( ESNUG 547 Item 4 ) -------------------------------------------- [02/20/15]

Subject: Samsung 14nm, ICC II, QCOM, Rich Goldman, Tabula, Lip-Bu, SPICE

FROZEN HELL: Here's the latest news & rumors I've heard about EDA and the
fabs this winter.  Feel free to comment/correct/scold anything!

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

  - You know the buzz about how the Qualcomm Snapdragon 810 lost it's socket
    inside the Samsung Galaxy S cellphone to Samsung's own Exynos CPU?

      "... our Snapdragon 810 processor will not be in the upcoming design
       cycle of a large customer's flagship device."

             - Qualcomm 1Q15 earnings release (01/18/15)

    For perspective, here's Strategy Analytics top global smartphone units
    shipped in Q2 2014:

             Samsung:  :######################### 74.5 M units (25.2%)
               Apple:  :############ 35.2 M units (11.9%)
              Huawei:  :####### 20.1 M units (6.8%)
              Lenovo:  :##### 15.8 M units (5.4%)
              Xiaomi:  :##### 15.1 M units (5.1%)
                  LG:  :##### 14.5 M inits (4.9%)

    That socket-loss news caused Qualcomm stock to drop from $72.61 down to
    $62.46 a share -- a painful 14% decline in just 4 days.
   
    But the untold gossip most people do NOT know is that this turned out to
    be a behind-the-scenes battle between Aart's favorite IC Compiler II vs.
    Aart's hated archrival Atoptech Aprisa/Apogee.
      
    That is, sources say Qualcomm's overheated Snapdragon 810 was laid out
    in TSMC 20 nm using IC Compiler II and the fully functional Exynos CPU
    was laid out in Samsung 14 nm using Atoptech.  Add the fact that Wall
    Street analyst Warren Lau of Maybank said in his stock report that the
    Exynos CPU has already yielded at 60% to 70% in 14Q4 and should hit
    80% yield in 1Q15 makes it: "Congrats to ATOP!... and sorry, Aart."

      Monday - On Friday I mistyped "TSMC 16 nm" above when I meant to
      type "TSMC 20 nm".  I've corrected it now to "TSMC 20 nm".  - John

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

  - FUNNY SIDE NOTE:

      "... Cadence and TSMC are collaborating on the 10 nm FinFET
       process, and Cadence solutions are ready to support 10 nm early..."

         - Cadence press release (09/26/14)

      "... With needed tool enhancements in place to meet 10-nm FinFET
       process requirements, customers can now use Synopsys tools..."

         - Synopsys press release (09/25/14)

      "... partnered to make the Mentor Olympus-SoC place and route
       system ready for TSMC's 10 nm FinFET requirements..."

         - Mentor press release (09/26/14)

      "... collaboration between ATopTech and TSMC ensures that joint
       customers realize the full potential of FinFET at 16 nm..."

         - Atoptech press release (09/30/14)

    Did you catch that?  Everyone other than Atoptech got TSMC endorsements
    for 10 nm!!!  ATOP only got 16 nm and that was 4 to 5 days after the
    Big Three.  My spies tell me this is probably Hsinchu's retribution for
    ATOP PNR helping make that Exynos CPU successful in Samsung 14 nm.

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

  - USER QUESTION:
   
      "We do mobile designs.  Currently an ARM house.  Atom uses
       too much power.  Kills batteries.  Do you know anyone using
       the Warrior I6400 in a real device?  What are they seeing?"

           - Anon Designer on Imgtec's new 64-bit MIPS core

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

    Mentor has announced that it's acquired Flexras, a French start-up that
    sold Wagas, a multi-FPGA partitioning tool for large chip prototyping.
    It's rumored to have sold for under $4 million, but since Flexras was a
    small 5 or 6 man company that's not bad.
      
    The untold rest-of-the-story is since MENT bought the Veridae Certus
    debugger from Tektronix (for a firesale price) -- and that MENT is also
    OEMing the Auspy multi-FPGA paritioner -- it's clear that Wally's FPGA
    guys are crafting a new FPGA tool possibly to launch at the next DAC.

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

  - Rumor is ChipEstimate.com is discontinuing its InCyte chip estimator
    that was both its namesake plus designer draw to the IP aggregator.
      
    Which probably is a result of the original founder of ChipEstimate.com,
    Adam Traidman, having left Cadence to found BreadWallet, a bitcoin
    wallet app on Apple's iTunes store.  "The money's in bitcoins, baby!"

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

  - A little bird tells me that during the latest SNPS Annual Sales Kickoff
    back in November 2014 -- which is the beginning of the Synopsys fiscal
    year -- Anton Domic announced internally that Synopsys was discontinuing
    all future sales of FineSim Pro.  And instead all the customers are to
    be pushed over to their Made-In-Synopsys CustomSim-XA.
   
    With this move, Anton effectively says to his fast SPICE customers:
   
      "Remember all those sweetheart deals you customers got from LAVA
       for FineSim Pro right before we absorbed them?  You won't get
       those low prices any more.  Prices are going up.  Time to pay."

    Which is dangerous because it creates a "Why-Don't-We-Look-At-All-The-
    Tools-In-This-Niche?" event where the customer does a complete re-eval
    of all SPICE vendors.  Thus creating an opportunity for BDA, Cadence,
    Silvaco, Tanner, and ProPlus to get a foot in the door.  Is that wise?
      
