T.R | Title | User | Personal Name | Date | Lines |
---|
488.1 | 21164/400 - another well-hidden Digital secret | GEMGRP::GROVE | | Tue Mar 05 1996 11:00 | 10 |
488.2 | thanks... i'll check... | CRONIC::semi3.hlo.dec.com::notes | the storyteller makes no choice... | Tue Mar 05 1996 11:18 | 3 |
488.3 | 21164-366 SPECfp92 estimate lower than 8400 5/350 SPECfp92 ? | MINNY::usr12.zuo.dec.com::DOLDER | The future isn't what it used to be | Tue Mar 05 1996 12:46 | 7 |
488.4 | | SMURF::MENNER | i'm gonna plant a weeEEping willow... | Tue Mar 05 1996 17:42 | 2 |
488.5 | What would be the point of estimating to 2 decimal places ? | KAMPUS::NEIDECKER | EUROMEDIA: Distributed Multimedia Archives | Wed Mar 06 1996 03:38 | 6 |
488.6 | OSF/1 -> UNIX(tm) | NAMIX::jpt | FIS and Chips | Wed Mar 06 1996 06:45 | 21 |
488.7 | | LANDO::OBRIEN | Give it a TRI | Wed Mar 06 1996 11:16 | 11 |
488.8 | | WRKSYS::INGRAHAM | Andy | Wed Mar 06 1996 11:45 | 6 |
488.9 | | AMCFAC::RABAHY | dtn 471-5160, outside 1-810-347-5160 | Wed Mar 06 1996 12:56 | 1 |
488.10 | We are debugging one here... | KAMPUS::NEIDECKER | EUROMEDIA: Distributed Multimedia Archives | Thu Mar 07 1996 02:42 | 1 |
488.11 | Talk is cheap. We deliver. | PERFOM::LICEA_KANE | when it's comin' from the left | Thu Mar 07 1996 07:43 | 16 |
488.12 | Tuesday? | SAYER::ELMORE | through the looking glass | Thu Mar 07 1996 10:27 | 6 |
488.13 | | QUARK::LIONEL | Free advice is worth every cent | Thu Mar 07 1996 11:47 | 3 |
488.14 | what EV56 systems were announced? | NPSS::WATERS | I need an egg-laying woolmilkpig. | Thu Mar 14 1996 11:16 | 3 |
488.15 | | QUARK::LIONEL | Free advice is worth every cent | Thu Mar 14 1996 11:51 | 4 |
488.16 | | YIELD::HARRIS | | Thu Mar 14 1996 12:56 | 10 |
488.17 | "Unprecidented Performance for the Desktop" | PERFOM::LICEA_KANE | when it's comin' from the left | Thu Mar 14 1996 17:26 | 15 |
488.18 | Bret SPEC is 11.3 int, 14e fp | I4GET::HENNING | | Thu Mar 21 1996 21:57 | 18 |
488.19 | A or not an A | ULYSSE::REVEMAN | Scan his brain, it must be there somewhere... | Fri Apr 05 1996 05:05 | 6 |
488.20 | | WRKSYS::INGRAHAM | Andy | Fri Apr 05 1996 09:19 | 5 |
488.21 | | QUARK::LIONEL | Free advice is worth every cent | Fri Apr 05 1996 09:49 | 5 |
488.22 | | WREATH::SNIDER | Because that's the way it IS! | Fri Apr 05 1996 10:07 | 13 |
488.23 | EV5 and EV56 machines are both sold at 366 MHz too | NPSS::WATERS | I need an egg-laying woolmilkpig. | Fri Apr 05 1996 10:34 | 6 |
488.24 | | SOAK2::neth | Craig Neth | Fri Apr 05 1996 10:36 | 4 |
488.25 | EI_STAT | LANDO::OBRIEN | Give it a TRI | Fri Apr 05 1996 12:42 | 13 |
488.26 | | WRKSYS::INGRAHAM | Andy | Fri Apr 05 1996 13:03 | 4 |
488.27 | | WRKSYS::INGRAHAM | Andy | Fri Apr 05 1996 13:07 | 5 |
488.28 | | PCBUOA::KRATZ | | Fri Apr 05 1996 13:57 | 2 |
488.29 | re .28 Uhu ?? | FRIXEU::OLOBARDI | Marco - E&RT Prod Mgr - Italy | Tue Apr 09 1996 09:32 | 7 |
488.30 | | WRKSYS::INGRAHAM | Andy | Tue Apr 09 1996 09:43 | 11 |
488.31 | What he said ;-) (a lot better than I did) | PCBUOA::KRATZ | | Tue Apr 09 1996 15:16 | 1 |
