mslater@cup.portal.com (Michael Z Slater) (06/19/89)
>Can someone explain the basic differences between the Cypress and Fujitsu >SPARC implementations? Are they binary code compatible? They are binary compatible, although the Cypress chip has a couple of instructions that the Fujitsu does not. There are two Fujitsu chips: the original gate array, now called the S-16, which runs at 16 MHz, and a standard-cell reimplementation, called the S-20 and S-25, at 20 and 25 MHz. The Cypress chip is a full-custom design, and runs at up to 33 MHz. All three designs have different pinouts. The Cypress design has a different hardware structure for the coprocessor interface. All of these designs are pretty primitive -- No on-chip floating-point, MMU, or cache control. Things will get more interesting in the next generation. Michael Slater, Microprocessor Report mslater@cup.portal.com
aglew@mcdurb.Urbana.Gould.COM (06/25/89)
>>Can someone explain the basic differences between the Cypress and Fujitsu >>SPARC implementations? Are they binary code compatible? > >They are binary compatible, although the Cypress chip has a couple of >instructions that the Fujitsu does not. Terminology check - what does this mean? Possibilities: (1) the Cypress chip and the Fujitsu chip share all user mode instructions, but there are a few priviliged instructions that differ. (2) the Cypress chip implements in hardware several instructions that are in the SPARC architecture definition, that the Fujitsu chip traps for, and for which a conforming implementation is required to provide emulation code. (3) the extra instructions are accessible from user mode and not emulated, but are flagged "don't use these instructions if you want to be portable to other SPARCs". (I would sure hope that there's a compiler switch in this case).
ingoldsb@ctycal.COM (Terry Ingoldsby) (07/22/89)
Can someone explain the basic differences between the Cypress and Fujitsu
SPARC implementations? Are they binary code compatible?
If this is a question that has been discussed 10**10 times before, then
just send me email. Thanks in advance.
--
Terry Ingoldsby ctycal!ingoldsb@calgary.UUCP
Land Information Systems or
The City of Calgary ...{alberta,ubc-cs,utai}!calgary!ctycal!ingoldsbkhb@chiba.Sun.COM (chiba) (07/23/89)
In article <351@ctycal.UUCP> ingoldsb@ctycal.COM (Terry Ingoldsby) writes: >Can someone explain the basic differences between the Cypress and Fujitsu >SPARC implementations? Are they binary code compatible? > From an application programmers point of view all existings (and all proposed) SPRC's are the same. Binary compatibility exists (try your sun4 code on a solb... :>) 4/2xx application code works fine on SS-1 and SS 330. There are one or two places where the OS may have to care (page size on the SS-1 is smaller, for example). Electrically the chips are different. The current Cypress set is faster than the current Fujitsu set. I am sure someone from Cypress (or their ROSS subdivision) will rise up to tell us all about the neat electrical stuff ... :> Keith H. Bierman |*My thoughts are my own. Only my work belongs to Sun* It's Not My Fault | Marketing Technical Specialist ! kbierman@sun.com I Voted for Bill & | Languages and Performance Tools. Opus (* strange as it may seem, I do more engineering now *)
mo@prisma (08/03/89)
Are they binary compatible: Yes, all SPARC machines are binary compatible at the user program level. OS code can tell the difference that some machines have different MMUs or different floating point units (number of functional units, for example, may effect fault recovery). All user programs should run identically. -Mike