Jump to content


Photo
- - - - -

USB probs @ fsb 278 and more


  • Please log in to reply
26 replies to this topic

#13 THE_BRAIN

THE_BRAIN

    New Member

  • Members
  • 43 posts

Posted 31 December 2005 - 10:16 AM

Anyway if 3X work good or not a HT multiplier of 4x is a no go for past 250 FSB .


i have 3x but it is the same prob!

#14 Raphie

Raphie

    Geek

  • Members
  • PipPipPip
  • 657 posts

Posted 31 December 2005 - 10:41 AM

BRAIN, it seems that you don't like any of out feedback. we tell you it doesn't run, you tell us IT MUST. this is not going to be a very fruitful discussion this way. You can also ask the OCZ Guys at Extreme Forums how far they think your memory will go. But i guess they have it wrong as well then.

#15 THE_BRAIN

THE_BRAIN

    New Member

  • Members
  • 43 posts

Posted 31 December 2005 - 11:01 AM

BRAIN, it seems that you don't like any of out feedback. we tell you it doesn't run, you tell us IT MUST. this is not going to be a very fruitful discussion this way. You can also ask the OCZ Guys at Extreme Forums how far they think your memory will go. But i guess they have it wrong as well then.



i thing you cant read.... :shake:
you speak always from ocz memory....but that ist NOT the prob!!! Ihave test it with some other ram ADATA PC600 GEIL and more.... AND WITH ALL THE SAME PROB!!!!

you have to learn to read.... first post .... i waste my time with you...

#16 Guest_Flip_*

Guest_Flip_*
  • Guests

Posted 31 December 2005 - 11:18 AM

Hmm...
Very strange, indeed.
Could you try to improve your chipset cooling?
Put some AS5 under it for example.. Or you could take a big fan (Standventilator) blow to it. You never know! Never heard of this problem with a NF4 Board.

NF2 had these problems, but NF4 is not common for this.

#17 THE_BRAIN

THE_BRAIN

    New Member

  • Members
  • 43 posts

Posted 31 December 2005 - 11:27 AM

Hmm...
Very strange, indeed.
Could you try to improve your chipset cooling?
Put some AS5 under it for example.. Or you could take a big fan (Standventilator) blow to it. You never know! Never heard of this problem with a NF4 Board.

NF2 had these problems, but NF4 is not common for this.



i use the default cooling, now i turn the fan on 100% but no chance....the same prob...
i want chage the wlp next time.

#18 syar2003

syar2003

    Member

  • Members
  • PipPip
  • 281 posts

Posted 31 December 2005 - 11:53 AM

Talkin' about reading.
You should read this :
http://www.xtremeres...ead.php?t=37010
(might give you some hints)

#19 Raphie

Raphie

    Geek

  • Members
  • PipPipPip
  • 657 posts

Posted 31 December 2005 - 12:27 PM

i think a can read, but i also think we missunderstand eachother. You will have a hard time with the AMD Memory controller to run ANY ram beyond 270Mhz 1:1 (at maximum multiplier) When lowering your multiplier you can go higher on your 1:1 timings since the lower MP will sort this out. It's all about memory bandwidth. Both your controller and Memoery will have a hard time processing any more data stable than let's say 7000+ in Sandra memory bandwidth benchmark. But that's just my 2cts...

#20 syar2003

syar2003

    Member

  • Members
  • PipPip
  • 281 posts

Posted 31 December 2005 - 12:37 PM

i think a can read, but i also think we missunderstand eachother. You will have a hard time with the AMD Memory controller to run ANY ram beyond 270Mhz 1:1 (at maximum multiplier) When lowering your multiplier you can go higher on your 1:1 timings since the lower MP will sort this out. It's all about memory bandwidth. Both your controller and Memoery will have a hard time processing any more data stable than let's say 7000+ in Sandra memory bandwidth benchmark. But that's just my 2cts...


Exactly , thats what that article/thread at extreme shed some lights on.(in the 2048 section)
I also believs that around 270 +/- is around the spot where you have to turn to using 2T .
2048 even with 2 modules is a lot for the memory controller compared to 2x512MB modules also .
There is also a tip that involves turning off bank interleaving might help higher speeds with high amount of ram .

#21 Raphie

Raphie

    Geek

  • Members
  • PipPipPip
  • 657 posts

Posted 31 December 2005 - 12:44 PM

Exactly , thats what that article/thread at extreme shed some lights on.(in the 2048 section)
I also believs that around 270 +/- is around the spot where you have to turn to using 2T .
2048 even with 2 modules is a lot for the memory controller compared to 2x512MB modules also .
There is also a tip that involves turning off bank interleaving might help higher speeds with high amount of ram .


ThnX Syar2003 :)

#22 J054

J054

    Member

  • Members
  • PipPip
  • 72 posts

Posted 31 December 2005 - 01:31 PM

plz read the first post complet!!! :rolleyes:
- i have some ram ADATA , GEIL.... and with all the same prob over 265mhz
- it MUST RUN with 1:1 !!!


It doesnt have to run 1:1, I encourage you not to run 1:1 so that you can bump up your cpu, that is the only way to do it, put the ram on a divider. Read Angry_Games' sticky on 1:1:

tight timings on a divider, or loose timings 1:1? look here for some results

The only thing that 1:1 275x10 beats divider (9:10) 275x10 in is memory bandwidth...which sorta proves my point that memory bandwidth is useless, and that it is all about cpu mhz, as well as proving that dividers are not always a performance penalty (unless, for some reason, memory bandwidth is that important to you...)



#23 THE_BRAIN

THE_BRAIN

    New Member

  • Members
  • 43 posts

Posted 31 December 2005 - 05:53 PM

Talkin' about reading.
You should read this :
http://www.xtremeres...ead.php?t=37010
(might give you some hints)


wtf ? this is only the "tccd" guid for 256mb & 512MB moduls and my adata is hynix d5 & GEIL ist Hynix D43 and my ocz is Infineon CE-6

i dont neet a tccd guid...


i think a can read, but i also think we missunderstand eachother. You will have a hard time with the AMD Memory controller to run ANY ram beyond 270Mhz 1:1 (at maximum multiplier) When lowering your multiplier you can go higher on your 1:1 timings since the lower MP will sort this out. It's all about memory bandwidth. Both your controller and Memoery will have a hard time processing any more data stable than let's say 7000+ in Sandra memory bandwidth benchmark. But that's just my 2cts...



thx but my sys runs 100% stable @ 278Mhz FSB 1:1 - 12h prime95 & OCCT and 5h MemTest
Only the USB sux !!!

#24 THE_BRAIN

THE_BRAIN

    New Member

  • Members
  • 43 posts

Posted 31 December 2005 - 05:56 PM

i think a can read, but i also think we missunderstand eachother. You will have a hard time with the AMD Memory controller to run ANY ram beyond 270Mhz 1:1 (at maximum multiplier) When lowering your multiplier you can go higher on your 1:1 timings since the lower MP will sort this out. It's all about memory bandwidth. Both your controller and Memoery will have a hard time processing any more data stable than let's say 7000+ in Sandra memory bandwidth benchmark. But that's just my 2cts...



thx but my sys runs 100% stable @ 278Mhz FSB 1:1 - 12h prime95 & OCCT and 5h MemTest
Only the USB sux !!!