Overclock.net › Forums › Intel › Intel - General › Squeezing E8400 with X38 and stuck - weird stuff happening
New Posts  All Forums:Forum Nav:

Squeezing E8400 with X38 and stuck - weird stuff happening

post #1 of 4
Thread Starter 
Hello awesome, members of an awesome forum. (My loss indeed that i wasn't active here).

I've upgraded my build with new RAM and VGA, in hopes of utilizing all the hidden potential of my Wolfie and bringing it to level of high quality gaming once again as well as binging myself to digital frontiers of overclocking.

Current overclock achieved is 4.5 GHz @ 500 FSB and 1500 MHz RAM but it isn't quite stable.
__________________________________________________ __________________

FSB - 500 MHz
Multi - 9
FSB strip to NB - 400 MHz
PCI-E - 111 MHz
RAM - 1500 MHz
Timings - 7 - 8 - 7 - 20 (N2)
Read and write controls - disabled
AI clock twister - light (not sure)
AI Transaction Booster - enabled, lvl 2

CPU - 1.448 v - load | 1.4 v - idle | 1.46875 v - bios (setting)
PLL - 1.64 v
FSB VTT - 1.42 v
NB - 1.67 v
SB - 1.05 v
RAM - 2.06 v
CPU GTL - x62
NB GTL - x61
Clock over-charging - 1.0 v (not sure)
CPU Spread Spectrum - auto
PCI-E Spred Spectrum - disabled
__________________________________________________ __________________

And here are the weird (and interesting, in some level) problems that i've encountered during the past week of tinkering, testing and benching:

I'm between some kind of voltage walls because if i go any lower then obviously, i lose stability but if i go any higher (with any given voltage) i lose stability as well as some windows performance.

The only voltage setting which i can play with a bit is NB voltage - the benefit for stability of increasing it is fighting the drawbacks quite evidently and the effect on stability of adjusting it is related with SB voltage in some weird way. It seems that there is some kind of voltage cap for both of them together.

So the best results are achieved with NB - 1.67 v, SB - 1.05 v or NB - 1.65 v, SB - 1.20 v. I can go much lower with VNB but it doesn't seem to improve stability.


Even more weird discovery: when i was thinking of giving up on 4.5 GHz and anywhere near it, (cause lower FSB didn't provide me with enough stability to accept the trade) i've tried out CPU Spread Spectrum for a last resort, despite all the info that it hinders stability and performance, so i've set it to "auto" (only "disable" and "auto" modes available) and it reignited the enthusiasm in me as the system doesn't crash after 5 mins but performs for 30 mins before doing so and instead of decreasing performance, it increases it a little bit. PCI-E Spread Spectrum, hoever, does seem to hinder stability.


Method of testing used: As soon as desktop loads, i open and launch Orthos, 1080p youtube video (on repeat), torrent client with active downloads and uploads and winamp, tuned to radio. Before switching CPU Spread Spectrum back to normal, the system was giving out even before me completing all these tasks.


Some more weird info - RAM doesn't like slower timings at all. If i loosen up at least one of the given timings a bit, it doesn't boot (Command rate is an exception but i didn't notice that i'd affect stability).

AI Transaction Booster increases CPU performance on lvl 2 and while on lvl -8 the performance is considerably worse, appearance of more stability isn't evident.

NB GTL "x61" yelds way more stability than "x67" (noticed only when enabled CPU Spread Spectrum).

Question: is my testing method appropriate?

And the last words to this overly long rookie thread - i assume these "weird" things mentioned above aren't so weird to you pro's so yeah, i'm waiting for any thoughts with anticipation
CPPC4_OVERRIDE
(15 items)
 
  
CPUMotherboardGraphicsRAM
e8400 P5E3 Deluxe Asus EAH6950 1GB Corsair DOMINATOR 2X 2GB DDR3 1600 MHz 
Hard DriveCoolingOSMonitor
WD 36GB 10k RPM, WD 1TB 7.2k RPM Gigabyte liquid Windows 7 Ultimate 64 bit Dell E198WFP 
KeyboardPowerCaseMouse
Logitech UltraX Media slim Corsair HX850W Antec Skeleton Cyborg R.A.T. 9 wireless 6400 dpi 
Mouse PadAudio
Leather Creative I-Trigue 3500 
  hide details  
Reply
CPPC4_OVERRIDE
(15 items)
 
