The E0 Q9550 SLB9V thread (finally!)

[askimo]

you can order them online. shipping wont be that much.

I don't think you can; the e-store says that it is only available for pickup in the store

What other options do I have?

[Unrealer]

still testing...

I think the temperature is good, when i bought it i expected around 60-70º at 4Ghz

Q9550 E0 L844C062
DFI LanParty P35 Dark
TRUE lapped & 2x1000rpms (around 22º of ambient temp)
471x8,5 (4Ghz)
vcore 1.33v (@ full)
NB 1.50v
VTT 1.47v
GTL 0/2 0.61v
GTL 1/3 0.61v
NB GTL 0.67

OCCT 30mins test LARGE DATA SET


I need to lower the vcore till gets stable.

I confused about GTLs and NB/VTT (when i have E4300 and E5400 i let these on AUTO ) so im trying all the posible values...

PD: i have paid 105€ (shipping included) for it, and im waiting another one for the same price....

[Kain665]

I just picked a SLB8V today (1.25v VID, Bah...), and fully expected it to go to 4ghz. Unfortunately, I didn't know about the X38 FSB wall with yorkies. So it's not stable at anything over 400mhz fsb... bleh. I'll pick up a UD3P or something sometime soon, but in the mean time, I'm seeing how low below stock VID I can run it. Here is my best power-sipping undervolt/overclock:



400x8.5 at 1.18275 set in bios. Have to love my load temps, Big Typhoon in a cramped case with a GTX275, with an 81 Degree Fahrenheit ambient (~27c) :P

[phsinc1]

@ rurik

What vcore you get under load? 1.264v?

[Unrealer]

4.25H Stable (i stopped the test)



Q9550 E0 L844C062
DFI LanParty P35 Dark
TRUE lapped & 2x1000rpms (around 22º of ambient temp)
471x8,5 (4Ghz)
Vcore 1.34v (@ full)
NB 1.50v
VTT 1.47
CPU GTL 0/2 0,61
CPU GTL 1/3 0,61
NB GTL 0,67





Still testing...

[rurik]

@ rurik

What vcore you get under load? 1.264v?

yep.In bios @1.275 or something...

[rurik]

@Unrealer
my chip do 20 loops of Linx at lower voltages,but Prime95 have a error at 5min.Try Prime too.

[truehighroller]

@Unrealer
my chip do 20 loops of Linx at lower voltages,but Prime95 have a error at 5min.Try Prime too.

If you can see what the GFlops that your CPU is putting out with Linx, or use Intelburntest which I know you can see the GFLops the CPU is pushing in that program "I just haven't used Linx mind you". You will be able to probably see that your FSB Term is set to low and your GFlops being put out by the CPU when cold or when it has been shut down all night or all day, are probably half of what they should be with your chip.

I am pushing 58.5 GFlops with my CPU @ 490 x 8.5 @ 1.296v under load, this is with utilizing 3084MBs of my Ram~ for the test and the test finish in around 86s a piece~. I had noticed this because when booting my system cold from being shut down all day or all night, I noticed that when running Intelburntest the GFlops being put out by my CPU were half of what they are when the system has been warmed up or after being rebooted a few times because of me tweaking settings. You can also use it to fine tune your GTL settings as well.

I noticed that if the GTLs are off then two of the cores on the CPU will get hotter then the other two while running Intelburntest or all cores will not get hot enough. This can significantly cut time you spend tweaking these settings because you don't have to finish the test you can just go ahead and stop the test reboot and tweak some more until you see the temps are even and about the right temperature under load taking in to consideration your cooling ofcourse.

Then once you think you have the cores hitting good temps and evenly for that matter, let the test finish and see what the GFlops being pushed out are and if the time finished is fast enough as well. If you get everything set just right BIOS setting wise, all cores will stay about the same temperature all the way through the test nice and stable like. I just caught on to this and was able to tweak mine to perfection landing me the following.

3DMark Vantage Score 17719
http://service.futuremark.com/compare?3dmv=1441908

Here is what my bandwidth is with Everest as well. This is the best I have been able to get this board to do in Everest, at this speed that is.



I had to set my FSB Term to 1.44 with a GTL Reference of .84 "on my motherboard mind you, your's might be different from mine GTL setting wise so be carefull there" to get everything to act correctly when first firing my PC up from being shut down all night or all day and the results speak for them selves. Hope this helps people because if I had known this awhile ago I could have saved my self some serious time tweaking BIOS settings.

[Unrealer]

@Unrealer
my chip do 20 loops of Linx at lower voltages,but Prime95 have a error at 5min.Try Prime too.

Testing...

