UNOFFICIAL BIOSTAR TPower I45 M/B

[PolRoger]

My first sub 10 second Super Pi 1M.

I used my new Kingston HyperX PC2-9200 2X1GB (spec 2.30-2.35v). I've been having some trouble getting them to run at high speeds (so far just DDR2-1098) but they hold tight timings at 500fsb 1/1 ratio. This was with RCOMP Config. 3

[_G_]

Got this board up and running and im liking it alot
The best thing is that Ghost works perfectly with this board, very usefull when you currupt the os lol.

[Ace-a-Rue]

My first sub 10 second Super Pi 1M.

I used my new Kingston HyperX PC2-9200 2X1GB (spec 2.30-2.35v). I've been having some trouble getting them to run at high speeds (so far just DDR2-1098) but they hold tight timings at 500fsb 1/1 ratio. This was with RCOMP Config. 3

that is wonderful!

now, where is your Super PI 32M?

[Mortis360]

I'm trying to get my hands on a Q9550 with E0 stepping right now!
Friend of me managed 471*8,5, resulting in 4003mhz on his Q9550, this is stepping C1, and E0 is even better!

~ Mortis360

[Rainmaker]

Hi guys,

I'm in the UK and bought this board last week Although I've been into PCs, games, Linux etc for years and am the unofficial Windows support guy for everyone I know, this was my first actual build. I've learnt a lot about hardware this year, and this is my first overclock

Specs are:

Biostar TPower I45, E8600 (Q820A670), 2x2GB OCZ Reaper HPC PC26400, GeForce 8600GT, 500GB Seagate Barracuda 7200.11. Cooled on air by OCZ Vendetta 2 (120mm) and running on an Akasa 500W Paxpower PSU.

Having set the RAM in the BIOS to its default 2.1v and set the timings slightly slack (5,5,5,15 instead of 4,4,4,15) I booted to Windows. Pushing the FSB alone got me to 10x420 and with +0.125 on the CPU volts to be 'safe' I was cruising at 4.2Ghz. This passed 10.5 hours of prime (blend) but failed prime (small ffts) after about half an hour Here I am 10.5 hours into prime blend:



No matter what I tried I couldn't get it past 10x420 without getting BSODs or failing prime within 60 seconds flat. I was advised on a UK forum to go back to stock, and push each bit of hardware separately so I understood my components more. With most settings set to auto, on a low multi (6x) I got 500FSB before Windows (Vista Ultimate) failed to boot. Then on 10x multi but underclocking by FSB (10x275 for example) I had my RAM on an upwards divider running 550Mhz (1100Mhz DDR2) so I know my RAM can clock well also.

So, I'm stuck Basically it feels like this board has a lot to offer, but being such an OC n00b I'm holding myself back! TBH I'd have been better off (in the short term) buying an Asus P5Q - everyone and his dog has those and there is a ton of easy "Here's the settings - go get a 4.5Ghz stable oc" posts around. But then, I'm more technically minded than that and don't want an oc just "giving" to me; I want to learn and earn mine!

I have spent a week reading, and I know the chipset volts etc will need playing around with. But as it's my first OC and I'm the only guy I know in the UK with a built Biostar rig, help hasn't been easy to find. I did find some settings posted by a guy on a forum, and they got me to 4.5Ghz (10x450) after setting the Northbridge latch to 1333 and upping the FSB and chipset volts by 0.3. I could cruise in Windows all day long at low temps, watch video, play games, etc, but it fails prime in about 30 seconds on core 2



I'm sorry this is long, I've left out my detailed BIOS settings on purpose for now until someone decides to take pity on me and help. I didn't want my first post on XS to be a great big essay... Too late?

Thanks in advance, great work so far Ace... this thread is awesome! Also, a huge thanks to Fugger
Cheers,

Lee

[DoomSayer]

Can you post your bios settings ?

Also on a side note.
Whenever i restart my pc instead of doing a shutdown i get the thing where the mobo starts the pc 3 times and fails.
Then all i got to do is save and it boots fine again.
Any solution to this ?

