"ASUS P5Q Deluxe discussion".

[Dalten]

I'm reluctant to go back to 0803 because i've been priming for 24+ hours with the setting at the default and so far no issues...

[BeastNotro]

Ok so now it brings me to maybe my Q6600 got screwed and if so then board may still be screwed and if so then my E8600 (i hope she is a good one) maybe become screwed and that would suck.. really really bad.

I am still not 100% sure what my issue is with the Q6600 as dropping 200mhz in 2-3 weeks of running is something i never seen on any CPU i ever had Stable before.... then again were only talking about Linpack which it was stable at 3.6ghz NP's runnin g linpack 1.4 many times as i describe previously and now i am at 3.4ghz to get it linpack 1.4 stable, so either CPU degraded and/or the max is 3.6ghz and suddenly after being broken in for 2-3 weeks .. it decided to give up on that ghz for linpack 1.4 stable.

Or its the mobo controller that is causing this wild degration and if so then i am worried about any CPU i put in it whether Quad or not, but it will be E8600 for sure .. again my luck has been not back for CPU's and i hope it not run out and that i get a good E8600.

[mac_69]

Yep what's your week markings on your chip though I'm curious. I will tell the manager tomorrow for you though and I will get his number and what not for you so you can call him and work with him directly because the other people I worked with were dumb..

Hmm seems to be working fine here, I had to take the heat sink off to read the chip markings, but ran prime95 small for a few hours okay

I have no idea how to read the chip. It says:

SLAWR MALAY
2.66 GHZ/12M/1333/05A
L8038553

and then what looks like e4 in a circle but my eyes are straining and CPUz says revision C1 stepping 7

Everest gives:

CPU Properties:
CPU Type QuadCore Intel Core 2 Q9450, 3600 MHz (8 x 450)
CPU Alias Yorkfield
CPU Stepping C1
Instruction Set x86, x86-64, MMX, SSE, SSE2, SSE3, SSSE3, SSE4.1
Original Clock 2667 MHz
Min / Max CPU Multiplier 6x / 8x
Engineering Sample No
L1 Code Cache 32 KB per core
L1 Data Cache 32 KB per core
L2 Cache 2x 6 MB (On-Die, ASC, Full-Speed)

Multi CPU:
CPU #1 Intel(R) Core(TM)2 Quad CPU Q9450 @ 2.66GHz, 3600 MHz
CPU #2 Intel(R) Core(TM)2 Quad CPU Q9450 @ 2.66GHz, 3600 MHz
CPU #3 Intel(R) Core(TM)2 Quad CPU Q9450 @ 2.66GHz, 3600 MHz
CPU #4 Intel(R) Core(TM)2 Quad CPU Q9450 @ 2.66GHz, 3600 MHz

CPU Physical Info:
Package Type 775 Contact LGA
Package Size 3.75 cm x 3.75 cm
Transistors 820 million
Process Technology 45 nm, CMOS, Cu, High-K Gate
Die Size 214 mm2
Core Voltage 1.150 - 1.250 V
I/O Voltage 1.150 - 1.250 V

CPU Manufacturer:
Company Name Intel Corporation
Product Information http://www.intel.com/products/processor

CPU Utilization:
CPU #1 / Core #1 0 %
CPU #1 / Core #2 0 %
CPU #1 / Core #3 0 %
CPU #1 / Core #4 0 %


CPUID

--------------------------------------------------------------------------------


CPUID Properties:
CPUID Manufacturer GenuineIntel
CPUID CPU Name Intel(R) Core(TM)2 Quad CPU Q9450 @ 2.66GHz
CPUID Revision 00010677h
IA Brand ID 00h (Unknown)
Platform ID 29h / MC 10h (LGA775)
IA CPU Serial Number Unknown
Microcode Update Revision 703
HTT / CMP Units 0 / 4
Tjmax Temperature 105 °C (221 °F)

[JohnLoph]

v.1306 seem like a little improve with DRAM Latency for my P5Q DLX
but DRAM overvolt still to much (0.11V - -*)

450x9 (Vcore 1.28750 bios)

V. 1103
Strap ---> 333MHz
DRAM Freq ---> DDR2-1081MHz
Timing ---> 5-5-5-5
DRAM static control ---> Disable
DRAM read training ---> Disable
Ai clock twister ---> Strong
PL ---> 8 [all disable]
DRAM voltage ---> 2.04 (2.13 real)

