Asrock 4CoreDual-SATA2

[jimmor]

Originally posted by: Microsystem
I set AGP voltage form AUTO to Normal. The system stability reduced and system restarted after 10 min Ortos ruining on the other hand:

E4500@vcore 1.51@3000Mhz VAGP= High: Stable
E4500@vcore 1.51@3000Mhz VAGP= Auto: Stable
E4500@vcore 1.51@3000Mhz VAGP= Normal: Not Stable

This show that Auto=High at my system setting and Increasing Vagp more than high(1.65 I think) Can help me more overclocking

But my motherboard warranty loss if I solder it. I hear anyone don it with pencil
Can anyone help me to increase my vagp by pencil mode.

The Vagp pencil mod is same for 775dual and 4coredual mobos, see link below,

http://forums.vr-zone.com/showthread.php?t=93127

 

[jimmor]

Originally posted by: Microsystem
Another question:
increasing vagp more than 1.65v can help me overclocking my Agp card (radeon 9600) more?

Increasing the mobo Vagp volts (northbridge volts) increases mobo FSB stability and overclocking. If you want to increase your vga cards overclocking then you must increase your vga cards GPU volts.

 

[Microsystem]

thanks
this means Agp volt=Northbrig volt and not Agp Graphic card volt?
why the name is AGP volt?

 

[Microsystem]

and another question
Cpu VTT volt described in this article is the same with Cpu Vcore?

 

[jimmor]

Originally posted by: Microsystem
thanks
this means Agp volt=Northbrig volt and not Agp Graphic card volt?
why the name is AGP volt?

It was introduced as part of the mobo/vga AGP communications interface design, which therefore means as far as the video communications interfaces are concerned, it's not applicable to either pci, or pci-e vga cards. However the volts you set mobo Vagp at is not same volts as at vga agp card interface. Apart from increasing Vagp (northbridge) to improve mobo fsb overclocking, you would also probably want increased Vagp volts if overclocking the apg communications speed from it's norm of 66.66MHz ---> meaning increasing the pci bus, since agp is directly derived from pci bus ---> where agp bus can typically be increased to around 80MHz without problems ----> or maybe more important, you should only ever increase the pci/agp bus speed to a level that does not interfere with the normal operation of anything else connected to the pci bus, ie, NET cards, HDD's, TV, etc ---> but that's another story !

 

[jimmor]

Originally posted by: Microsystem
and another question
Cpu VTT volt described in this article is the same with Cpu Vcore?

CPU Vtt is not same as cpu Vcore !!!

And for this asrock mobo, I recommend you don't play around with Vtt since it can easily become a cpu killer !

 

[Peav]

Originally posted by: jimmor

Originally posted by: Microsystem
and another question
Cpu VTT volt described in this article is the same with Cpu Vcore?

CPU Vtt is not same as cpu Vcore !!!

And for this asrock mobo, I recommend you don't play around with Vtt since it can easily become a cpu killer !

LOL even when I dont play with any voltages VTT is still killing my cpus.

Lost two sata2 boards and a q6600 and e6600. Both times the mosfets near the ATX Power connector are burning. So hot I've actually burned my finger on them :S

Good news is that the companies have issued me RMA numbers.

Bad news is I'll prolly get an asrock back

 

[jimmor]

Originally posted by: Peav

Originally posted by: jimmor

Originally posted by: Microsystem
and another question
Cpu VTT volt described in this article is the same with Cpu Vcore?

CPU Vtt is not same as cpu Vcore !!!

And for this asrock mobo, I recommend you don't play around with Vtt since it can easily become a cpu killer !

LOL even when I dont play with any voltages VTT is still killing my cpus.

Lost two sata2 boards and a q6600 and e6600. Both times the mosfets near the ATX Power connector are burning. So hot I've actually burned my finger on them :S

Good news is that the companies have issued me RMA numbers.

Bad news is I'll prolly get an asrock back

You are quite unique then, since I have never heard of any instances of Vtt killing a cpu outside people playing around with Vtt volts ---> typically can suddenly fail when Vtt has been increased above 1.5v ?

The Vtt MOSFET's should normally run very cool, since at default levels (1.25-1.35 range) very little power is drawn from that line. Your experience of them running "very hot" was presumably because your CPU had failed and become a shortcircuit on that line.

