[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference 7.286::digital

Title:The Digital way of working
Moderator:QUARK::LIONELON
Created:Fri Feb 14 1986
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:5321
Total number of notes:139771

3687.0. "Does Alpha Chip run on CDC Cyber?" by MROA::CGRIFFIN () Thu Feb 09 1995 17:50

    I am looking for information regarding whether or not our Alpha Chip
    runs on a CDC Cyber mainframe.  I work in the MRO1 Library and have 
    a Digital employee looking for this info.  If anyone knows the answer,
    please call me at DTN 297-6564. 
    
    Thanks,
    Carol Griffin
T.RTitleUserPersonal
Name
DateLines
3687.1QUARK::LIONELFree advice is worth every centThu Feb 09 1995 18:415
The question, as phrased, is meaningless.  I tried calling you to get
clarification, but got no answer.  Perhaps you can get the employee to
restate the question with more detail as to what is wanted.

					Steve
3687.2 Missed my chance...POBOX::CORSONHigher, and a bit more to the rightThu Feb 09 1995 19:056
    
    	Aaw, Steve - I wanted to ask if Alpha ran on SUN.
    
    
    
    		the Greyhawk
3687.3OTOOA::MOWBRAYThis isn't a job its an AdventureThu Feb 09 1995 19:4516
    If the question is "does CDC use ALpha's in their systems (Cyber)" then
    the answer is    NO.
    
    If the question is "does OSF/1 run on a CDC Cyber " then the answer is
    NO.
    
    In fact for alomst any question that combines CDC-Cyber and Alpha the answer
    is NO, except for things like "are Cyber's and Alpha's both computers "
    in which case a "Maybe" is perhaps OK. or "Does Alpha kick a Cyber's
    butt"? in which case the ans. is YES.
    
    re. Greyhawk's question, the answer is also NO, Sun uses a proprietary
    architecture which (although RISC) is not an open industry standard and
    consequently there is no INTEROPERABILITY between these systems.  
    
    Hope this clears things for everyone.
3687.4Mis-phrased question or not, here's an answer...ATLANT::SCHMIDTE&RT -- Embedded and RealTime EngineeringThu Feb 09 1995 22:078
  A third party *DID* build a Cyber emulator, although I don't
  remember the details.  I could probably dig up a reference
  that would get you to the third party in another one or two
  "hops".  Bob Supnik may also be aware of this effort.

  In addition, Digital-CSS was designing an emulator for a much-
  lower-end CDC system.
                                   Atlant
3687.5MAIL2::CRANEFri Feb 10 1995 12:262
    I know this probably isn`t the place to ask this but how fast is the
    Alpha chip in MIPS? 
3687.6TLE::REAGANAll of this chaos makes perfect senseFri Feb 10 1995 12:5012
    MIPS measurements are not considered accurate in this day and age.  We
    talk about SPECmarks and SPECrates these days.
    
    And which Alpha chip are you talking about?  You question is like
    asking "How fast does a Chevrolet go?"  Unless you mention which
    model and engine, we have no idea where to start!
    
    You can find some SPEC comparisions of assorted Alpha models at:
    
    PERFOM::CSG_REPORTS:SPEC*COMPARISON*.PS
    
    				-John
3687.7MAIL2::CRANEFri Feb 10 1995 12:564
    .6
    Could I have a mean speed?
    
    Thanks.
3687.8AXEL::FOLEYRebel without a ClueFri Feb 10 1995 13:519
RE: .7

	From fast to wicked fast.

	There are many variations of Alpha chips from the 166 Mhz LCA chip
	to the 275 Mhz (and supposedly running today, a 300+ Mhz) Alpha.

							mike
3687.9MAIL2::CRANEFri Feb 10 1995 13:562
    .8
    Thanks.
3687.10QUARK::LIONELFree advice is worth every centFri Feb 10 1995 14:045
Note also that you can't use the MHz rating alone as an indicator of
performance when comparing to other architectures (or even to other Alphas
in some cases.)

					Steve
3687.11AlphaCyber pointerSSDEVO::PARRISRAID-5 vs. RAID-1: n+1 << 2n, in $$$Fri Feb 10 1995 14:122
Info about the software to ease the Cyber-to-Alpha migration is contained
in VAXAXP::NOTES$ARCHIVE:ALPHANOTES_V1 note 3293.1.
3687.12ATLANT::SCHMIDTE&RT -- Embedded and RealTime EngineeringFri Feb 10 1995 15:037
Keith:

> -< AlphaCyber pointer >-

  "AlphaCyber" is the one I was thinking of -- thanks for the pointer!

                                   Atlant
3687.131 BIPS (1000 MIPS)ANGLIN::SULLIVANTake this job and LOVE itMon Feb 13 1995 03:048
>                       <<< Note 3687.5 by MAIL2::CRANE >>>
><
>    I know this probably isn`t the place to ask this but how fast is the
>    Alpha chip in MIPS? 


