MTDEW
Diamond Member
Ok, doing some experimenting here and getting some odd results.
I have two types of ram here to experiment with.
2x512 of Kingston DDR433
2x1024 of Corsair Value select DDR400
Now to find my max HTT/FSB i set the ram to 100, multiplier to 5 and HT to 2x.
This is to take cpu and ram out of the equasion and isolate the fsb of course.
Then i boot to my desktop.
open prime95, and start an "in-place fft" torture test.
increase the FSB (HTT) with clockgen 2-3mhz every 15-30 seconds
With the corsair ram my fsb locks or reboots at 228fsb!
With The Kingston i get up 280fsb before lock or reboot.
So i experimented a bit farther using the Kingston and changed the timings from spd to
8-3-3-3 2t to see if the fsb would go farther.....nope with the more relaxed timings it tops out at 240fsb.
Odd how when i'm actually trying to take the memory out of the equasion that it still seems to have an effect.
My best "guess" is it has to due with my cpu's memeory controller.
So from seeing this it seems that the "type" of memory could limit OCing potential of the cpu. (a higher fsb can result in higher overclocks due to not being able to raise the multiplier)
Can anyone else who has extra sticks to try verify that this happens on their setup also.
I just dont fully understand how having the memory WAY underclocked it still has an effect on the HTT/FSB!
I have two types of ram here to experiment with.
2x512 of Kingston DDR433
2x1024 of Corsair Value select DDR400
Now to find my max HTT/FSB i set the ram to 100, multiplier to 5 and HT to 2x.
This is to take cpu and ram out of the equasion and isolate the fsb of course.
Then i boot to my desktop.
open prime95, and start an "in-place fft" torture test.
increase the FSB (HTT) with clockgen 2-3mhz every 15-30 seconds
With the corsair ram my fsb locks or reboots at 228fsb!
With The Kingston i get up 280fsb before lock or reboot.
So i experimented a bit farther using the Kingston and changed the timings from spd to
8-3-3-3 2t to see if the fsb would go farther.....nope with the more relaxed timings it tops out at 240fsb.
Odd how when i'm actually trying to take the memory out of the equasion that it still seems to have an effect.
My best "guess" is it has to due with my cpu's memeory controller.
So from seeing this it seems that the "type" of memory could limit OCing potential of the cpu. (a higher fsb can result in higher overclocks due to not being able to raise the multiplier)
Can anyone else who has extra sticks to try verify that this happens on their setup also.
I just dont fully understand how having the memory WAY underclocked it still has an effect on the HTT/FSB!