Biostar Tpower I45 , +600mhz FSB easy

[cstkl1]

This is a reply from biostar

i find it as a big issue but .. they practically said not really for them..
well up to u guys..

email to biostar fae carter@biostar.com.tw

my question

1. FSB voltage set in bios will never be lower than NB voltage. Example

Set NB to +0.1
Set FSB voltage to +0.025
Will result in

NB voltage 1.196
FSB voltage to 1.216.. ( see the issue here) Suppose to be 1.12 and not
1.216

only can be adjusted via TPower 2 utility


His reply

after whole tested, we check that
the system is more stable and tunable under FSB voltage higher than NB voltage
that the reason FSB Voltage is always higher than NB Voltage.

so whats the point even in having a voltage adjustment in the fsb.
since its always going to be sky high

also this is not possible then since its ideal for 500fsb c2d




the reason i posted the everest score..

biostar has this function called ddr2 rcomp configuration
how it does it what i am going to tell i have no clue
but it will reduce read and copy of the above reading to 7k/5k with latencies to 60+.. thats how u can clock up to 600fsb easy...with a tRD readings very low..
so this board on high fsb clocks.. has no gain in mem.. for my testing compared to asus p5q which has higher tRD's but atleast respectable bandwidth scores..

i rate this board 5/5 for cpuz posting of 600fsb
other than that its rubbish.. compared to asus p5q

[VIctorj]

Is it me, or Biostar is going the Asus way with these tools and layouts?

[rusty]

This is a reply from biostar

i find it as a big issue but .. they practically said not really for them..
well up to u guys..

email to biostar fae carter@biostar.com.tw

my question

so whats the point even in having a voltage adjustment in the fsb.
since its always going to be sky high

also this is not possible then since its ideal for 500fsb c2d




the reason i posted the everest score..

biostar has this function called ddr2 rcomp configuration
how it does it what i am going to tell i have no clue
but it will reduce read and copy of the above reading to 7k/5k with latencies to 60+.. thats how u can clock up to 600fsb easy...with a tRD readings very low..
so this board on high fsb clocks.. has no gain in mem.. for my testing compared to asus p5q which has higher tRD's but atleast respectable bandwidth scores..

i rate this board 5/5 for cpuz posting of 600fsb
other than that its rubbish.. compared to asus p5q

do you mean p5q or p5q3?

[cstkl1]

biostar solved a lot of issues in the last two bioses..

way better
now only with disabling the ddr2 performance thingy ure rams performance suffer

other than that up to 600fsb stable is possible but depending on what trd performance u want to aim at...

biostar awesome

http://forums.vr-zone.com/showthread.php?t=292119

[mop231]

this board is ddr3 or ddr2 different sources different versions

[mrcape]

DDR2, but I hope they make a DDR3 version.

[mop231]

or a combo ddr2+ddr3 2+4 slots or somethin

[dinos22]

i am testing this Biostar P45 board as well atm

so far it was really really buggy with first release bios and after flashing to a new one that one of the beta testers shared on XS here (Pt1t)-(bios 618) i must say it has improved a lot (it took me about an hour to get my head around the bios)......i was doing some 590MHz*8 (4720MHz) 32M SuperPi runs with 1.59v on air cooling >>>> this new E8400 CPU is a monster >> efficiency wasn't the best but it wasn't completely terrible. I think i managed to reduce PL in windows to 9 from memory and timings were 5-4-4-9-5-30. final time was 9m53s. i was also booting at 4.8GHz but could not complete any runs

I do have a warning for anyone thinking of going to Biostar for high FSB clocking and besides some bugs like being forced to run auto RAM timings etc etc, the way they are achieving the FSB is that they actually tie vFSB (FSB volts) with vNB (NB volts) so when you start pushing to +0.4 or more on MCH your vFSB is starting to reach dangerous levels which according to Intel will kill your CPU very quick. Biostar hardware health reporting software was showing the following:

• Bios set NB +0.35 >> Actual NB voltage reported: 1.37v
• +0.35 on NB makes vFSB jump to 1.42v (setting any value in bios does NOT change vFSB)
• Bios set NB +0.4 >> Actual NB voltage reported: 1.42v
• +0.4 on NB makes vFSB jump to 1.47v (setting any value in bios does NOT change vFSB)
• Next Voltage setting in bios for NB is +0.5 >> Actual NB voltage reported: 1.47v
• +0.5 on NB makes vFSB jump to 1.52v (setting any value in bios does NOT change vFSB)

Intel says that running 1.5v on vFSB will kill your CPU very quick!!! I haven't had any boards yet that gave me an option to run more than 1.45v from memory and i have run 1.45v for benching on most of these boards without any issues but who knows.........this was recently reported on a website as well and it's now got me spooked as i currently have a CPU going back to Intel for warranty (first time i had a dead CPU ever)

youngpro says he used +0.7 which will result in vFSB at around 1.72v EEEEEEEEEEEK >>i donno how long that chip will last but if all these sources are correct not very long at all..........

