"ASUS P5Q Deluxe discussion".

[Mumid]

I have just added 2x1GB Cellshock pc8000 4-4-4-12 ram to my board for a total of 4GB to go with my Ballistix. Board booted fine straight away with 4gig mem at 600mhz 5-5-5-15 PL8 with 2.2v. So running 4gig ram 500fsb / 1200 mhz ram fine over here. I love this board! The xtra ram even boosted my bandwidth scores. Will post benchmarks when i get home from work

[RAdoxx]

What settings would be optimal for a Q9550 running at stock? Which unnecessary features can I disable etc?

[CryptiK]

PortCitySlim - That sunbeam cooler seems to perform quite well, I wish they had the TRUE fan speed stated in that comparison table though. I do like the TRUE (hence owning one) but I think either one will do you well, and outperform you current cooler.

Mumid - what vNB do you require to be linpack (Intel Burn Test) stable with 4 x 1GB?

Oh BTW - Tulu - Intel burn test is an excellent stability tester, and it only takes 8 mins to do a 5 run pass, and 15 runs shows excellent stability. However be aware that no one program can show complete stability.

**Small update - Looks like I have found stability at long last! Not completely done testing but this is a massive improvement. At my original settings, I passed only 2 out of 5 tests, but with some adjusting of vcore and ram timings, I got this result at 4GHz @ 1.32v load, DDR 1186 MHz @ 5-5-5-15 @ 2.12v vdimm (bios), PL8 @ 1.26v vNB:

[PortCitySlim]

Here goes, I know this is mainly a memory question but this is all done on the P5Q Dlx and I need to understand the Ram speed vs Everest Benchmark relationship. Here are the two stable ram tests I have done with my PC. The first one was done on my old 3.4ghz(8x425) overclock and 1133mhz Ram with Static read control, Performance Level 7/ Strong and all pull in enabled. The second one is my latest overclock at 3.52ghz(9x390) 1173mhz Ram with Static read control, Performance Level 6/ Strong and all pull ins enabled. Which of these is better performance wise? I don't understand how the faster Ram speed of 1173mhz benefits if it performs worse in Everest benchmarks.


1133mhz 1.98v PL7/Strong


1173mhz 1.96v PL6/Strong

[Zeus]

Aight so 1201 is to blame for crappy LLC?

I flashed back to 1103, Asus update wouldn't let me so i used EZ-flash with my USB stick.

First time i ever used this and it worked like a charm i have to say, 1103 seems great.

Unfortunately i couldn't overclock my cpu any further than with 1201 but at least now i could boot at 600FSB, opposed to a lousy 570 with 1201.

Tried to stabilize 600fsb but had a hard time.

1.28Vfsb, 1.58Vpll, 1.38Vnb and 0.565 for NBgtl seemed best.

After trying with memtest 86+ i booted into windows with these settings but stability was non-exsistent, although i managed to pull off a 1M still.



Now i'm gonna see if LCC works better with this bios, 1201 gave exactly the same crap Leeghoofd is talking about, rockstable at 4.3GHz with just 1.208V but bluescreens at starting or shutting down windows or just in explorer.

[CryptiK]

Here goes, I know this is mainly a memory question but this is all done on the P5Q Dlx and I need to understand the Ram speed vs Everest Benchmark relationship. Here are the two stable ram tests I have done with my PC. The first one was done on my old 3.4ghz(8x425) overclock and 1133mhz Ram with Static read control, Performance Level 7/ Strong and all pull in enabled. The second one is my latest overclock at 3.52ghz(9x390) 1173mhz Ram with Static read control, Performance Level 6/ Strong and all pull ins enabled. Which of these is better performance wise? I don't understand how the faster Ram speed of 1173mhz benefits if it performs worse in Everest benchmarks.

It's the lower FSB speed that gives you the performance deficit, even though your ram is running faster. Higher FSB & lower ram speed will outperform lower FSB & higher ram speed. Nice clocks though mate

Here's mine at 445 FSB/DDR1186/PL8



@ Zeus - nice work there too mate, max screen I got is 590 FSB but that is because I refuse to go above 1.36v on my e8400

[Polizei]

My E6600 with these settings failed Orthos Small FFT after 30 minutes. Advice?

