DDR Boosters @ FrozenCPU

[deathstar13]

here is some timings ive tested to max's the lower 231 isnt tested to max fsb so 2.5-4-3-5 may go higher than 231
also i ran these tests at 2.5-4-4-7 and can still lower the timiengs yet

"edit" http://www.xtremesystems.org/forums/showthread.php?s=&threadid=35343
http://www.xtremesystems.org/forums/showthread.php?s=&threadid=35147

one has bandwith tests with my old epox p4ca3+ at 261 1:1 2.5-4-4-7
i didnt ever run 5:4 and 231 on that mobo as i couldnt max the ram out and thats at 264fsb 1:1
so its gonna be a guess if 231 and these score are in fact reproducable.

i dont want to derail this thread mabey we should start another?

i did some newer testing at 231fsb and 2.5-4-3-5 and 234fsb 2.5-4-3-5
and also i ran 264fsb 1:1 and 3.0-4-4-8
the only odd thing i see is at 5:4 my efficancy is 84-85% and at 1:1 its 74%
can that be what it is?

also i just ran 231fsb 1:1 2.5-4-3-5 same everything as the 61-6200scores and i get 5593.last pic
the efficiancy dropped when going from 1:1 over 5:4 so that explains why im boosting so much at 231fsb.

have you guys tried 5:4?

 

[batboy]

I've been telling people for a year now that there's nothing wrong with the 5:4 ratio.

Here is what I just now got with my Hyperram PC4200 using the 5:4 ratio and 2.5,4,4,7 timings running at 289 FSB just like Deathstar. This is with my 2.4C in an IC7-G mobo (no volt mods) with GAT settings of A,A,A,D,D.

 

[batboy]

I tried to run that same 289 FSB using the 1:1 ratio to see what the difference was, but my RAM won't quite do it with a 2.8v limit. But here's 285 FSB with the same settings and system.

 

[deathstar13]

hmmm looks exactly what i thought it would
same exact settings also.

i like the 5/4 ratio myself and been sticking to it as i cant fine ram to run 1:1 290 yet but this new cpu should help me get back to 1:1

again ty batboy! big help and problem solver!

 

[deathstar13]

I tried to run that same 289 FSB using the 1:1 ratio to see what the difference was, but my RAM won't quite do it with a 2.8v limit. But here's 285 FSB with the same settings and system.

wooooo sexxayy!
thats what im talking about! im in love all over again!

 

[eva2000]

guys post your unbuffered bandwidth scores that's the real speed of the ram

uncheck 9 boxes listed in this image to run unbuffered

 

[krag]

I tried to run that same 289 FSB using the 1:1 ratio to see what the difference was, but my RAM won't quite do it with a 2.8v limit. But here's 285 FSB with the same settings and system.

Daaaa-aaang bro!! *krag runs to bathroom to change shorts*

Must be nice! THat is some crusher mem bandwidth!!

 

[krag]

hmmm looks exactly what i thought it would
same exact settings also.

i like the 5/4 ratio myself and been sticking to it as i cant fine ram to run 1:1 290 yet but this new cpu should help me get back to 1:1

again ty batboy! big help and problem solver!

Wait a sec....you got 6100mbs in Sandra on 1:1 at 231fsb. Right? With the Hynix mem. set to 2.5-4-4-7???

 

[deathstar13]

Wait a sec....you got 6100mbs in Sandra on 1:1 at 231fsb. Right? With the Hynix mem. set to 2.5-4-4-7???

no i got 5900 at 231 1:1
its the last screenshot i posted in that set of 4 or 5

its the bendwith efficiancu that goes up at 5:4 that does this.thats why i asked batboy to try and duplicate these tests and prove its just not my sticks.

 

[batboy]

Kraig, the 231 is memory frequency, I think he's running 289 FSB at the 5:4 ratio same as what I did.

 

[deathstar13]

Kraig, the 231 is memory frequency, I think he's running 289 FSB at the 5:4 ratio same as what I did.

correct.
289fsb cpu 5:4 231mem 2.5-4-4-7 =6200 bandwith
231fsb cpu 1:1 231mem 2.5-4-4-7 =5500 bandwith still not bad but not as good.

what i do find odd is my 231fsb 5:4 2.5-4-4-7 beating your bh-5 250fsb 1:1 with 2.0-2-2-4
it seems odd its running faster but it is.

