Oldschool slot hardware questions.

[benjagan91]

Hi,

I'm just reading up on some old Intel Tech... and i figure some of you veterans might be the guys to ask my questions to... since some of you probably used the tech I'm about to discuss?

Which slot was best?

So what is better... slot1 or slot2? From what i can tell... the slot2 cpu cache ran at 100% speed where as the slot1 cpu cache was at 50% Also.. The FSB protocol for the slot1 is GTL+ where as the slot2 is GTL+ and later AGTL+... whatever that means?


Which slot2 cpu was best?

Reading wikipedia reveals to me that 1000mhz was the fastest P3 xeon released. It had 256KB L2 cache and ran on a 133mhz FSB. BUT then there is a lower clocked 900mhz P3 Xeon with 2048KB L2 cache which ran on a 100mhz FSB... which is better? From what i can understand... the higher FSB will equal more memory bandwidth? Higher the clockspeed equals higher the performance? And the same again with cache? Argh confusing...


Comparison of best slot solutions

So whatever you answered to be the best cpu for slot2... does it beat a P3 1000mhz with 256kb cache running on a 133fsb? How does a Tualatin core P3 via a slotket compare? (Tualatin @ 1.4ghz / 512kb L2 / 133 FSB)

And to make things exciting... imagine all options are being run in SMP (Dual CPU boards).

Thanks for your time =)

Regards, Benjagan91

[terrastrife]

we arent allowed to do your homework on ocau

for anything else, theres wiki.

all the slots sucked, why do you think we went back to sockets/arrays.

cache speed was dependant on the cpu, not the slot.

ignore the bus.

i doubt very few of us worked with slot 2

requires chipset/voltage support first.

as above.

[benjagan91]

I graduated last year

Thanks for the reply

[Utetopia]

The 900MHz CPU with 2MB cache (on-die??) will still be slower than a 1400MHz Tualatin CPU with 256KB for most apps. Very few apps back in those days would utilize the extra cache of the 900MHz Xeon.

IIRC, no Slot (1 or 2) CPU could run the cache on-die. If there were Socket 370 chips with more than 2MB of on-die cache it would be interesting to see how much cache actually gets used for the OS & applications.

edit: Slot or socket. Cache speed is dependent on the CPU, as Terrastrife has mentioned above. L2 cache for a 900MHz Xeon would've run at either 450 or 300MHz. Inefficient if you ask me.

[BuuBox]

Originally Posted by terrastrife

all the slots sucked, why do you think we went back to sockets/arrays

Cost reasons due to the expensive packaging?