomplish the same, as well as the failed efforts of past design houses (cyrix/centaur/trasnmeta), and looking at ARM's revenue/cashflow...it is not clear to me how ARM (the company) could ever be expected to find the finances needed to resource a team of thousands of engineers as needed to begin the 4-5yr uphill climb.

    To me, the money aspects of the challenge is one very valid reason why ARM couldn't develop such a cpu uarch.

    Now someone with silly deep pockets that was managed by someone who is either a fool or a visionary could do that if they had a design license, say Apple, Nvidia, Qualcomm, or Samsung.

    But those companies are successful because they aren't wasteful like AMD in chasing the pot of gold at the end of the rainbow, so I don't see any of them deciding to pursue such a resource-intensive development path.