Skip to navigation

PCPro-Computing in the Real World Printed from www.pcpro.co.uk

Register to receive our regular email newsletter at http://www.pcpro.co.uk/registration.

The newsletter contains links to our latest PC news, product reviews, features and how-to guides, plus special offers and competitions.

// Home / Blogs

Posted on September 28th, 2009 by Steve Cassidy

How many MIPS is Honeyball holding?

One more wafer-thin Nehalem? This picture shows PC Pro contributing editor Jon Honeyball cuddling an entire fabrication wafer of Intel’s Nehalem generation of Xeons.

Pardon the overall fidelity and quality of the shot but these opportunities don’t come up often, and it was in the middle of an auditorium of journalists, waiting their chance to fondle the wafer case and go cross-eyed counting how many processors are here: we thought it might be $200,000-worth but on the other hand we couldn’t figure out which revision or clock-rate these are.

The man from Intel messed up his own presentation by bringing this lot out, because he wanted to keep a beady eye on the thing as it passed through so many hands round the hall – though if I’m honest I had no independent way to verify that this was a wafer of Nehalems; seen one naked chip, you’ve very nearly seen them all.

But it all left me wondering which year in computing history this number of Nehalems represents? That is, how many MIPS are there in that wafer, and what point would that have equalled the total number of MIPS for the entire planet. My guess is 1957, but I’m willing to be corrected…

Posted in: Hardware, Real World Computing

Permalink

Follow any responses to this entry through the RSS 2.0 feed.

Social Bookmark this article: What is this?

You can skip to the end and leave a response. Pinging is currently not allowed.

7 Responses to “ How many MIPS is Honeyball holding? ”

  1. Jesse Marcel Says:
    September 28th, 2009 at 5:14 pm

    Early in the morning of July 8th 1947 would be my guess the total number of MIPS for the entire planet.

     
  2. Ian Says:
    September 28th, 2009 at 9:21 pm

    Nothing like a bit of late-night maths:

    Assume 300mm diameter wafer, 261mm^2 die. Wikipedia formula gives maximum dies on the wafe to be ~230, so lets take 90% ish of that – 205

    Information at this time of night is sketchy, but lets guess (again) at 48 GFlops per die (seems to be low-end for the beefiest 45nm xeons).

    So ~9.8 TFlops, or there about.

    (Sorry couldn’t find xeon scores in Mips)

     
  3. Jim Says:
    September 28th, 2009 at 9:48 pm

    I worked for Intel until 12 weeks ago and it is amazing how blaze you get about wafers, chips etc.

     
  4. Steve Cassidy Says:
    September 29th, 2009 at 9:27 am

    This seemed to me like a question for Wolfram Alpha. It snootily says that Gflops and “megainstructions per second” are not compatible units, presumably because a floating point op may require a variable number of instructions… so maybe MIPS is the wrong unit. Taking Ian’s late night maths as a starting point, I found this: http://boinc.berkeley.edu/talks/singapore_public.pdf and page 10 suggests the year 2000!

     
  5. milliganp Says:
    October 1st, 2009 at 5:09 pm

    I’d be happy to guess something in the late 70’s or early 80’s. The Vax was the machine that defined Mips (they used to be called VAX mips). In 1978 DEC were building less than 1000 VAX systems per year -and that was the most powerful minicomputer available. IBM s370’s were capable of 20-30 MIPS, but those sold in hundreds of units a year so the two biggest manufactures together were probaly shipping less than 20,000 mips per year. The wafer probably does 2+ million mips -up to 100x the total worlwide computer production in the mid 70’s.
    Moore’s law has brought us a long way!

     
  6. Steve Cassidy Says:
    October 2nd, 2009 at 12:20 pm

    I remember booting VAXes off tiny tape cartridges! this was the period when the disk controller subsystem test program played chess with you (because it was a complete PDP-11). The idea that all the world’s TFlops might produce nothing more interesting than Twitter, was a long way away back then…!

     
  7. Mr, A .Wake. Says:
    October 2nd, 2009 at 2:22 pm

    You do know that The Roswell ‘incident’ occurred on or about July 8, 1947 and was initially ‘exposed’ by Major Jesse Marcel. I think the first poster is hinting that after the recovery of alien technology the planets total MIPS was a lot larger than was publicly admitted or believed.

     

Leave a Reply

* required fields

* Will not be published

Categories

Authors

Archives

advertisement

SEARCH
SIGN UP

Your email:

Your password:

remember me

advertisement


Hitwise Top 10 Website 2008