JumperFree Configuration Settings
AI Overclock tuner: MANUAL
CPU Ratio Setting: 9
FSB Strap to North Bridge: 333
FSB Frequency: 378
PCI-E Frequency: 105
DRAM Frequency: 757
DRAM CLK Skew on Channel A1: AUTO
DRAM CLK Skew on Channel A2: AUTO
DRAM CLK Skew on Channel B1: AUTO
DRAM CLK Skew on Channel B2: AUTO
DRAM Timing Control: MANUAL

1st Information :

CAS# Latency: 5
DRAM RAS# to CAS# Delay: 5
DRAM RAS# Precharge: 5
DRAM RAS# Activate to Precharge: 15
RAS# to RAS# Delay : AUTO
Row Refresh Cycle Time: AUTO
Write Recovery Time: AUTO
Read to Precharge Time: AUTO

[2nd and 3rd information for RAM is ALL AUTO]

DRAM Static Read Control: Disabled
DRAM Read Training: AUTO
MEM. OC Charger: AUTO
AI Clock Twister: AUTO
AI Transaction Booster: AUTO

CPU Voltage: 1.5
CPU GTL Voltage Reference (0/2): .695
CPU GTL Voltage Reference (1/3): .695
CPU PLL Voltage: 1.58
FSB Termination Voltage: 1.26
DRAM Voltage: 1.96
NB Voltage: 1.2
NB GTL Reference: .665
SBridge Voltage: 1.10
PCIE SATA Voltage: 1.50

Load Line Calibration: ENABLED
CPU Spread Spectrum: Disabled
PCIE Spread Spectrum: Disabled
CPU Clock Skew : AUTO
NB Clock Skew : AUTO
[In the 1201 BIOS, I have something here set to "Optimization"]

Advance CPU Settings
CPU Ratio Setting: 9
CPU VID: Default
C1E Suppport: Disabled
Max CPUID Value Limit: Disabled
Intel® Virtualization Tech: Disabled
CPU TM Function: Enabled
Execute Disable Bit: Enabled

[SuporterPoli]

Guys i'm having a rather weird problem. One of my cores crashes in Prime95 after about 1h10-1h20minutes but the other 3 go on for hours. Each time is another core(pretty much random) but usually thread#1 and #2.
I tryed upping the vcore to 1.395 and vtt to 1.36,nb to 1.5,vpll to 1.58 just to be sure but its the same thing as with vcore 1.375,vtt 1.28,nb to 1.28 and vpll to 1.52.
My rig is this:
Q9450 L812B264
P5Q Deluxe 0803 bios(not the last version)
Thermalright Ultra120Extreme+2xScythe SFF21F(1600rpm)
Palit 8600GT+ Sonic
2x2GB Mushkin Redline PC8000
Corsair 1000HX

ASUS P5Q DLX Template :

JumperFree Configuration Settings
AI Overclock tuner: MANUAL
CPU Ratio Setting: 8
FSB Frequency: 475
PCI-E Frequency: 101
FSB Strap to North Bridge: AUTO
DRAM Frequency: AUTO
DRAM CLK Skew on Channel A1: AUTO
DRAM CLK Skew on Channel A2: AUTO
DRAM CLK Skew on Channel B1: AUTO
DRAM CLK Skew on Channel B2: AUTO
DRAM Timing Control: MANUAL

1st Information :

CAS# Latency: 5
DRAM RAS# to CAS# Delay: 5
DRAM RAS# Precharge: 5
DRAM RAS# Activate to Precharge: 15
RAS# to RAS# Delay : AUTO
Row Refresh Cycle Time: AUTO
Write Recovery Time: AUTO
Read to Precharge Time: AUTO

2nd Information :

READ to WRITE Delay (S/D): AUTO
Write to Read Delay (S): AUTO
WRITE to READ Delay (D): AUTO
READ to READ Delay (S): AUTO
READ to READ Delay (D): AUTO
WRITE to WRITE Delay (S): AUTO
WRITE to WRITE Delay (D): AUTO

3rd Information :

WRITE to PRE Delay: AUTO
READ to PRE Delay: AUTO
PRE to PRE Delay: AUTO
ALL PRE to ACT Delay: AUTO
ALL PRE to REF Delay: AUTO
DRAM Static Read Control: AUTO
DRAM Read Training: AUTO
MEM. OC Charger: AUTO
AI Clock Twister: AUTO
AI Transaction Booster: AUTO

CPU Voltage: 1.395
CPU GTL Voltage Reference (0/2): AUTO
CPU GTL Voltage Reference (1/3): AUTO
CPU PLL Voltage: 1.58
FSB Termination Voltage: 1.36
DRAM Voltage: 1.98
NB Voltage: 1.5
NB GTL Reference: AUTO
SBridge Voltage: 1.2
PCIE SATA Voltage: 1.5

Load Line Calibration: Enabled
CPU Spread Spectrum: Disabled
PCIE Spread Spectrum: Disabled
CPU Clock Skew : AUTO
NB Clock Skew : AUTO

Advance CPU Settings
CPU Ratio Setting: Manual
CPU VID: Default
C1E Suppport: Disabled
Max CPUID Value Limit: Disabled
Intel® Virtualization Tech: Disabled
Vanderpool Technology: Disabled
CPU TM Function: Disabled
Execute Disable Bit: Disabled

Can anyone help on this?