488.32 | | NYOSS1::MONASCH | I wrote the DECmate games | Fri May 10 1996 10:59 | 7 |
488.33 | Yup. It stinks | WRKSYS::INGRAHAM | Andy | Fri May 10 1996 12:56 | 8 |
488.34 | Ignore byte/word - that's the ticket... ;-( | GEMEVN::GLOSSOP | Alpha: Voluminously challenged | Fri May 10 1996 13:00 | 5 |
488.35 | | DECCXL::OUELLETTE | To err is human, to moo bovine | Fri May 10 1996 13:55 | 5 |
488.36 | Party Line? | XDELTA::HOFFMAN | Steve, OpenVMS Engineering | Mon May 13 1996 15:10 | 23 |
488.37 | When is a 21164 not a 21164? | XDELTA::HOFFMAN | Steve, OpenVMS Engineering | Mon Feb 03 1997 11:36 | 77 |
|
::What it boils down to, is this: Digital Semiconductor made a decision.
::EV56 is NOT a different generation. It's just an EV5 with a faster clock
::cycle time.
That's right, it's not a different generation. It is, however,
a different member of the family, one with unique characteristics,
characteristics that need to be uniquely identified to customers,
engineers, etc...
: We (programmers, system builders, and custoemrs) have to be able to
: tell the two apart, because no matter what the folks at DS say, the
: EV5 and EV56 are different. (Because I suspect we're going to wear
: some egg for this decision otherwise, it would be very useful to hear
: the party line around this particular decision.)
Well, the first `egg' arrived here this morning...
Quoting the customer question:
}Which Alpha (server) models uses this ev56 CPU
}
}1 1000A 300 MHz? 333 MHz? 400 MHz???
}
}2 2000/2100A 250 MHz? 300 MHz? (375 MHz???)
}
}3 4000/4100 All ????????????????????
I looked in the SOC and -- guess what -- they're all listed as the
`21164'. Argh.
--
/*
// ALPHA_IMPLVER_AMASK_FEATURES.C
// Retrieve the Alpha microprocessor family and features
//
// Author: Steve Hoffman, Digital Equipment Corporation
// [email protected], 16-Jan-1997.
//
// DEC C V5 and later under OpenVMS.
// Also operates under DEC C on Digital UNIX V4.0 and later.
//
// See the Alpha Architecture Reference Manual for the current
// microprocessor family codes, as well as the instruction set
// extension definition codes.
*/
#include <stdio.h>
#include <c_asm.h>
main()
{
int implver = asm("implver %r0;");
int amask = asm("amask %a0,%r0;", -1 );
int definedbits = 0;
printf("Alpha Microprocessor Family:\n");
/*
// The "microprocessor family" includes the named microprocessor,
// and various microprocessors (if any) derived from it.
*/
printf(" DECchip 21%1.1d64 or variant\n", implver );
printf(" (EV%1.1d microprocessor core)\n", implver + 4 );
/*
// The bit is *clear* if the capability is *present*...
*/
printf("Alpha extensions available:\n");
if ( !~amask ) printf(" none.\n");
definedbits |= 1L<<0;
if ( ~amask & 1L<<0 ) printf(" byte-word, et al.\n");
definedbits |= 1L<<1;
if ( ~amask & 1L<<1 ) printf(" square root, et al.\n");
definedbits |= 1L<<8;
if ( ~amask & 1L<<8 ) printf(" Audio and video, et al.\n");
if ( ~amask & ~definedbits ) printf(" Unknown capability bit present.");
return 1;
}
|