[Info-vax] OT: news from the trenches (re: Solaris)

[John Reagan]

On Wednesday, March 11, 2015 at 7:21:05 PM UTC-4, Richard Maher wrote:
> On 3/12/2015 7:03 AM, johnson.eric at gmail.com wrote:
> > Yes it has become
> > complicated, but 30+ years of success against a forever changing landscape will do that.
> 
> What? Compared to EPIC?
> 
> I'd be interested in what JR thinks but IMHO he'd be getting his jollies 
> back on a CISC chip.

Instruction sets don't really do much for me anymore.  So many of them...  CDC 6500, Z80, PDP-11, PDP-8 (does that count, there are so few?) PDP-10, VAX, Mips, Alpha, PRISM, Itanium, x86, PA-RISC, etc.

Personally, I don't like all the complicated addressing modes on the VAX and x86.  I actually sorta like the Itanium instruction set.  I like big register sets and I cannot lie.

> 
> Didn't floating-point operations look up the answer in some table somewhere.
> 

Whose floating-point?  Itanium's floating reciprocal is somewhat table lookup for many floating numbers but not for all.

> Won't we (and Rdb) be able to go back to generating machine code on the 
> fly, forget about alignment faults, and be generally lazy again?

I thought Rdb still generated Itanium on the fly.  Norm and I traded lots of email about that back 'in the day'.