V. 1201
Strap ---> 333MHz
DRAM Freq ---> DDR2-1081MHz
Timing ---> 5-5-5-15
DRAM static control ---> Disable
DRAM read training ---> Disable
Ai clock twister ---> Stronger
PL ---> 8 [all disable]
DRAM voltage ---> 2.08 (2.18 real)

V. 1306
Strap ---> 333MHz
DRAM Freq ---> DDR2-1081MHz
Timing ---> 5-5-5-10
DRAM static control ---> Disable
DRAM read training ---> Disable
Ai clock twister ---> Stronger
PL ---> 8 [all disable]
DRAM voltage ---> 2.04 (2.15 real)

-------------------------------------------------------------------

500x8 (Vcore 1.28750 bios)

V. 1103
Strap ---> 333MHz
DRAM Freq ---> DDR2-1201MHz
Timing ---> 5-5-5-15-Auto-42-Auto-Auto
DRAM static control ---> Disable
DRAM read training ---> Disable
Ai clock twister ---> Moderate
PL ---> Auto
DRAM voltage ---> 2.22 (2.31 real)

V. 1201
Strap ---> 333MHz
DRAM Freq ---> DDR2-1201MHz
Timing ---> 5-5-5-18-Auto-50-Auto-Auto
DRAM static control ---> Disable
DRAM read training ---> Disable
Ai clock twister ---> Moderate
PL ---> 8 [all disable]
DRAM voltage ---> 2.26 (2.35 real)

V. 1306
Strap ---> 333MHz
DRAM Freq ---> DDR2-1081MHz
Timing ---> 5-5-5-18-Auto-50-Auto-Auto
DRAM static control ---> Disable
DRAM read training ---> Disable
Ai clock twister ---> Strong
PL ---> 8 [all disable]
DRAM voltage ---> 2.20 (2.31 real)
-------------------------------------------------------------

Conclude (for my P5Q DLX only) :
> DRAM overvolt
v.1103 = 0.09, v.1201 = 0.10, v.1306 = 0.11

> DRAM Latency : little improve with higher bios version

> Timing : unfortunate

[CERO]

shoudnt Performance level reduce to 7 when I manual input common performance level 8, and with all phase-pulls enabled? I tried this with enabled/disabled the performance level was still 8 in memset.

[CryptiK]

It really is very strange that the voltages for vdimm we are all measuring are so different. These boards seem to be very poorly made, with very little consistency. Perhaps it's time to throw it in the bin, and get something decent? It's a pity, because I thought the 'Deluxe' would be a high quality board. From all the problems, and variations between voltages etc, it appears they are in fact not high quality at all.

shoudnt Performance level reduce to 7 when I manual input common performance level 8, and with all phase-pulls enabled? I tried this with enabled/disabled the performance level was still 8 in memset.

No, it stays at whatever the PL is bios is set to, but in effect reduces latencies so it is equivalent to the next lower PL level. However the pullins actually reduce my boards performance, I have tried it numeerous times and with them all disabled, I get better performance. Go figure.

[Martin.v.r]

E8600 @ 5106 Q820A606 fun on Watercooling
Bios 1306
http://www.hwbot.org/result.do?resultId=766897

[CryptiK]

E8600 @ 5106 Q820A606 fun on Watercooling
Bios 1306
http://www.hwbot.org/result.do?resultId=766897

Nice mate, try 4-4-4-4 on the memory at that speed, it sill give you a bit of a boost.

[Martin.v.r]

ok i can try

[CERO]

No, it stays at whatever the PL is bios is set to, but in effect reduces latencies so it is equivalent to the next lower PL level. However the pullins actually reduce my boards performance, I have tried it numeerous times and with them all disabled, I get better performance. Go figure.

Hey cryptik, in what way? U get higher latency with phase-pulls enabled? Yea this is the first asus board i had with such a inconsistent behaviour, but i dont have real problems so far, only each time after BIOS update, I need to take out my DIMM and meassure the vDIMM to be sure.

[ci2kla]

From what I know is the pull ins is like
"Pull in phase adjust" in other boards
when enabled, that particular phase on particular channel should decrease by 1
which means when some of them enabled, some of them disabled
the performance level in each phase should differ but ASUS BIOS did nothing here...

[CryptiK]

Yes, I get worse performance with the pull-ins enabled. Strange huh.