[davidoff619]

help please
Can get stable at 1.45 for hours on my qx9650 but 1.42.5 fails after 130mins both of these setting don't seem very responsive I'm using 400x10 but felt 465x8.5 was far better but didn't play around with voltages to get stable. Plus with 400x10 at 1.45.5 it gets as hot as 77 degrees is this safe could the responsiveness be due to the ram being sets at 800 rather than at 900 with 465x8.5. I just want to get my rig to 4ghz and Stable so I can start my gaming and stuff please help me as I am in need to get this sorted thanks David

[NBF]

I have this LLC issue with other BIOSes, not only 1201. In fact I had that with other ASUS boards too.

After finding lowest stable load Vcore with LLC enabled, I choose one of the 2 aproaches to deal with light task/idle BSOD.

- Raise Vcore (usually 2 notches)
- Enable C1E and let Multi on Auto

The second option is what I set for daily usage, that way I can keep load Vcore as low as possible and have idle stability, I also have lower idle temps and power consumption.

[Mumid]

Mumid - what vNB do you require to be linpack (Intel Burn Test) stable with 4 x 1GB?

Not tried it mate, only had half an hour with my new mem as i had to go to work, but will have a look at linpack later.

p.s glad you are making some headway with your ram clocking attempts, really weird that you can post at 1199 and 1277 but nothing in between.

I have this LLC issue with other BIOSes, not only 1201. In fact I had that with other ASUS boards too.

After finding lowest stable load Vcore with LLC enabled, I choose one of the 2 aproaches to deal with light task/idle BSOD.

- Raise Vcore (usually 2 notches)
- Enable C1E and let Multi on Auto

The second option is what I set for daily usage, that way I can keep load Vcore as low as possible and have idle stability, I also have lower idle temps and power consumption.

But then you would have to use default multi wouldnt you?

[NBF]

But then you would have to use default multi wouldnt you?

Yeap, default multi. But you can always have a more agressive OC profile saved in BIOS for the hard tasks

[truehighroller]

My E6600 with these settings failed Orthos Small FFT after 30 minutes. Advice?

JumperFree Configuration Settings
AI Overclock tuner: MANUAL
CPU Ratio Setting: 9
FSB Strap to North Bridge: 333
FSB Frequency: 378
PCI-E Frequency: 105
DRAM Frequency: 757
DRAM CLK Skew on Channel A1: AUTO
DRAM CLK Skew on Channel A2: AUTO
DRAM CLK Skew on Channel B1: AUTO
DRAM CLK Skew on Channel B2: AUTO
DRAM Timing Control: MANUAL

1st Information :

CAS# Latency: 5
DRAM RAS# to CAS# Delay: 5
DRAM RAS# Precharge: 5
DRAM RAS# Activate to Precharge: 15
RAS# to RAS# Delay : AUTO
Row Refresh Cycle Time: AUTO
Write Recovery Time: AUTO
Read to Precharge Time: AUTO

[2nd and 3rd information for RAM is ALL AUTO]

DRAM Static Read Control: Disabled
DRAM Read Training: AUTO "make this disabled"
MEM. OC Charger: AUTO "make this enabled"
AI Clock Twister: AUTO "make this lighter until you get stable"
AI Transaction Booster: AUTO "you need to controll this your self"

CPU Voltage: 1.5
CPU GTL Voltage Reference (0/2): .695
CPU GTL Voltage Reference (1/3): .695
CPU PLL Voltage: 1.58
FSB Termination Voltage: 1.26
DRAM Voltage: 1.96
NB Voltage: 1.2 "this needs to be 1.26 probably should be good"
NB GTL Reference: .665 "leave this at auto"
SBridge Voltage: 1.10 "set this to 1.2"
PCIE SATA Voltage: 1.50 "set this to 1.6"

Load Line Calibration: ENABLED
CPU Spread Spectrum: Disabled
PCIE Spread Spectrum: Disabled
CPU Clock Skew : AUTO
NB Clock Skew : AUTO
[In the 1201 BIOS, I have something here set to "Optimization"] 
Advance CPU Settings
CPU Ratio Setting: 9
CPU VID: Default
C1E Suppport: Disabled
Max CPUID Value Limit: Disabled
Intel® Virtualization Tech: Disabled
CPU TM Function: Enabled "disable this"
Execute Disable Bit: Enabled "disable this"

Try what I put in quotes.

[PortCitySlim]

It's the lower FSB speed that gives you the performance deficit, even though your ram is running faster. Higher FSB & lower ram speed will outperform lower FSB & higher ram speed. Nice clocks though mate

Ok, I understand now. I can run 3.52ghz(440x8) and 1103mhz but the problem I was having is that I could not get it stable with Static read control. My everest benchmarks were like 9100/8000/8800/56.9ns so I thought that the latency increase was worse then the hit I was taking in the bandwith area on my write/copy mb/s. So if I can run at a higher FSB even though my latency is worse but bandwith is better am I getting better performance even though I can not run Static read control? I dropped to PL9 and enabled SRC, then my Everest benchmarks were even worse. Maybe better put this way, which is better...