 

[WDR]

The odds for burning 2 boards with 2 procs are pretty slim imo.
Maybe is some issue with your specific hardware, like the power source or maybe the case
touching the mobo or something else.
What are the common parts you used on both times?

 

[Microsystem]

Originally posted by: jimmor

Originally posted by: Microsystem
and another question
Cpu VTT volt described in this article is the same with Cpu Vcore?

CPU Vtt is not same as cpu Vcore !!!

And for this asrock mobo, I recommend you don't play around with Vtt since it can easily become a cpu killer !

Thank again
No I don?t want to play with VTT
So I increase Vcore to 1.51 (that is above 1.5 mentioned in this article) I worry about it.
All things is ok but the max cpu core temp (75 C while running ORTOS and 68 while running dual FORTRAN code) worry me.
Cpu temp is less than 55 C in all cases.

 

[jimmor]

Originally posted by: Microsystem

Originally posted by: jimmor

Originally posted by: Microsystem
and another question
Cpu VTT volt described in this article is the same with Cpu Vcore?

CPU Vtt is not same as cpu Vcore !!!

And for this asrock mobo, I recommend you don't play around with Vtt since it can easily become a cpu killer !

Thank again
No I don?t want to play with VTT
So I increase Vcore to 1.51 (that is above 1.5 mentioned in this article) I worry about it.
All things is ok but the max cpu core temp (75 C while running ORTOS and 68 while running dual FORTRAN code) worry me.
Cpu temp is less than 55 C in all cases.

Vcore of 1.51v is quite high for the 3GHz overclock you previously indicated. So if intending to stick with 3GHz, setting Vcore to something like 1.41-1.45 would be more appropriate ---> and would reduce your running temps ?

Also, if you are currently still using stock intel heatsink, I suggest changing to a more efficient cpu heatsink for overclocking at 3Ghz or above levels ---> something like the cheap but excellent Arctic Freezer 7 Pro. Or for the "serious" overclocker, something like a Thermalright Ultra-120 extreme or even water cooling ?

 

[cpmee]

Originally posted by: cpmee

Originally posted by: Steffche
Well, I sent en email to Asrock complaining about how I cant my BSEL modded E2160 and my DDR550 RAM to run at 1:1 ration (and I have tried EVERYTHING...including dropping the FSB to 200Mhz)...it still wants to run at 4:3. Prior to the BSEL mod I could run 1:1...

Anyway, their response was so typical..."Sorry sir, but our BIOS doesnt support that ratio option".

I replied with a very stern email..making it quite clear that I was not impressed.

Anyway, last night I managed to get the FSB to 288Mhz, RAM still at 4:3, but timings are tight at 2.5-3-3-6, and my RAM is running at a measly 215Mhz or there abouts...

Like others have mentioned, thats the spooky automatic ratio divider kicking in at anything above about 240fsb with ddr1. Some Asrock bios programmer certainly knows the specifics of it.

DDR2 800 pc6400 at $30 for 2gb is a viable option. Get it while you can. :laugh: (Of course I been generally wrong at guessing future ram prices in the past, but I fail to see how it can get much cheaper )

I guess I was right this time. The ram I bought at $18.99 a gig is now $24.99 a gig.:Q :laugh:
http://www.newegg.com/Product/...x?Item=N82E16820211062

 

[Alex0915]

Thanks! So all I should do is set fsb to auto/266, flexibility off, ddr2-533 and the agp voltage. Thanks for the help!

 

[Peav]

In reply to WDR and Jimmor.

My X1950Pro is the only thing that got used in both.

First board failed, I replaced that, the cpu and the psu just in case. Ram at the time was ddr1.

Second time, different cpu, obviously different board. PSU was rma'd altho not found to be faulty ( different serial numbers ) DDR2 800 was used.

It is a strange one, but this board did go off for a week not long after I had it RMA'd. Was boxed up to go back and kicked back in. Put it down to a quirk but it did then fail 2 months later.

I've lost plenty of boards in my time due to mods gone wrong and I never rma them. Its a risk I take. MSI, ABit and a few Epox boards to name a few.