Hey I just did this @ 2.29v (real):


My board is so weird, it's really unstable between 1277 and 1300, but higher than that it becomes OK again, and I simply cannot even post at all between 1200 and 1277. I cannot explain it at all. As you can see from the screenie @ DDR 1333, it's obviously not my ram at fault.

[CERO]

hmm on the DFI X48 and Rampage X48, it did worked, with such way u can bypass to a lower PL, I remembered on a certain FSB on the Rampage I coudnt boot on a certain commen PL , but with phase pulls it worked and u basicly still decreased the PL

[CERO]

Yes, I get worse performance with the pull-ins enabled. Strange huh.

Hey I just did this @ 2.29v (real):


My board is so weird, it's really unstable between 1277 and 1300, but higher than that it becomes OK again, and I simply cannot even post at all between 1200 and 1277. I cannot explain it at all.

Yea thats very strange, I am still testin my cpu, next step i will try if mine board also got this behaviour...also doesnt make sense that CL4 doesnt make a "difference" like @ CL5 , but still required the extra voltages to achieve it.

[truehighroller]

Hmm seems to be working fine here, I had to take the heat sink off to read the chip markings, but ran prime95 small for a few hours okay

I have no idea how to read the chip. It says:

SLAWR MALAY
2.66 GHZ/12M/1333/05A
L8038553

You will have no issues on 0803 at all trust me that cpu margin enhancement thing is not coded in on the 0803 and earlier BIOSes there for the issue goes away. I just wanted to see what week chip you were working with. You have a 803 and mine is a 813 totaly different weeks I just wanted to make sure it wasn't something with the CPUs them selves but, it obviously isn't. I am going to be calling them today and telling them I found another person with the same issue as me. I will IM you the contact info if you want it or you can just stick with the 0803 it's up to you no biggie but, still I found you which is good news for me because I'm not the only one having this issue and I have actualy chatted with someone with the same issue now and that makes me feel alot better about the whole thing.

The issue is not,, popping up on dual cores so dude with the e8600 on the way you shuld be fine.

[JohnLoph]

Yes, I get worse performance with the pull-ins enabled. Strange huh.

Hey I just did this @ 2.29v (real):


My board is so weird, it's really unstable between 1277 and 1300, but higher than that it becomes OK again, and I simply cannot even post at all between 1200 and 1277. I cannot explain it at all. As you can see from the screenie @ DDR 1333, it's obviously not my ram at fault.

I think that 1333MHz CL5 the Latency should lower that 52.0ns

because my RAM PL 8 all Disable (Enable / Disable that make no difference)
1200 @5-5-5-15 ---> 52.2 - 52.4ns
1081 @5-5-5-10 ---> 53.4 - 53.6ns

[CryptiK]

Yes I agree the latency should be lower @ 1333, I think my board has problems and I want to RMA it and get a replacement. I still can't post at DDR 1200, and it wont post all the way up to DDR 1277, where suddenly it does. I can't explain the issues my board has, it frustrates me a lot, I have tried absolutely everything.

[Mumid]

Yes I agree the latency should be lower @ 1333, I think my board has problems and I want to RMA it and get a replacement. I still can't post at DDR 1200, and it wont post all the way up to DDR 1277, where suddenly it does. I can't explain the issues my board has, it frustrates me a lot, I have tried absolutely everything.

Daft question perhaps but have you tried using different dimm slots? Also have you tried running single channel or with only 1 stick to see if you can post between ddr 1200 and 1277?

[truehighroller]

I'm amazed that he is at 1333. What memory do you have man? What size as well?

[SiGfever]

You will have no issues on 0803 at all trust me that cpu margin enhancement thing is not coded in on the 0803 and earlier BIOSes there for the issue goes away. I just wanted to see what week chip you were working with. You have a 803 and mine is a 813 totaly different weeks I just wanted to make sure it wasn't something with the CPUs them selves but, it obviously isn't. I am going to be calling them today and telling them I found another person with the same issue as me. I will IM you the contact info if you want it or you can just stick with the 0803 it's up to you no biggie but, still I found you which is good news for me because I'm not the only one having this issue and I have actualy chatted with someone with the same issue now and that makes me feel alot better about the whole thing.

The issue is not,, popping up on dual cores so dude with the e8600 on the way you shuld be fine.

Would you please ask if the Q9550 is having the same issues? I am on bios 1201 but have it set for "Compatible" after reading your post.
Thanks.

[Big Lar]

1 would think that if this CPU Margin Enhancement was causing problems with alot of folks, they would have recoded bios by now, and not kept it in a new bios version.