3.51ghz (390x9) 1173mhz PL6/Strong, SRC enabled
9119/7118/7973/54.5

or

3.52ghz (440x8) 1103mhz PL8/Moderate, SRC Disabled
9100/8000/8800/56.9


Thanks

[Tulû]

Oh BTW - Tulu - Intel burn test is an excellent stability tester, and it only takes 8 mins to do a 5 run pass, and 15 runs shows excellent stability. However be aware that no one program can show complete stability.

Thank you for your advice.
That Linpack made my cores 5C hotter than prime .(small FFTs)
Ran 5 times and said " passed 5 out of 5 (100%) "
Thanks again.

[truehighroller]

Thank you for your advice.
That Linpack made my cores 5C hotter than prime .(small FFTs)
Ran 5 times and said " passed 5 out of 5 (100%) "
Thanks again.

Yah I think it is the best IMO.

[CryptiK]

Not tried it mate, only had half an hour with my new mem as i had to go to work, but will have a look at linpack later.

p.s glad you are making some headway with your ram clocking attempts, really weird that you can post at 1199 and 1277 but nothing in between.

No problem, just interested to see what it takes for 4 x 1GB vs 2 x 1GB. And thanks, my issue certainly is strange. I even can post at up to and beyond DDR 1333 MHz on 2.24v bios. Damned if i can figure out why 1200 - 1277 is a black hole causing no post.

Ok, I understand now. I can run 3.52ghz(440x8) and 1103mhz but the problem I was having is that I could not get it stable with Static read control. My everest benchmarks were like 9100/8000/8800/56.9ns so I thought that the latency increase was worse then the hit I was taking in the bandwith area on my write/copy mb/s. So if I can run at a higher FSB even though my latency is worse but bandwith is better am I getting better performance even though I can not run Static read control? I dropped to PL9 and enabled SRC, then my Everest benchmarks were even worse. Maybe better put this way, which is better...

3.51ghz (390x9) 1173mhz PL6/Strong, SRC enabled
9119/7118/7973/54.5

or

3.52ghz (440x8) 1103mhz PL8/Moderate, SRC Disabled
9100/8000/8800/56.9 Thanks

I'd be leaning towards the second config with more bandwidth but lower latency, but to test it, use 3D benches and superpi That should give you an indication what works best.

[Polizei]

Try what I put in quotes.

Thanks... I'll give that a go once I get off work.

[truehighroller]

Thanks... I'll give that a go once I get off work.

Those are the main things I noticed. You can still prob tweak your voltages for gtl ref and cpu voltage and the other one the cpu pll voltage but, you need to def do what I quoted IMO. You might want to bump your nb volts up to 1.28 if your are more stable but not stable all the way still with 1.26. I had to bump mine up to 1.28 to get 100% intelburntest stable.

[CryptiK]

Yeah!

Still perfecting things but I have made a lot of progress tonight:

[Polizei]

Those are the main things I noticed. You can still prob tweak your voltages for gtl ref and cpu voltage and the other one the cpu pll voltage but, you need to def do what I quoted IMO. You might want to bump your nb volts up to 1.28 if your are more stable but not stable all the way still with 1.26. I had to bump mine up to 1.28 to get 100% intelburntest stable.

I know I have 65nm, but I didn't know how high I could bump up my GTLref voltages or anything before hurting it. How high can I go?

[Leeghoofd]

Aight so 1201 is to blame for crappy LLC?

I flashed back to 1103, Asus update wouldn't let me so i used EZ-flash with my USB stick.Tried to stabilize 600fsb but had a hard time.
1.28Vfsb, 1.58Vpll, 1.38Vnb and 0.565 for NBgtl seemed best.
After trying with memtest 86+ i booted into windows with these settings but stability was non-exsistent, although i managed to pull off a 1M still.
Now i'm gonna see if LCC works better with this bios, 1201 gave exactly the same crap Leeghoofd is talking about, rockstable at 4.3GHz with just 1.208V but bluescreens at starting or shutting down windows or just in explorer.

Now let's see a 32mb Pi master of the seas, I know you can do it...

[truehighroller]

I know I have 65nm, but I didn't know how high I could bump up my GTLref voltages or anything before hurting it. How high can I go?

I will try to see real quick for you.

[CryptiK]

Not sure on GTL voltage upper safe limits, I started a thread about this but got no definitive answer other than each processor class prefers a certain value (none were as high as your 0.695x though). Have you tried all values from 0.61x to 0.695x? There is apparently a 'sweet spot' that takes trial and error to find.

[Pntgrd]

I seem to be seeing a trend as to the 1103 bios being better for most than 1201. Is this correct?