[truehighroller]

Testing...

Can you see what kind of GFlops you are pushing with Linx? Thank you for your time.

[Unrealer]

Can you see what kind of GFlops you are pushing with Linx? Thank you for your time.

XP 32Bits SP3

[truehighroller]

XP 32Bits SP3

Your FSB Term is set to low, or Intelburntest is better optimized one of the two. I would look into that. Look at my numbers compared to yours. There shouldn't be that big of a difference in speed between me and you. Your CPU can be stable and still not be performing the way it should be. Bump up your FSB Term and fine tune your CPU GTL and test again and I bet it goes up. Not only that, your system will act alot more responsive once you hit the right FSB Term voltage for what speed you are at.


Edit: Sorry I just noticed that I didn't post a pic of my testing with Intelburntest but, I am hitting 58.5 GFlops compared to your 42 and we are pretty close speed wise. I think you should be hitting 55 GFlops.

[-n7-]

nvm testing

[truehighroller]

That's not why his GFlops are lower.

His are lower because he's testing less RAM than you.

Are you sure because I swear if, I remember correctly mine was pushing the same GFlops no matter what amount of memory I was doing the only thing that changed was the time spent testing. Maybe I'm wrong, maybe I'm not I'm not near my PC right now.

[Zucker2k]

Are you sure because I swear if, I remember correctly mine was pushing the same GFlops no matter what amount of memory I was doing the only thing that changed was the time spent testing. Maybe I'm wrong, maybe I'm not I'm not near my PC right now.

Don't swear so fast, ram-size (problem-size) does affect g-flops.

[-n7-]

As soon as i said it's a RAM thing, i doubted myself lol, hence the previous post edit.

It does make a difference though.

You also have a higher FSB & likely higher RAM speed than he does, & have tweaked it properly (like lowered tRD, etc.).

Okay, this is just quick testing of different RAM sizes, which in turn increases problem size.

1024MB


2048MB


4096MB


6144MB

[truehighroller]

As soon as i said it's a RAM thing, i doubted myself lol, hence the previous post edit.

It does make a difference though.

You also have a higher FSB & likely higher RAM speed than he does, & have tweaked it properly (like lowered tRD, etc.).

Okay, this is just quick testing of different RAM sizes, which in turn increases problem size.

1024MB


6144MB

Ok but, that was not that big of a difference though right? Like a tiny difference there compared to the 10GFlops I am thinking he is losing. Thank you for checking by the way. Not only that but, going from 470ish to the 490 I am sitting at now I think increased it only by 2 at the most I beleive.

[Unrealer]

Bump up your FSB Term and fine tune your CPU GTL

Do you mean VTT? what value should i use then?

And im confused about GTLs, i tried 0,67 (on both) and it was stable during 2mins, i tried 0,63 and it was stable during 8mins.... i tried with 0,61 (also on both) and is OCCT stable for 5 hours or more (i stopped the test).
The last value i need to test is 0,57 and i also need to test North Bridge GTL, cause i let it default (0,67).

[truehighroller]

What I did was get it semi stable or boot-able then tweak one setting up and down until I could see / feel what spot was more, stable again just by watching my temps while running burntest for the most part then leave that setting at the best spot and go to the next one and do the same thing. The FSB Term has different names, it depends on the board. I'm sorry I don't know all the names.

It was the fastest I have ever been able to set all my settings, honestly. So, if your CPU GTLs were the most stable say at 63 because .67 was 2 minutes and lets just say you went to .59 as well and it was one minute then stay at .63 and move to the next one. I think if you were to watch your temps while benching you would see what I am talking about.

Edit: Here I am home now, this is at 490 x 8.5.




This is at 470 x 8.5. I didn't feel like doing the big one again so I just did 1024. I think I could actually use some GTL tweaking again here since I changed the FSB and I could probably get better here as well I think.




I know my temps are a little high but, this cooler can't handle much more. I need to get water. I figure no other programs or games get it this hot any way. Prime95 only gets it to 75C tops I believe.

[phsinc1]

What I did was get it semi stable or boot-able then tweak one setting up and down until I could see / feel what spot was more, stable again just by watching my temps while running burntest for the most part then leave that setting at the best spot and go to the next one and do the same thing. The FSB Term has different names, it depends on the board. I'm sorry I don't know all the names.

It was the fastest I have ever been able to set all my settings, honestly. So, if your CPU GTLs were the most stable say at 63 because .67 was 2 minutes and lets just say you went to .59 as well and it was one minute then stay at .63 and move to the next one. I think if you were to watch your temps while benching you would see what I am talking about.