It would be good to know Exactly what problems it brings on and symptoms, before eveyone gets in a panic about it. Perhaps truehighroller or someone else can shed some light about this and thier expierience?

Larry

[BeastNotro]

They should just do away with it all together.

When i get my new CPU i hope a good one and i will then know if my MOBO is screwy or my Q6600 degraded in 2-3 weeks of folding, linpacking, orthosing and whatever else.... to only be able to 3.4ghz with it now linpack stable and not my 3.6ghz in 2-3 weeks THE THING WAS NEW .. so in three weeks of it time at 3.6ghz it now can't do it stable linpack that is and i not know if it can orthos 12 hours as i care not to try.

My new one E8600 dual-core yeah . hope she is good and if i can do 4.0-4.3 that will be nice, but is this MOBO screwing up CPU's and/or degrading them FAST.

[mac_69]

You will have no issues on 0803 at all trust me that cpu margin enhancement thing is not coded in on the 0803 and earlier BIOSes there for the issue goes away. I just wanted to see what week chip you were working with. You have a 803 and mine is a 813 totaly different weeks I just wanted to make sure it wasn't something with the CPUs them selves but, it obviously isn't. I am going to be calling them today and telling them I found another person with the same issue as me. I will IM you the contact info if you want it or you can just stick with the 0803 it's up to you no biggie but, still I found you which is good news for me because I'm not the only one having this issue and I have actualy chatted with someone with the same issue now and that makes me feel alot better about the whole thing.

The issue is not,, popping up on dual cores so dude with the e8600 on the way you shuld be fine.

PM sent - big thanks for your help in this, appreciate it

So, basically if we have Q9450 quad then we have to stick with bios 803 and cannot benefit from any future Asus bios updates. That's a bad situation to be in given that the board is basically brand new to market

Here I was cursing marvel drive expert and bios corruption and it was the margin enhancement making the system fall over

[Mumid]

They should just do away with it all together.

When i get my new CPU i hope a good one and i will then know if my MOBO is screwy or my Q6600 degraded in 2-3 weeks of folding, linpacking, orthosing and whatever else.... to only be able to 3.4ghz with it now linpack stable and not my 3.6ghz in 2-3 weeks THE THING WAS NEW .. so in three weeks of it time at 3.6ghz it now can't do it stable linpack that is and i not know if it can orthos 12 hours as i care not to try.

My new one E8600 dual-core yeah . hope she is good and if i can do 4.0-4.3 that will be nice, but is this MOBO screwing up CPU's and/or degrading them FAST.

Do you think that you might be stressing your mobo and cpu too much running linpack etc all the time when it is brand new? I mean this program puts on abnormal amount of load onto a system. I like to treat my new hardware like a new car lol, give it a bit of time to settle in before pushing it like mad! For me though memtest, orthos and occt is enough.

[BeastNotro]

Do you think that you might be stressing your mobo and cpu too much running linpack etc all the time when it is brand new? I mean this program puts on abnormal amount of load onto a system. I like to treat my new hardware like a new car lol, give it a bit of time to settle in before pushing it like mad! For me though memtest, orthos and occt is enough.

In never ran linpack until about 2 weeks of having the CPU as i never even knew about linpack .. so linpack i only been using maybe 1 week to 1 1/2, but not everyday .... but i did do it for 3-4 days in a row and not 24/7, but )3-6) times 5-10 loops each .... then maybe 3 days later i ran again newer version 1.4 at that time and she failed, so i upped the PPL and NB a little and she passed ... then i tried what a few days or so later 1.5.1 and she failed.. hmm .. i try 1.4 again and she failed....

But of course 3.4ghz SHE pass again with flying colors.

CPU degraded is what i am thinking and it may have taken 1-2 weeks of me testing and FOLDING the crap out of it ... that it did (what would of happened in 3-6 months) in 1-2 weeks is all.. is what i am thinking.

Now as for the E8600 i want to find it MAX ASAP as i hate taking 2 weeks to find out a CPU is a DUD...

Hey 3.4ghz is nothing to be sad about on a Q6600, but i want 4.0 ghz or betetr ... on any CPU right now and it appears the dual-cores are more proned fro thsi to happen.

So Q6600 is SOLD still in my rig until new CPU arrives .. so long Q6600 and one day i will either get another Quad-Core or upgrade to Nehalem, next year is my goal to upgrade to Nehalem ... so i just having fun for now on my rig and CPU's.