"EDIT" i didnt get 5900bw its 5500 at 231fsb 1:1
i corrected the above! dang smal attachments make it hard to read em.

 

[eva2000]

correct.
289fsb cpu 5:4 231mem 2.5-4-4-7 =6200 bandwith
231fsb cpu 1:1 231mem 2.5-4-4-7 =5900 bandwith still not bad but not as good.

what i do find odd is my 231fsb 5:4 2.5-4-4-7 beating your bh-5 250fsb 1:1 with 2.0-2-2-4
it seems odd its running faster but it is.

it ain't odd as sandra buffered doesn't show real speed of ram.. sandra unbuffered shows real speed of ram... buffered takes into account chipset/cpu cache and system etc

 

[krag]

correct.
289fsb cpu 5:4 231mem 2.5-4-4-7 =6200 bandwith
231fsb cpu 1:1 231mem 2.5-4-4-7 =5900 bandwith still not bad but not as good.

what i do find odd is my 231fsb 5:4 2.5-4-4-7 beating your bh-5 250fsb 1:1 with 2.0-2-2-4
it seems odd its running faster but it is.

Awwwww, I get it. ~ding!~ *The little light bulb over kraggy's head pops on*

Yeap! THat be ODD!

So 231fsb at 1:1 is the max for that mem at 2.7v? It don't like 2.8v?

 

[SportyGuy220]

i have to say sandra mem benchmarking is inacurate. Notice my ram in my sig and it says that the SiS pc3200 is tied with my ram. Now i know for a fact that this isnt true. I understand that dual channel is probly worthless on AMD socket A cpu's, but there is no way this is factual.

 

[deathstar13]

Awwwww, I get it. ~ding!~ *The little light bulb over kraggy's head pops on*

Yeap! THat be ODD!

So 231fsb at 1:1 is the max for that mem at 2.7v? It don't like 2.8v?

no dude 264 1:1 3.0-4-4-8 2.7v
260 1:1 2.5-4-4-7

231fsb on the mem 5:4 just happens to be where my cpu maxes out at 289-293fsb
in order to run my cpu maxed i have to use 5:4 as as im 30mhz over my rams fsb limits.

what you not getting is the bandwith efficency goes up 10-12% with 5:4 over running 1:1 couseing ram at 231fsb 5:4 run as fast as ram at 260fsb 1:1

 

[krag]

no dude 264 1:1 3.0-4-4-8 2.7v
260 1:1 2.5-4-4-7

231fsb on the mem 5:4 just happens to be where my cpu maxes out at 289-293fsb
in order to run my cpu maxed i have to use 5:4 as as im 30mhz over my rams fsb limits.

what you not getting is the bandwith efficency goes up 10-12% with 5:4 over running 1:1 couseing ram at 231fsb 5:4 run as fast as ram at 260fsb 1:1

231fsb 5:4 is faster/more efficient than 260fsb 1:1?

Uuuhhh, not in my benchies

250fsb 1;1 gets me 5860
25ofsb 54 gets me 5150

 

[SportyGuy220]

ight i took an ss. here it is.

 

[deathstar13]

231fsb 5:4 is faster/more efficient than 260fsb 1:1?

Uuuhhh, not in my benchies

250fsb 1;1 gets me 5860
25ofsb 54 gets me 5150

if i was able to get my cpu/mobo over 300fsb id be able to run 5:4 250fsb mem and would get ungodly bandwith.but problem is these mobos dont like 5:4 over 300fsb.

when your running 250fsb 5:4 your cpu is at 250fsb yes but the mem is at 200fsb

1:1 doesnt start looking good until over 250fsb

 

[deathstar13]

ight i took an ss. here it is.

i appreciate you helpfullness but i just dont see what this has got to do with what we are discussing. dont think im ignoring you tho im listning.

 

[ForeverUnknown]

are you guys doing anything special to cool your ram?