Archived from groups: alt.comp.periphs.mainboard.abit,alt.comp.periphs.mainboard.asus (
More info?)
In alt.comp.periphs.mainboard.abit, Kokoro ordered an army of hamsters
to type:
> In alt.comp.periphs.mainboard.abit, GlassVial ordered an army of
> hamsters to type:
>
>> Hey Wes, do you have any websites with benchmarks of xp vs sempron
>> cpu's of the same "speed" ? Like an Athlon XP 2600 vs a Sempron
>> 2600? All I've seen is the Sempron vs Celeron info, and Sempron
>> appears to destroy the celery handily.
>>
>> -GV
>>
>
>
>
> Stop pestering poor Wes.
>
> Its easy to work out the performance of a cpu, forget the stupid names
> they are given for a minute.
>
> Lets take your two choices a 2600+ Sempron and AthlonXP (32 Socket A)
>
> The faster the Bus speed the faster the performance and the faster the
> total speed the faster the performance (total speed being multiplier
> multiplied by the bus speed)
>
> Sempron 2600+ has a bus speed of 166mhz, not 333mhz as a lot of people
> say. there is a reason for the 333 but its irrelevent for the time
> being.
>
> This Sempron speed is 1.83ghz, thats 1830mhz, 1830/166 = 11 which is
> the multiplier but its not important, the other two figures are.
>
> Athlon XP 2600+ (Barton) speed is 1.92GHz (1920mhz) but also runs at a
> bus speed of 166. Because of the higher total speed of the Athlon it
> is a faster cpu.
>
> An Athlon XP 2600+ Throughbred is 2.07ghz but the smaller cache size
> hinders performance.
>
> It is also worth noting that Semprons have a debilitating 256k L2
> cache like the Throughbreds and the Bartons have 512k to play with.
>
> If wanted a true 2ghz cpu then another choice is a 2400+ AthlonXP32
> but, its bus speed is 133mhz and because of this it wouldn't be as
> good performer as a Sempron 2800+
>
> the 2600+ Sempron however is meant to mean it has a similar
> performance to a 2.6Ghz Celeron and a 2600+ AthlonXP is meant to have
> similer performance to a 2.6Ghz Pentium4... in practice the
> performaces maybe different.
>
> All you need is a little information to work out a cpu's
> performance.... not the stuff a pr department churns out.
>
Does anyone else think there should be a website with information like this
that can be found easily and at a glance?