Edit: Here I am home now, this is at 490 x 8.5.




This is at 470 x 8.5. I didn't feel like doing the big one again so I just did 1024. I think I could actually use some GTL tweaking again here since I changed the FSB and I could probably get better here as well I think.




I know my temps are a little high but, this cooler can't handle much more. I need to get water. I figure no other programs or games get it this hot any way. Prime95 only gets it to 75C tops I believe.

Thanks for this tip, I just tried and found my GFlops was a bit low, increased CPU Termination to 1.30v from 1.28v and GFLops are now normal. I also played with GTL, I was using 0.848, I increased it to 0.863 and my temps are more even now.

using 475x8.5
CPU Termination 1.30v
CPU Ref 0.863
MCH Core 1.32v
MCH Ref 0.863
CPU PLL 1.50v

[truehighroller]

I just happened to be watching my temps and started noticying the annomaly and was like wow and started utilizing it and it was so easy to tweak everything this way, I figured I just had to tell every one. Your welcome. I was expieriencing a little, tiny, lag when playing NFS Shift as well before noticying this and now it is gone and my system is a lot more snappy.

[phsinc1]

I just happened to be wtaching my temps and started noticying the annomaly and was like wow and started utilizing it and it was so easy to tweak everything this way, I fugred I just had to tell every one. Your welcome. I was expieriencing a little, tiny, lag when playing NFS Shift as well before noticying this and now it is gone and my system is a lot more snappy.

Can I know what is your settings for your current OC?
CPU Termination
CPU Ref
MCH Core
MCH Ref

Thanks

[truehighroller]

Can I know what is your settings for your current OC?
CPU Termination
CPU Ref
MCH Core
MCH Ref

Thanks

Ok I will ramble them off and just assume I am hitting close here because I might be off a little but, surely not to much.


Robust Graphics Booster - Turbo
CPU Clock Ratio - 8X
Fine Clock Tune: .5x
CPU Host Frequency - 490
PCI Express Frquency - 101
C.I.A.2 - Disabled
CPU Clock Drive - 1000mV
PCI Express Clock Drive 1000mV
Clock skews are all set to 0PS.

****DRAM Performance Control****
Performance Enhance - Turbo
Extreme Memory Profile - Disabled
(G) MCH Frequency Latch - 333MHz
System Memory Multiplier - 2.40B
Memory Speed: 1176MHz
DRAM Timing Selectable - Manual
>>>Sandard Timing Control<<<
CAS Latency Time - 5
tRCD - 5
tRP - 5
tRAS - 15
>>>Advanced Timing Control<<<
tRRD - auto
tWTR - auto
tWR - auto
tRFC - auto
tRTP - auto
Command Rate - auto
>>>>Channel A Timing Settings
Static tRead Value - 8
Everything else is Auto.
>>>>Channel A Driving Settings
Driving Strength Profile - Auto
Everything else is Auto.
>>>>Channel B Timing Settings
Same as Chanel A.
>>>>Channel B Driving Settings
Same as Channel A.

****Motherboard Voltage Control****
>>>>CPU
Load Line Calibration - Enabled

CPU Vcore - 1.3475V "Load is actual 1.296 and 1.312 idle"

CPU Termination - 1.42V "I think I actually bumped this down one notch because my CPU liked it better at 1.42. I was going to look last night but was to busy testing and then playing NFS Shift, then I went to bed."

CPU PLL - 1.57V
CPU Reference - .83 or .84
>>>>MCH/ICH
MCH Core - 1.52V
MCH Reference - .912
MCH/DRAM Refence - 1.000v
ICH I/O - 1.50V
ICH Core - 1.10V
>>>DRAM
DRAM Voltage - 2.100V
DRAM Termination - .75v
Channel A Reference - Auto
Channel B Reference - Auto

[rurik]

truehighroller



is this normal?system running 7days without restart. but my temp too high.@3952 at moment vcore is 1.280V(Im too lazy to lower it ) 75 is max with my IFX-14+Noctua NF-P12 in closed case

[truehighroller]

truehighroller



is this normal?system running 7days without restart. but my temp too high.@3952 at moment vcore is 1.280V(Im too lazy to lower it ) 75 is max with my IFX-14+Noctua NF-P12 in closed case

That's not to high of temps for air cooling and when you run games etcedera your temps will not be nearly that high, no program tortures CPUs like Linx code. I think if you hone everything in a little more you will be suprised about how your temps go up, I'm guessing you will hit 80C~. I think you might have some settings off a little but, you are pretty close to honed in though. I am basing this off just looking at that screen shot though. You need to watch your temps like I had mentioned earlier, they speak volumes.