[rescue] who here is familiar with turbosparc 170's?

[James Lockwood]

On Wed, 8 May 2002, Loomis, Rip wrote:

> > What can you tell about these?
> They're pretty stinkin' fast--slightly faster than
> an SS20/single SM71 based on SPEC-CINT95.  Otherwise
> they're pretty much like any other sun4m, just with
> TurboSPARC instead of MicroSPARC/SuperSPARC.  It's been
> a while since I had one around to play with.  Only
> negative thing I remember someone else complaining about
> is that they tended to run hot compared to 75/85/110
> SS5 systems.

They also have an interesting exploitable CPU implementation bug: any
userland program can hard-hang the system.  The code necessary to do this
is not generated by a compiler except under rare circumstances, but it is
possible.  There is no known workaround.

I find it interesting that this received so little press compared with
(relatively) minor errata in the x86 family.

-James