[Rainmaker]

For the 4.2Ghz as above (10.5hrs prime blend stable, fails prime small ffts):

Intel SpeedStep: Disabled
Ratio CMOS Setting: x10
CPU Frequency Setting: 420
Over Clock Retry Count: 2
FSB (Bsel) to Northbridge Latch: Auto
DRAM Frequency: DDR2 840Mhz
DDR2 Enhance Mode: Disabled

DDR2 RCOMP Configuration: Auto Config
Static tREAD Value: Auto
Configure DRAM Timing by SPD: Disabled (4,4,4,15)

PCIE Frequency Setting: 100
CPU Clock Skew Control: Auto
(G)MCH Clock Skew Control: Auto
CPU Clock Driving Control: Auto

CPU Voltage: +0.125 (VID is 1.25, this gives 1.328 under load in Windows)
DDR Voltage: +0.300 (giving the 2.14v for my RAM which is between the rated 2.1 to 2.2v)
Chipset Voltage: Auto
FSB Voltage: Auto
CPU Core1 GTL Ref Voltage: 0.63*VTT
CPU Core2 GTL Ref Voltage: 0.67*VTT
MCH GTL Ref Voltage: Auto
CPU PLL Voltage: 1.5v

For the 4.5Ghz as above (Windows stable, passes wprime stability, fails Orthos/prime95 in seconds, fails Intel Burn Test with a BSOD)
Intel SpeedStep: Disabled
Ratio CMOS Setting: x10
CPU Frequency Setting: 450
Over Clock Retry Count: 2
FSB (Bsel) to Northbridge Latch: FSB 1333
DRAM Frequency: DDR2 900Mhz
DDR2 Enhance Mode: Disabled

DDR2 RCOMP Configuration: Configuration 2
Static tREAD Value: 10T
Configure DRAM Timing by SPD: Disabled (5,5,5,18)

PCIE Frequency Setting: 100
CPU Clock Skew Control: 200ps
(G)MCH Clock Skew Control: Auto
CPU Clock Driving Control: Auto

CPU Voltage: +0.2 (VID is 1.25, this gives 1.4 under load in Windows - too high for 45nm??)
DDR Voltage: +0.300 (giving the 2.14v for my RAM which is between the rated 2.1 to 2.2v)
Chipset Voltage: +0.275
FSB Voltage: +0.325
CPU Core1 GTL Ref Voltage: 0.63*VTT
CPU Core2 GTL Ref Voltage: 0.63*VTT
MCH GTL Ref Voltage: Auto
CPU PLL Voltage: 1.5v

I hope this is what you required, thanks in advance for your time and efforts

[Ace-a-Rue]

Can you post your bios settings ?

Also on a side note.
Whenever i restart my pc instead of doing a shutdown i get the thing where the mobo starts the pc 3 times and fails.
Then all i got to do is save and it boots fine again.
Any solution to this ?

i had the same and it is either your voltage(s) is low or could be you CPU GTL REF or MCH GTL REF...try raising your vcore first and then see if you can duplicate the problem....if that does not correct it try another voltage and setting...the whole idea is do only ONE change at a time....if you do multiple changes at one time, you will never know what is causing that.

[Ace-a-Rue]

For the 4.2Ghz as above (10.5hrs prime blend stable, fails prime small ffts):

Intel SpeedStep: Disabled
Ratio CMOS Setting: x10
CPU Frequency Setting: 420
Over Clock Retry Count: 2
FSB (Bsel) to Northbridge Latch: Auto
DRAM Frequency: DDR2 840Mhz
DDR2 Enhance Mode: Disabled

DDR2 RCOMP Configuration: Auto Config>>>>try 1-4
Static tREAD Value: Auto>>>>>use 9 or 10, possibly 11 stay away from Auto
Configure DRAM Timing by SPD: Disabled (4,4,4,15)>>>>loosen these up