I'm still not 100% sure the cpu has died, I doesn't boot in my asrock 4coredual-vsta as the board is a bit weird. Sometimes fails to boot the Celey D thats in it if I remove and reseat it. Mosfets seemed to warm a little with the q6600 in it but nowhere near the temps on the sata2.
I've posted voltages before and the CPU was receiving around 1.8v on the Vtt side. Yes cpu could have failed. However there should be safeguards inplace to stop it killing the board.
Never voltmoded the boards, more so with the q6600 as I knew they don't clock at all in the sata2 board. I was contemplating the vagp one but never did. Board died before hand.

I havent seen anything like this since I bought an ECS bourd for my Celeron500 years ago.
Might just be a bad batch, it happens.
My vsta seems stable other than the quirk. My sis has a sata2 with a Pentium D chip and hers is fine.

Hopefully all sorted by the end of the week, cannot game on this celey d and x800gt card

Hmmm..
I'm wondering trying the knackered e6600 in it before I got the q6600, just to make sure it was dead, has caused something to fail with time?? Forgot about that e6600 was never rma'd due to intel not accepting my proof of purchase as valid.

 

[Alex0915]

I meant manual/266* to get normal Q6600 speeds

 

[Peav]

Originally posted by: Alex0915
I meant manual/266* to get normal Q6600 speeds

For me I set these,

CPU FSB = 266
PCI-e = 100
pci/agp = 33/66

Memory to 533Mhz
Flex = off.
All Ram timings to auto
Memory to Dual Channel ( 2 sticks )
Command Rate = 1T

AGP and DDR Volts at normal.

Then set AGP options to suit.

If set to auto on the cpu fsb, my q6600 would default to 255 FSB.

 

[Alex0915]

is pci-e 100 the same as auto? I have two monitors hooked up and for some reason my other monitor cant be detected until I get into windows. Im wondering if my pcie auto setting has anything to do with that, although it didnt happen with my other pcie card or any other video card Ive used on this board for that matter( 6600GT, 7950GT, x1950Pro).

Ok, I guess I have my board "up to spec" now. The Q6600 is running at 2,400.9MHz and the memory is running at DDR2-533 (4-4-4-12-16T), 1:1 with the cpu fsb. This increased my 3DMark06 scores to 9,300 at default and 8,600 with 4xaa/16xaf at high quality settings (averaging our three runs of each) with supersampling transp. aa.

The really big thing here is that the gpu simply doesnt overheat. My room temp is at 90F thanks to my radiator overheating my room; and the gpu starting temp was 55C but the gpu simply couldnt get past 65-67C in any of the tests, even though its factory oc-ed at 575/1700! I think I should start doing some oc-ing of it myself, to see how high I can get that kind of stability to persist.

 

[cpmee]

Originally posted by: Alex0915
is pci-e 100 the same as auto? I have two monitors hooked up and for some reason my other monitor cant be detected until I get into windows. Im wondering if my pcie auto setting has anything to do with that, although it didnt happen with my other pcie card or any other video card Ive used on this board for that matter( 6600GT, 7950GT, x1950Pro).

No, 100mhz is 100, auto is ? Its normal for the other monitor not to be detected until you get into windows on a lot of cards.

Ok, I guess I have my board "up to spec" now. The Q6600 is running at 2,400.9MHz and the memory is running at DDR2-533 (4-4-4-12-16T), 1:1 with the cpu fsb. This increased my 3DMark06 scores to 9,300 at default and 8,600 with 4xaa/16xaf at high quality settings (averaging our three runs of each) with supersampling transp. aa.

Time to get it higher.

The really big thing here is that the gpu simply doesnt overheat. My room temp is at 90F thanks to my radiator overheating my room;

Open the window ! :laugh:

Print out and try these settings next. If successful, try 280fsb next.

