Re: Good news for SPARC

From: Casper H.S. Dik (Casper.Dik_at_Sun.COM)
Date: 01/28/04 

Date: 28 Jan 2004 22:42:27 GMT

"Christopher L. Estep" <pghammer21@comcast.net> writes:

>Actually, more than fourteen years. (The SPARC consortium dates back to
>1988.) Among the original consortium members that still exist is Texas
>Instruments (who actually made *most* of the SPARC processors used by
>Sun themselves in the mid-1990s SPARCstation5 and later SPARCstation20).

TI produces them but doesn't design them.

>The microSPARC subarchitecture was developed by TI, not Sun.

Are you sure?

>UltraSPARC was designed/developed by the consortium itself (lead
>designers: TI and Hitachi); however, the only company that wanted to
>*market* UltraSPARC was Sun (and that was largely for MP servers, not
>workstations, at least in North America). Later UltraSPARC *did* appear
>in workstations.

That's definitely wrong; the UltraSPARC was designed by "Sun Microelectronics"
which is part of Sun.

As far as I know only the following SPARC cpus were designed by companies
other than Sun:

        - HyperSPARC (Ross)
        - SPARClite (Fujitsu?)
        - Weitek "PowerUp" (IPX/SS2 double clocked chip)
        - SPARC64 (HAL, later Fujitsu; SPARCv9 implementation)

The others I'm fairly certain were designed by Sun:

        SS1/SS2 SPARC CPUs (names?)
        MicroSPARC
        SuperSPARC, I & I
        UltraSPARC, I, II, IIi, III, IIIi (& IV & V)

>Part of the problem wasn't the *restrictive* licensure of SPARC (as
>several posters have pointed out, there were far fewer restrictions on
>SPARC than there were on Intel's x86 architecture), but more on the
>*commoditization* of the x86 architecture, and Sun's far more
>restrictive licensure of Solaris at the time.

Indeed; giving away Solaris/Intel a long time ago would have put Sun
in a much stronger position today.

>While it cost little to license SPARC, it cost a mint to license
>*ancillary* architectures (and while Sun did have the controls to some
>of those licenses, the major lincesees often were more egregious in
>terms of fees than even Sun) needed to make a SPARC-based computer work.

Which architectures would those be?

>Since the early 1990s, there has been little *honest* interest in
>building a *commodity* SPARC computer (and no interest at all in
>building one based on an IBM-standard design). Further, I know of no
>vendor that has tried to design (let alone *succeeded*) in designing a
>SPARC mainboard for IBM-style AT or ATX cases.

Sun (Sun Microelectronics) sells/sold ATX form factor SPARC boards.

Casper 

-- 
Expressed in this posting are my opinions.  They are in no way related
to opinions held by my employer, Sun Microsystems.
Statements on Sun products included here are not gospel and may
be fiction rather than truth.
Quantcast