    The other weird side effect is that Aart is sending a signal inside his
    own Synopsys R&D telling the old Magma digital FineSim Pro SPICE R&D
    guys that they're "no longer needed, thank you".  Ouch!

    That's the rumors on fast SPICE.  For Full SPICE, the word is that Aart
    is discontinuing his (Nassda) HSIM and pushing customers over to HSPICE
    and/or FineSim.  Supporting 9 flavors of SPICE is expensive.  Aart wants
    to consolidate SPICEs to cut costs.

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

  - Rumor is that over the past 4 months Cadence has been reorganizing
    their Custom IC Business Unit such that many of their key Virtuoso
    and Spectre R&D and marketing positions are moving to Pittsburgh, PA.
      
    The reason why is that's where Tom Beckley and his old loyalist core
    of Neolinear guys live.  Got that?  Cadence is moving all their key
    top analog/custom work out of Silicon Valley.  Weird.

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

  - In related bad news for the TSMC 20 nm process folks who are struggling
    to make planar fab still work:

      "We have learned that high performance discrete GPUs will never
       make it to 20 nm, simply as process has too much power leaks and
       poor yields with big cores."

           - Fuad Abazovic (Fudzilla 02/09/15)

    Which, if true, means that GPU houses like AMD and Nvidia will wait
    until 2016 for TSMC 16 nm or GlobalFoundries/Samsung at 14 nm.  Planar
    is pretty much dead.  It means those smaller TSMC customers like AMD
    and Nvidia are waiting and might jump.  And it gets worse:

      "Apple is likely to switch its chipset for its next generation
       iPhone to Samsung on 14 nm.  Qualcomm could transition to 14/16 nm
       quickly, making 20 nm node just a transitional node due to several
       challenges."

           - Warren Lau, analyst at Maybank (Barron's 01/11/15)

    Big customers like Apple and Qualcomm make up 30% to 40% of TSMC's
    business are also waiting and might jump -- which is why Warren Lau
    of Maybank downgraded TSMC shares from "Neutral" to "Sell".  And also
    why both Credit Suisse and BNP Paribas downgraded TSMC to "Neutral".

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

  - Don't look now, but there are two SNPS bigwigs suddenly now missing from
    the Synopsys.com web site:
      
    Rich Goldman, VP of Marketing and Strategic Alliances, and the former
    CEO of Synopsys Armenia -- is suddenly missing -- plus all traces of
    his biography, etc. -- even though he's been with Synopsys for an
    incredible 22 years since 1993!!!

    John Chilton, GM of Coverity, and former Sr. VP of Marketing, Strategy,
    and Corporate Development -- is suddenly missing -- plus all traces
    of his biography, etc. -- even though he's been with Synopsys for an
    whopping 20 years since 1995!!!

    It's kind of sad to put in 20 years and then get disappeared like that.

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

  - FUN HISTORY QUOTE:

      "System Verilog is, in truth, a bit of a Frankensteinian language.
       Its many new constructs on top of Verilog came from several
       different sources."

           - Tom Anderson on Cadence.com (09/15/11)

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

    EEtimes.com reports that the FPGA start-up, Tabula, is rumored to be
    going out of business in 5 weeks at the end of March.
      
    One of those rare programmable logic start-ups that occasionally pops
    up to take on the Xilinx and Altera duopoly, Tabula was a Wall Street
    darling which got $214 M in VC funding.  Intel Capital even invested!
    Their ABAX2 chip had a unique 12 layer 3-D architecture that shortened
    internal interconnects which enabled "efficient place-and-route,
    simpler timing closure, and higher resource utilization."
   
    The WSJ says Tabula and Achronix -- also an FPGA company -- was one of
    Intel Custom Foundry's early first customers of their 22 nm fab line.
    
            ----    ----    ----    ----    ----    ----   ----

   - LIP-BU'S PUZZLE: Four months ago during the CDNS 3rd quarter 2014
     earnings call, Lip-Bu Tan boasted (on record):

      "I think this large contract we referred to is a marquee global
       company.  We cannot say more than that.  But clearly it's a very
       big contract for us and clearly across product line and you know,
       clearly, digital is a very good portion of that and then also
       our verification, our IP business."

           - Lip-Bu Tan, CDNS CEO (3Q14 Earnings 10/20/14)
      
     Last time that this happened, it turned out that Lip-Bu had stolen
     LG away from SNPS.  (See ESNUG 541 #3.)  So now inquiring minds are
     super curious about Lip-bu's new not-so-quiet puzzle...

     I'd love it if anon sources told me who this "marquee company" is.
     (Hint! Hint!)

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

And again, feel free to comment/correct/scold/question anything you see here
by just emailing me directly.  And, yes, you'll be ANONYMOUS!

    - John Cooley
      DeepChip.com                               Holliston, MA

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

Related Articles:

    17 guesses on which tier one P&R customer Lip-Bu stole from Aart

Join    Index    Next->Item






   
 Sign up for the DeepChip newsletter.
Email
 Read what EDA tool users really think.














Feedback About Wiretaps ESNUGs SIGN UP! Downloads Trip Reports Advertise

"Relax. This is a discussion. Anything said here is just one engineer's opinion. Email in your dissenting letter and it'll be published, too."
This Web Site Is Modified Every 2-3 Days
Copyright 1991-2024 John Cooley.  All Rights Reserved.
| Contact John Cooley | Webmaster | Legal | Feedback Form |

   !!!     "It's not a BUG,
  /o o\  /  it's a FEATURE!"
 (  >  )
  \ - / 
  _] [_     (jcooley 1991)