Date: Fri, 20 Apr 2001 10:35:34 -0700
Reply-To: "William W. Viergever" <wwvierg@ATTGLOBAL.NET>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: "William W. Viergever" <wwvierg@ATTGLOBAL.NET>
Subject: Re: Dual Pentium IIIs or P4?
At 04:00 PM 4/20/2001 +0000, Puddin' Man wrote:
>William W. Viergever wrote:
> > The only other issue -and with SAS dataset size limitations, it's a
> > non-issue- is the overall ram afforded by a 64-bit P4 vs. a 32-bit PIII ...
>Just curious: you "see" P4 as a 64 bit architecture? No doubt you've heard
>about Merced, Itanium, IA-64 (or whatever they are calling it nowadaze)??
Yeah - the ephemeral "Itanium", 64-bit Linux, etc. *real* 64-bit machines,
but .... I guess I was thinking more of the bus imporvemnets vis-a-vis the
>400MHz System Bus
>The fastest Pentium III available is designed to work with a 133MHz front
>side bus (FSB). With approximately 1GBps of bandwidth, this FSB acts as a
>bottleneck between the CPU and the rest of the system. Intel will remove
>this bottleneck with the Pentium 4 by using a 100MHz FSB quad-pumped to
>the equivalent of 400MHz. 400MHz on a 64-bit bus yields 3.2GBps of data
>bandwidth, three times what the Pentium III's 133MHz FSB can handle. This
>bandwidth will keep the CPU well connected to the i850 chipset, codenamed
>Tehama, which will support dual RDRAM channels.
What's interesting, is that Intel calls the P4 an IA-32, which is different
than the PIII which is still a P6 architecture?!?!?
> > terabytes (?) compared to gigabytes (don't quote me on this last part). But
>You are talking about the ceiling on physical memory? It is actually
>determined by the chipset (not the cpu)?
Semantics, me thinks, Puddin' Man: for a 32-bit machine can *theoretically*
address 4GB of memory space; thus a 64-bit machine should be into terabytes
IIRC, the original Motorola 68000 chip, had 32-bit registers and was thus
*essentially* a 32-bit processor, and yet it had only 16 external data pins
and 24 external address pins. It's smaller brother, the 68008 chip,
addressed only 20 bits of address space and had the same 1 megabyte memory
limitation as the 8086.
If this pin-out stuff is what you mean by the chipset, then I'd say we're
I always, get this word size (bits and bytes), address space, phsical
memory, registers, etc. stuff confused <g>.
IIRC, Intel's 1st (early?) 16-bit 8086 processor had 8 16-bit registers and
could manipulate 16 bits of data at a time. It could also address 16-bit of
address space at a time (or 64K, similar to the Atari 800 and Apple II of
the time). Via "segment" registers, a program could simultaneously address
4 such 64K segments at a time and have a total of 1 megabyte of addressable
memory in the computer. It's from here, I believe, we inherited our famous
640K RAM limitation of DOS, since the remaining 384K was used for hardware
and video drivers/etc. So here, as you explained above, we do have a
difference between *addressable* and *physical* memory.
Yikes! - you're right!
>No! You must be talking about memory addressing, eh?. Well, to my po'
>knowledge, P4 does 32 bit memory addressing like P3.
See above - you're correct.
> > given that P4's *only* can use RAMBUS memory, which is quite expensive
> right now (monopoly in fact), this is not that much of an advantage
> either (Note: PIII's can use PC-133 SDRAM *or* RAMBUS).
>Some might consider it a distinct disadvantage.
RAMBUS considers it advantage <g>. But I'll rephrase: for *me* the high
cost of RAMBUS makes any RAMBUS performance advantage moot.
>I think Intel is expected to support SDRAM and/or DDRAM with P4 _eventually_.
their "Brookdale" chipset described in the link below? -- true, but it
won't work w/ the quad channel 400 mhz bus that gives RAMBUS a good chunk
of it's stellar performance differential.
> > FWIW, I have a dual-550 PIII and a single PIII 800. The 800 is a consistent
> > 10-15% faster for disk based stuff (e.g., creating a 32MB file from two
> > nested do 1 to 1000 loops which outputs 4 numeric vars) and even more so
> > (25 - 30%) for the more CPU oriented tasks (do freqs, means, summaries off
> > that 32MB file) than the dual-500 and both are using Ultra 160 SCSI hard
> > drives.
>This is 100% consistent with what I would've expected.
>To my knowledge, SAS is _not_ "effectively multi-threaded applications
>software" on Intel-based systems.
Good place to QUIT; STOP; - for here we are in agreement! <vbg>
10-4 - over & out, or from my Wylbur days:
>Peter Ivanick wrote:
> > See: http://www.sharkyextreme.com/hardware/roadmaps/p4_brookdale_update/
> > & the ff. quote:
> > Intelís Thermal Design Guide has revealed that the absolute maximum
> > power dissipation of the 1.5GHz P4 is actually 72.9 watts. This
> > is 33% higher than the published system design specification, and
> > essentially identical to the 1.33 GHz Athlon. In order to prevent
> > the CPU from exceeding 54.7 watt, thermal throttling is used. If
> > performance critical applications drive CPU power above its
> > artificially low 54.7 watt limit, the CPU is halted with a 50% duty
> > cycle (alternating 2 microseconds on; 2 microseconds off) until
> > it cools down. This effectively turns your 1.5GHz processor into
> > a 750MHz processor ≠ just at the moment you demand peak
> > performance. On the other hand, you will probably still be able
> > to check your email at 1.5GHz.
>Does anybody remember the earliest Pentiums (i.e. 60, 66 mhz)?
>They ran on 5 volts, generated tons of heat, and had numerous
>problems. Took 'em a while to sort it all out.
>Even if my rich uncle died, I don't think I would be buying
>P4 at this point in time (just kiddin' about the rich uncle :-)).
>*** Puddin' Man *** Pudding_Man@postmark.net *****
>"I don't believe you really, really love me ...
> You just like the way my music sound."
> - John Lee Williamson, maybe 1937
William W. Viergever Voice : (916) 483-8398
Viergever & Associates Fax : (916) 486-1488
A SAS Institute Quality Partner (USA) E-mail : firstname.lastname@example.org
Sacramento, CA 95825