PCIE Frequency Setting: 100
CPU Clock Skew Control: Auto>>>>100
(G)MCH Clock Skew Control: Auto>>>>300
CPU Clock Driving Control: Auto>>>>>900 or 1000

CPU Voltage: +0.125 (VID is 1.25, this gives 1.328 under load in Windows)>>>>higher Vcore
DDR Voltage: +0.300 (giving the 2.14v for my RAM which is between the rated 2.1 to 2.2v)
Chipset Voltage: Auto>>>>start with +0.30 add and then go higher
FSB Voltage: Auto>>>>actual voltage should be around 1.3v
CPU Core1 GTL Ref Voltage: 0.63*VTT
CPU Core2 GTL Ref Voltage: 0.67*VTT>>>>>63%
MCH GTL Ref Voltage: Auto>>>>>64%
CPU PLL Voltage: 1.5v

For the 4.5Ghz as above (Windows stable, passes wprime stability, fails Orthos/prime95 in seconds, fails Intel Burn Test with a BSOD)
Intel SpeedStep: Disabled
Ratio CMOS Setting: x10
CPU Frequency Setting: 450
Over Clock Retry Count: 2
FSB (Bsel) to Northbridge Latch: FSB 1333
DRAM Frequency: DDR2 900Mhz
DDR2 Enhance Mode: Disabled

DDR2 RCOMP Configuration: Configuration 2
Static tREAD Value: 10T
Configure DRAM Timing by SPD: Disabled (5,5,5,18)>>>>good but what is your tRAS...how about showing all your memory settings

PCIE Frequency Setting: 100
CPU Clock Skew Control: 200ps>>>>100 or 200
(G)MCH Clock Skew Control: Auto>>>>300 or 400
CPU Clock Driving Control: Auto

CPU Voltage: +0.2 (VID is 1.25, this gives 1.4 under load in Windows - too high for 45nm??)>>>>not necessairly
DDR Voltage: +0.300 (giving the 2.14v for my RAM which is between the rated 2.1 to 2.2v)
Chipset Voltage: +0.275>>>>higher
FSB Voltage: +0.325
CPU Core1 GTL Ref Voltage: 0.63*VTT
CPU Core2 GTL Ref Voltage: 0.63*VTT
MCH GTL Ref Voltage: Auto>>>>>64%
CPU PLL Voltage: 1.5v

I hope this is what you required, thanks in advance for your time and efforts

put MCH GTL REF to 64%...i believe that will go a long way to stabilize the system.

don't be afraid of raising the voltages..the whole idea is find a stable overclock and once you find that, you can try to reduce ONE voltage at a time to see if it gives you the same results.

other recommendations in red within your quote.

thanks for the pat on the back; very much appreciated.

[Rainmaker]

Wow. I'm seriously impressed, and I genuinely mean that. Thank you very much indeed for the input Ace 5* thread rating from me I'll let you know how I get on, thanks again.

[Ace-a-Rue]

Wow. I'm seriously impressed, and I genuinely mean that. Thank you very much indeed for the input Ace 5* thread rating from me I'll let you know how I get on, thanks again.

you can also raise CPU PLL to either 1.6 or 1.7v

good luck!

[Rainmaker]

Again, Ace, thank you. Had you been in the UK you'd have been drinking beer on me tonight! I'm currently 25 mins into Orthos (small FFTs) at 4.4Ghz. 4.5 fell over even under fairly high CPU/chipset/fsb voltage (vcore = 1.49 under load in Windows) and even then the temps were hitting 70oC - too high.

So I'm hoping with tweakage I should at least get 4.45Ghz stable even once the CPU vcore volts come down to more sensible levels. You've helped me over a huge hurdle, and I tip my hat to you. Before now I was thinking maybe my RAM was too low spec (should'a got the 1066 I thought) but now it seems I just needed someone clever enough to tweak the I45 itself

I will, of course, report back tomorrow with some nice long stability shots for you provided all goes well

[Ace-a-Rue]

keep chug-a-lugging!

i've been at this board long enough to know where to tweak...some of my tips have come from other posters in this thread...so...it is not all my brain-storming

[scorpion]

ace what departmemt did you find the fiber washer in home depo
.