CPU Configuration
==================================================
CPU Host Frequency [Manual]
Actual Frequency (MHz) [277]
PCIE Clock [100 MHz]
AGP/PCI Clock [66/33 MHz]
Spread Spectrum [Disabled]
Boot Failure Guard [Disabled]

Ratio Acutal Value 9

Max CPUID Value Limit [Disabled]
No-Execute Memory Protection [Disabled]






Chipset Configuration
==================================================
DRAM Frequency [266MHz (DDRII533) ]
Flexibility Option [Disabled]
DRAM CAS# Latency [4]
DRAM Bank Interleave [4-Way]
Precharge to Active (Trp) [4T]
Active to Precharge (Tras) [12T]
Active to CMD (Trcd) [4T]
REF to ACT/REF to REF (Trfc) [28T]
ACT (0) to ACT (1) (Trrd) [2T]
Read to Precharge (Trtp) [Auto]
Write to Read CMD (Twtr) [2T]
Write Recovery Time (Twr) [4T]
DRAM Bus Selection [Dual Channel]
DRAM Command Rate [2T Command]
> Advanced Memory Configuration [Untouched]
> Advanced Host Configuration [Untouched except Pipeline DRQCTl = Enabled]

DRAM Voltage [High]
AGP Voltage [High]

Primary Graphics Adaptor [PCIE]
AGP Mode [8x]
AGP Fast Write [Enabled]
AGP Aperture Size [256MB]
AGP Staggered Delay [Auto]
AGP GADSTB Output Delay [Auto]
AGP 3.0 Calibration [Enabled]
AGP 3.0 DBI Function [Auto]
DBI Output for AGP Trans. [Enabled]
DBI Output for FRAME Trans. [Auto]

V-Link Speed [Normal]
V-Link disable [Enabled]

PCI Delay Transaction [Enabled] (Disable for Audigy Sound Card)
IDE Drive Strength [Normal]
PCIE Downstream Pipeline [Disable] (Disable helps prevent artifacting on pcie cards)
OnBoard LAN [Disabled] (Enable if using it)
OnBoard HD Audio [Enabled] (Disable if onboard sound)
Front Panel [Disabled] (Enable if using it)
CD-In [Disabled] (Enable if using it)
Echo TPR Disable [Auto]

 

[Alex0915]

Thanks, Im off to try these settings now! Oh, about the pcie setting, I see 100 and I see a sync setting as well. is it preferable to have it set at 100 over sync?

 

[cpmee]

Originally posted by: Alex0915
Thanks, Im off to try these settings now! Oh, about the pcie setting, I see 100 and I see a sync setting as well. is it preferable to have it set at 100 over sync?

For stabilty, yes.

 

[Peav]

Originally posted by: Alex0915
Thanks, Im off to try these settings now! Oh, about the pcie setting, I see 100 and I see a sync setting as well. is it preferable to have it set at 100 over sync?

Yes Sync will set it to a divider of the CPU FSB.

As you increase FSB the PCI-e bus will increase aswell. You want PCI-e to always be at 100Mhz.

Same goes with PCI-AGP. You wand pci to be at 33.3Mhz and agp at 66.6.

 

[Alex0915]

Gotchya. So with fsb at 266MHz (stock), if pci-e is kept at sync, then it will be 133 and if I start increasing the cpu fsb to oc, the pci-e bus will increase as well? Just reading up on this in a few places, and they do recommend locking it at 100 to maintain stability, my question was, when its set as a divider of the cpu fsb is it at 133MHz when the cpu fsb is 266MHz?

 

[Alex0915]

Sorry for all the questions lol. Just trying to figure out what all this stuff does. Whats the difference between command rate 2T and 1T. I have it at 2T right now. Is 2T the default for DDR2-533?

 

[Alex0915]

Ok been reading up on the 1T/2T command rate issue, and it seems as though a setting of 1T increases benchmark performance quite a bit... my question is, would it make up for my Samsung DDR2-667 running at 533? If it makes the timings tighter, it might well make it perform like faster memory!

 

[brokencase]

Originally posted by: Alex0915
Ok been reading up on the 1T/2T command rate issue, and it seems as though a setting of 1T increases benchmark performance quite a bit... my question is, would it make up for my Samsung DDR2-667 running at 533? If it makes the timings tighter, it might well make it perform like faster memory!

You can try it, but 1T more than likely will be unstable. I have DDR2-667 and I cannot run at 1T.

For this type of experimenting it is better to get the memtest bootable CD than to boot into OS/hd.

Let's you confirm memory is good before letting Windows do the fandango on core.

http://files.extremeoverclocking.com/file.php?f=60

Download it, Burn it, You'll be glad you have it going forward.