Thier was a press release that the latest/fastest Alpha runs at 1 BIPS
that is 1000 MIPS.
3687.14Mighty fast he is ...RTOEU::KPLUSZYNSKIDon't wait too long ...Mon Feb 13 1995 05:4810
    The 21164 is qudruple-issue (can handle up to 4 instructions simultaneously)
    At 300 MHz clockrate the peak rate is 1.200 MIPS. (300 Mhz x 4)
    
    Note that this is a theoretical rate. Real world performance depends on 
    the application and the level of optimization a compiler does.
    
    SPECmarks or TPC ratings are better suited for comparing systems from
    different vendors for a given task.
    
    Klaus
3687.15MAIL2::CRANEMon Feb 13 1995 11:451
    Thanks all for your input and to Mr. Hoffman who took this off line.
3687.16GANTRY::ALLBERYJimMon Feb 13 1995 19:565
>>    The 21164 is qudruple-issue (can handle up to 4 instructions simultaneously)
>>    At 300 MHz clockrate the peak rate is 1.200 MIPS. (300 Mhz x 4)
    
    			That would be	    1.200 BIPS
    
3687.17cultural myopia?KLAP::porterthe mantra of the walls and wiringMon Feb 13 1995 21:2815
>>>    The 21164 is qudruple-issue (can handle up to 4 instructions simultaneously)
>>>    At 300 MHz clockrate the peak rate is 1.200 MIPS. (300 Mhz x 4)
>    
>                        That would be       1.200 BIPS
>    

Nonsense.  The '.' is the thousands delimiter, so "1.200" clearly
reads "one thousand two hundred".

Are you perchance getting confused with the delimiter which separates
the whole part from the fraction?  That's ",".

So, 1.200 MIPS is equal to 1,200 BIPS.


3687.18NOVA::FISHERnow |a|n|a|l|o|g|Tue Feb 14 1995 09:375
    or 1 1/5 BIPS to be unambiguous?
    
    :-)
    
    ed
3687.19Or maybe GIPS?BHAJEE::JAERVINENOra, the Old Rural AmateurTue Feb 14 1995 10:151
    
3687.20KLUSTR::GARDNERThe secret word is Mudshark.Tue Feb 14 1995 10:467
	uggg....not another BIPS vs GIPS rathole????

	1.2 Billion Instructions Per Second

	at best ;-)

	_kelley
3687.21Rathole warning!BHAJEE::JAERVINENOra, the Old Rural AmateurTue Feb 14 1995 12:122
    Just as the meaning of "." and "," is somewhat dependent on where you
    are, so is the meaning of 'billion'...
3687.22KLAP::porterthe mantra of the walls and wiringTue Feb 14 1995 12:165
re .-1

Yeah, although it appears the British news media (well, the
BBC world service at least) have given in to
American hegemony in that regard.  Sigh.
3687.23BHAJEE::JAERVINENOra, the Old Rural AmateurTue Feb 14 1995 12:295
    yeah, but I'd hazard a guess that in most of the rest of Europe, a
    billion is still 10**12...
    
    Of course, I'd have nothing against a machine that performs at 1,2
    [German] billion instructions per second... :-)
3687.24Re: The Billions versus Billions rathole...ATLANT::SCHMIDTE&amp;RT -- Embedded and RealTime EngineeringTue Feb 14 1995 13:493
  Why don't you guys go settle it out in the Mill Yard?

                                   Atlant
3687.25REMQHI::NICHOLSTue Feb 14 1995 13:594
>  Why don't you guys go settle it out in the Mill Yard?

    'twould be a shame to do that before the senior citizens get settled in
    there to enjoy the entertainment.
3687.26Attempted Rathole Redirection AlertHLDE01::VUURBOOM_RRoelof Vuurboom @ APD, DTN 829 4066Tue Feb 14 1995 14:271
    Its not about BIPS vs GIPS, its about "." vs ",".
3687.27QUARK::LIONELFree advice is worth every centTue Feb 14 1995 14:376
Re: .22

No they haven't - the BBC World Service consistently uses "thousand million"
for what we in the US refer to as "billion".

				Steve
3687.28"World" is another one of those words...KLAP::porterthe mantra of the walls and wiringTue Feb 14 1995 15:393
Hmm.  Maybe I've been listening to the World (as in "World
Series") Service instead of the World (as in "planet Earth")
Service!
3687.29scientific notation impliedMPGS::STANLEYI'd rather be fishingTue Feb 14 1995 16:092
    If one were to use "." as has been discussed here is there an
    assumed X 10 to the +3 on the tail end? If so I agree. 8>)
3687.30Who really cares!GLDOA::WERNERTue Feb 14 1995 18:396
    Rathole is too grand a term for the current state fo this pathetic
    sting, perhaps vermin nit or mouse pimple or rodent spec would be more
    appropriate. 
    
    
    -OFWAMI-
3687.31World Series???MIMS::SANDERS_JTue Feb 14 1995 18:535
    re. 28
    
    I don't think you have been listening to the "World Series" lately and
    maybe not in the future.
    
3687.32GEMGRP::gemnt3.zko.dec.com::winalskiCareful with that AXP, EugeneWed Feb 15 1995 21:318
RE: .24

>  Why don't you guys go settle it out in the Mill Yard?

Or better yet, settle it out in the milliard (the British term for a 
thousand million).

--PSW
3687.33COVERT::COVERTJohn R. CovertWed Feb 15 1995 22:251
Duh.