[SGVLUG] Linux Sonoma (Centrino) Support
Chris Smith
cbsmith at gmail.com
Thu Sep 22 14:53:45 PDT 2005
On 9/20/05, Dustin <laurence at alice.caltech.edu> wrote:
> On Tue, 20 Sep 2005, Chris Smith wrote:
> > On 9/20/05, Dustin <laurence at alice.caltech.edu > wrote:
> > > On Tue, 20 Sep 2005, Chris Smith wrote:
> > ...I think the poser value of having a flashy
> > looking laptop won out over the "how much can I do with Linux?"
> > factor.
>
> me->CrossOffList("Sony Vaio");
>
> I'm not that into form over function. :-)
Me either. There are some Sony's that score very high on the function
side of things, so don't rub them out entirely.
> > about such things are very much aware of it. We just aren't that
> > excited because when it comes to integer performance, the x86 chips
> > still beat it.
>
> Recall in another life I was a physicist. Floating point speed gives me a
> warm feeling all over. :-)
So you must really love the Itanium2. ;-)
> > Yeah, and when you look at the PPC ISA, it's only slightly less ugly than x86.
>
> I haven't really. I don't have a PPC chip to my name, for one thing. I
> recall MIPS was quite clean, and I kind of think Alpha was reasonable
> though now that I think of it I don't know why I think that.
Think of PPC is being like MIPS or Alpha, only their first version
didn't work too well (IBM RT), so they kind of hacked back in some of
the CISC-y instructions they knew and loved.
> > >From what I understand, cache is actually fairly cheap compared to
> > just about anything else you can do with your transistor budget. They
> > tend to be far more compact and fault-free than other parts of a chip.
>
> My understanding was that the cache was what killed the PPro in terms of
> price, but it had some funky system where the cache was bonded onto the
> CPU or some such wierdness that drove the cost up.
Well, it had a weird design where the cache was on it's own die, but
it was bonded to the chip early on, so a flaw in either die meant you
had to scrap both.
The seperate die thing though is a big clue about how the PPro is
different from even the Itanium. The cache was so massive that it
actually had it's own die that was about the same size as the main
CPU. This is particularly impressive, given that you can pack
transistors for a cache more tightly than for the rest of the CPU. To
give you an idea of just how much of an impact this had on transistor
budget, the PPro's transistor budget sans cache was 5.5 million (for
comparison a modern single-core Pentium4 has about 125 million
transistors). The 256KB cache was 15.5 million transistors. The 1MB
cache was 62 million. On top of that the cache was locked to the clock
rate of the CPU core, which is more than a little crazy (in a lot of
ways it was almost like Intel added in some slightly slower 64,000
anonymous general purpose registers ;-).
So yeah, the cache did kill the price, because at a minimum it
quadrupled the transistor budget to a size. In fact, the transistor
budget for the 1995 Pentium Pro w/512KB L2 cache was comperable with
budgets of CPU's (Celerons and the like) produced a decade later, and
the 1MB model that was introduced later has a transistor budget that
is actually not too far off from a top of the line Pentium M with 1MB
of L2 cache. From that perspective, the price for a Pentium Pro was
pretty amazing. How often do you get to buy a CPU with a transistor
budget from a decade in the future?
> It might not have any relevance to current chips, but given that the HPC guys tell
> me that nothing really matters any more besides avoiding cache misses, I assume
> cache sizes will only go up, maybe faster than main memory.
Oh there are other things besides cache misses, but they do turn out
to be a huge deal.
> > Having a shorter pipeline really, really helps. The P4's deep pipeline
> > requires almost the same kind of challenging thinking needed for EPIC.
> Epic? I assume you don't mean the IRC client. :-)
EPIC is effectively Intel's proprietary name for VLIW. There are some
subtle differences that someone explained to me once, but which seemed
so trivial I promptly forgot.
--
Chris
More information about the SGVLUG
mailing list