if im not mistaken,the ram's higher clock speed will benefit much more on P4 platforms than tighter timings.there was an article on this that i read, that showed quite a significant difference in SiSoft memory benchmark.but that article was published before the days of the A64..

for me, i can see that higher ram clock speed is better that tighter timings because my rams are not capable of 2-2-2-5 at higher clocks.IIRC, my rams can take 2-2-2-5 only till DDR410.i did a superpi comparison with the rams at DDR410/2-2-2-5 with DDR500/2.5-4-3-5, and the second setting gave better results (cpu speed was kept the same).

yeah, you're right but he stated to do little oc.

opps.sry,my earlier post didnt mean to answer the topic starter.just wanted to share some info in this discussion

anyway,im sure tighter timings will be better than higher fsb since he wants do to little oc

both playing much roles in overclocking.high FSB but loosen sometimes can be defeat by lil' lower FSB wif tight timmings

yes,,, both with same spec!

wrong

lower fsb + tight timming never beat high fsb + loose timming

but then again... it depends on how high the fsb and how tight/loose the timing is

hehe. what kind of statement is that - in order to compare - there must be a base figure for comparison.

what timings are you referring to?
what speed are you referring to?

high fsb loose timing vs low fsb tight timing <=--- this is too general to even be a comparison.

give a specific speed vs a specific timing.

so let's go based on what the thread starter posted

200mhz 2-2-2-5 vs 217mhz 3-3-3-6
or how about like
200mhz 3-3-3-8 vs 226mhz 3-4-4-11 (yes, i'm using 3-4-4-11 )

let's stop the nonsense statements like the above quoted. be specific on what you meant by high / low and tight / loose.

depends on the platform i say

on intel machines, higher timing higher fsb wins than lower timing since intel needs more mhz as the memory controller aint as efficient as the nforce

in amd xp platform via the nforce 2, tighter fsb timings is proven to be faster than higher clock speeds.


how ever, all this is useless if you have only 256mb or 512 mb ram.....the more ram you have will solve what ever problem you are facing

ok...lets make it like this then...

typical BH's can go up to 260Mhz retaining is super tight timings which is 2-2-2-x

and a typical TCCX can go up to 315Mhz with 2.5-4-4-8....

assume the clockspeed is diffrent between those two but is is the same processor...the battle should be won by the TCCD in most benchies except maybe math test by SPI and Sandra....which gives us a result in making a judgement..High FSB wins....

but when the test really needs no diffrence between clockspeed(i.e lowered the multiplier if run on TCCD)....then only IMHO the tight timings can show a diffrent in results....

so, the best thing is low timing running on divider........
speaking on divider, i really awnt to know that running on divider for long time can harm anything such as the chipset ??

no such thing ler... don't be afraid

okok, thanks for the info

all the same !

no changed!

This post has been edited by Mr_47: Dec 31 2005, 12:11 PM

as for me i vote tight timing becoz to achive high mhz with tccd (e.g 315mhz) u need to lower some option in bios in order to achive that..

but i have tried same dual channel bh5:

1. mushkin lvl 2
2. twinmoss sp bh5 utt

both at same speed 255mhz.. the mushkin give a little higher score that twinmoss. the chip also influnce the score in sandra...

by the way.. i like number 2 rather 4...hehe

be specific please

aihh........ since we're getting nonsense like those from Mr_47

why not everyone just go to..........
http://www.madshrimps.be/?action=getarticl...313&articID=325

have fun reading. draw your own conclusion.

did anybody ever read this!...

thats in term of gaming POV....well we might be in to gaming...but not with our overclocked rig...Overclocked speed is specially for:

-CPU-z
-Sandra Mem Bandwidth Test, CPU arithmatic...
-Super PI
-Aquamark

but the deviders available for decent boards are 200,166,133,100

lets have a fix speed of 9x300
can someone with tccd do some bench 1:1 ddr600 n do 200:166 ddr500 n post here?

i got tccd but mobo wont allow dividers on high HTT....so cant really tell....

but surely lost of marks is visible IMHO...because i tried once by just changing the tref to 100mhz 15.xxns oso drop kaw-kaw even with the same HTT (300)

This post has been edited by antonio_zth: Jan 4 2006, 04:02 PM

LOL. my OCed speed is for gaming. the rest is just for comparison / measurement.