Looks like engineers from Biostar are aware of this fact but i disagree that fixing vFSB to vNB is the only way to get high FSB as other boards have shown. I think is a bad position to take if vFSB does in fact pose such danger..........we will however find out what sort of impact it is having in future when more and more overclockers try these out and particularly the ones going for extreme clocks and high volts.

i feel that they should let us decide what volts to set rather than have it fixed....

my bad luck continues >> one of the sticks from my BEST dominator kit is dead

[cstkl1]

i am testing this Biostar P45 board as well atm

so far it was really really buggy with first release bios and after flashing to a new one that one of the beta testers shared on XS here (Pt1t)-(bios 618) i must say it has improved a lot (it took me about an hour to get my head around the bios)......i was doing some 590MHz*8 (4720MHz) 32M SuperPi runs with 1.59v on air cooling >>>> this new E8400 CPU is a monster >> efficiency wasn't the best but it wasn't completely terrible. I think i managed to reduce PL in windows to 9 from memory and timings were 5-4-4-9-5-30. final time was 9m53s. i was also booting at 4.8GHz but could not complete any runs

I do have a warning for anyone thinking of going to Biostar for high FSB clocking and besides some bugs like being forced to run auto RAM timings etc etc, the way they are achieving the FSB is that they actually tie vFSB (FSB volts) with vNB (NB volts) so when you start pushing to +0.4 or more on MCH your vFSB is starting to reach dangerous levels which according to Intel will kill your CPU very quick. Biostar hardware health reporting software was showing the following:

• Bios set NB +0.35 >> Actual NB voltage reported: 1.37v
• +0.35 on NB makes vFSB jump to 1.42v (setting any value in bios does NOT change vFSB)
• Bios set NB +0.4 >> Actual NB voltage reported: 1.42v
• +0.4 on NB makes vFSB jump to 1.47v (setting any value in bios does NOT change vFSB)
• Next Voltage setting in bios for NB is +0.5 >> Actual NB voltage reported: 1.47v
• +0.5 on NB makes vFSB jump to 1.52v (setting any value in bios does NOT change vFSB)

Intel says that running 1.5v on vFSB will kill your CPU very quick!!! I haven't had any boards yet that gave me an option to run more than 1.45v from memory and i have run 1.45v for benching on most of these boards without any issues but who knows.........this was recently reported on a website as well and it's now got me spooked as i currently have a CPU going back to Intel for warranty (first time i had a dead CPU ever)

youngpro says he used +0.7 which will result in vFSB at around 1.72v EEEEEEEEEEEK >>i donno how long that chip will last but if all these sources are correct not very long at all..........

Looks like engineers from Biostar are aware of this fact but i disagree that fixing vFSB to vNB is the only way to get high FSB as other boards have shown. I think is a bad position to take if vFSB does in fact pose such danger..........we will however find out what sort of impact it is having in future when more and more overclockers try these out and particularly the ones going for extreme clocks and high volts.

i feel that they should let us decide what volts to set rather than have it fixed....

my bad luck continues >> one of the sticks from my BEST dominator kit is dead

eh .. try 0619

dinos for 0619..
not sure whether u noticed this

on high fsb on ddr2 configuration 3
try messing with
write to precharge delay and read to precharge delay..
thats how i got a lot of fsb stable at acceptable tRD's and voltages

example for 566 fsb tREAD +1... if u use 13/10 14/11 , 15/12 ( ps u cant go lower than 13/10.. eventhough the options allow u to)
u end up with pull tRD from 11 to 10 aka channel A 10, channel B 11
but if u use 13/12 or 14/12 u end up tRD 10 on both channels..

same case as 533 fsb to get tRD 9/10 stable is difficult
so use the pull from channel A, B by changing

this is what i noticed on the p45's u can have rank delays 4 stable all the way to 600fsb.. which was impossible for x48/x38
i think this is to compensate for the higher tRD

just my thoughts.. on the biostar 0619 bios
everytime i hit a fsb wall .. i find changing those two delays solved the issue
usually the problem wont be bsod.. graphic card goes haywire..

[dinos22]

what is different with 0619?

[cstkl1]

what is different with 0619?

i think its the same...just date mod difference but not too sure..
but the gpu issue seems to be less

coolice is the one that gave me the link from that chinese website

taken from here
http://bbs.ocer.net/thread-219119-1-1.html
http://rapidshare.com/files/12393807...9.BSP.rar.html

so far i can say for fsb 533-566 tREAD needs to be +1 to maintain a good ratio increment from vNB
one step lower will result a big jump in vNB to get it stable...
all this with 4 sticks mind u.. kekeke...

[Nosfer@tu]

CPU Configuration section of BIOS looks very well done... Prefetch, etc.



PS: (0280) Frequency to (0286) Ratio Actual Value aren't set to Refresh access by default; I did it.

I Understand Im looking at bios programming or ? Ca you please explain a bit more before ?
Thx in advance

It did 515mhz fsb on my qx , fsb wall of the cpu

515 is the same for Quard in anothe thread.

We did 612 fsb with air same as windwithme and 633 fsb as max superpi stable, windwithme did 635.

I think these are quite close, maybe somthing common ?

[Unhooked]

up up