[scorpion]

the dummy here didnot know what a fiber washer was.

[Ace-a-Rue]

the dummy here didnot know what a fiber washer was.

it is in the row that has the pre-package nuts, screws and bolts...in that row or isle, they also have the drawers of hardware parts...i got the fiber washer out of one of the drawers.

[scorpion]

thanks ace!!!!

[Rainmaker]

Hm, either I have the worst E8600 to date or I didn't set it up right yet. To summarise, Ace's settings have helped me overcome a huge hurdle so far. I couldn't boot past 10x425 without huge issues. With the new settings as above I can run much better. I noticed though that even now, 10x425 boots fine as does 10x435 but 10x430 plain refuses to even post! Is this what I've read to be called 'FSB hole'?? Anyway, I'm priming away (almost 2 hours now) on Orthos with small FFTs. Great so far. The ONLY issue (and it's a huge one for me) is that the CPU volts required are 1.44 in Windows under load



There's only one guy in the UK on our database so far with volts even nearly that high, and he's running at 5Ghz (yes 5Ghz) never mind 4.3!! Did I break it? LOL Temps are still OK though, surprisingly. I have a lot of chipset and FSB volts at the mo (I can't check without killing Orthos before the run finishes but they're on the last 'yellow' setting before they go 'red' if that helps lol).

Argh, just my luck to save for two months and end up with the worst chip in the UK for overclocking That or I'm missing something obvious in the mobo config? I'll post settings when I've got an 8 hour Orthos screenshot.

Could incorrect/low mobo settings somehow be counter-acted by a high CPU voltage? I.E. maybe it doesn't NEED these huge volts if I set the mobo up differently? Or do I definitely just have a cruddy clocker? Shame though as I'm 24 hour "Windows Vista stable" at 4.6Ghz on 1.39 volts. Puzzling tbh, but then I'm new

[freaksavior]

can i get some help?

i did the setting you did ace for the QX9650, and it didn't work for my 9550, its ok though, i palyed with it some more and i got 3.91 but had to push about 1.5 v, wich i know for a 45nm is to much.

I need help getting this thing to 4.0, i know not all chips will do 4, but i am hoping mine will with 1.4ish volts.

[Ace-a-Rue]

can i get some help?

i did the setting you did ace for the QX9650, and it didn't work for my 9550, its ok though, i palyed with it some more and i got 3.91 but had to push about 1.5 v, wich i know for a 45nm is to much.

I need help getting this thing to 4.0, i know not all chips will do 4, but i am hoping mine will with 1.4ish volts.

think about it realistically...you have a 1100 mhz boost, right?...i'd say you are doing good...my 9650 makes 1000 mhz fairly easy...but as soon as i try to go higher then it starts breaking down.

[Ace-a-Rue]

for all those guys worrying about overdoing your voltages on your 45nm cpu's you need to read this thread...very interesting study going on right now:

http://www.xtremesystems.org/forums/...d.php?t=198614

[makaka]

Just for reference i was too lazy to RMA the board after being dead with FF code , Now already 2 week passed yesterday , i powered it up and the motherboard worked again , so go figure ,
CMOS and battery removal for one week do not help , but letting the motherboard for one week without any try result in a new system alive .

[Ace-a-Rue]

strange things happen when you totally leave it alone

[PolRoger]

that is wonderful!

now, where is your Super PI 32M?

Well Ace-

I tried to go back and use those same settings and now my overclock is failing. That was the first time I used +0.250 on my cpu vcore and I was able to get into windows so I ran a quick 1M and I got out. I'm still doing some limited exploring of stability around 1.45v but I don't think it is going to be 4.75Ghz.

P.S. Thanks for the link to BenchZowner's 45nm high voltage experiment.

[Solarfall]

hey guys any tips on getting over 600fsb with this board.. i've tested dozen different settings and best i've been able to get was 590 (i know for a fact that my E8600 can do over 600) there has to be something im overlooking here