Archived from groups: alt.sys.pc-clone.dell,comp.sys.laptops (
More info?)
>>>>> P Ruetz writes:
P> "Tom Scales" <tomtoo@softhome.net> wrote in message
news> qf_Ve.69130$p_1.44160@tornado.tampabay.rr.com...
>>
>> "P Ruetz" <parmailbox-news@yahoo.com> wrote in message
news> CoZVe.2581$Ob2.1510@newssvr12.news.prodigy.com...
>>>
>>> "Rob Nicholson" <rob.nicholson@nospam_informed-direct.com> wrote in
>>> message news:dg8joc$jd$1@nwrdmz02.dmz.ncs.ea.ibs-infra.bt.com...
>>>>> The rule of thumb is to add 1 GHz to the P-M processor to get the P4
>>>>> equivalent - so a 3 GHz P4 will run like a 2 GHz P-M.
>>>>
>>>> Thanks, Rob.
>>>>
>>>>
>>>
>>> I've heard similar rough estimates before and they strike me as quite
>>> odd. To first order, the processing power of a given architecture is
>>> proportional to the clock rate, so I would expect a multiplicative factor
>>> rather than an offset.
>>> Peter
>>>
>>
>> I think the original response was a reasonable rough estimate, but agree a
>> factor is more reasonable. I also believe the 1.5 is too low. I believe
>> it is closer to 1.8.
>>
>> Tom
>>
P> That would be pretty impressive (factor of 1.8)! A 2G PM would then perform
P> like a 3.6G P4.
As one point of comparison:
Dell Dimension 8250, 3.06 Pentium. Hyperthreading. Runs 2 Seti's at a time,
long term average, 12 units calc'ed/day.
Dell Inspiron 9200, 2.1 P-M, no hyperthreading, runs 1 seti at a time.
Long term average, 12 units calc'ed/day.
By this benchmark, the multiple is about 1.5.
P> I did hear that the floating point performance of the PM suffers
P> significantly, though, in comparison to the P4. The reasoning was that it
P> was for laptops doing light work (i.e. non-scientific) rather than for
P> workstations. I run a lot of floating point intensive programs and was put
P> off by that claim.
I expect seti does a lot of floating point. Perhaps Tom's 1.8 is a better
measure for more general tasks.
--
Andrew Hall
(Now reading Usenet in alt.sys.pc-clone.dell...)