  
CPUMotherboardGraphicsRAM
e8400 P5E3 Deluxe Asus EAH6950 1GB Corsair DOMINATOR 2X 2GB DDR3 1600 MHz 
Hard DriveCoolingOSMonitor
WD 36GB 10k RPM, WD 1TB 7.2k RPM Gigabyte liquid Windows 7 Ultimate 64 bit Dell E198WFP 
KeyboardPowerCaseMouse
Logitech UltraX Media slim Corsair HX850W Antec Skeleton Cyborg R.A.T. 9 wireless 6400 dpi 
Mouse PadAudio
Leather Creative I-Trigue 3500 
  hide details  
Reply
post #2 of 4
thats a nice overclock you have there. i got my E8400 to 5ghz but that was on a 10+2 Phase P45 board so it was a little more advanced with its settings and would have provided cleaner power.

as for the "weird" things happening CPU Spread Spectrum has been shown that in some cases when a high overclock is being used enabling it can improve stability. so this is a fairly normal finding.

id think that if you are finding stability can not be improved by increasing voltages then either your board or your CPU is at its max and you have hit a FSB wall. a 500FSB for a X35 board is pretty good work even to get it to boot at that. also the fact that any voltage you increase causes instability may be a sign that your Mosfet chips are getting to their limit of output, just a stab at another reason.

only thing i can really think of doing if voltages have no effect is playing around with the CPU GTL settings either increasing or decreasing to find its sweet spot. i found that the CPU GTL got me 100mhz more out of my E8400 and Q9400 when you found a good setting, however like i said i was using a more advanced board that had 4 GTL settings for the cpu, so not sure how this GTL settings work on your board.
Finally Finished!
(14 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7 3930K Batch: 3135B495T Gigabyte X79 UD7 GTX 570 3-way SLi Corsair Vengeance 1600mhz CL9 
Hard DriveCoolingOSMonitor
60GB OCZ Vertex II / 3x 500GB Samsung Spinpoint EK Supreme HF Nickel, EK DCP4.0, Stealth GT 360... Windows 7 Ult 64bit Acer 22" G225HQ 
KeyboardPowerCaseMouse
Razor Lacosa Mirror Edition Silverstone Strider 1500w  Aerocool X-Preditor  CM Storm Sentinel Advanced 
Mouse Pad
Razer Goliathus Speed edition 
  hide details  
Reply
Finally Finished!
(14 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7 3930K Batch: 3135B495T Gigabyte X79 UD7 GTX 570 3-way SLi Corsair Vengeance 1600mhz CL9 
Hard DriveCoolingOSMonitor
60GB OCZ Vertex II / 3x 500GB Samsung Spinpoint EK Supreme HF Nickel, EK DCP4.0, Stealth GT 360... Windows 7 Ult 64bit Acer 22" G225HQ 
KeyboardPowerCaseMouse
Razor Lacosa Mirror Edition Silverstone Strider 1500w  Aerocool X-Preditor  CM Storm Sentinel Advanced 
Mouse Pad
Razer Goliathus Speed edition 
  hide details  
Reply
post #3 of 4
Thread Starter 
CPU GTL settings available for me are "x62", "x63", "x65", "x67". I didn't notice a difference but i'll try to play with them more with current settings.

Are these Mosfet chips related to each others performance? If yes, then maybe lowering RAM voltage would let me increase other voltages? I'll play with that as well.

Yes so it's most likely that my x38 chipset is at it's limit, but could it be that 1600 MHz (set to 1500 MHz) RAM are limiting OC because they aren't 1:2 to FSB?

Thanks.

----------
UPDATE
----------

Now this is really weird. I've just got considerably more stability with RAM voltage of 1.66 (from 2.06 v) then upped NB to 1.75 v but it turned out quite unstable and after system crash, i've accidentaly loaded defaults, then i've set everything to it's previous values again and it couldn't even load windows normaly. Changed RAM voltage back to 2.06 and i'm getting same stability results as before with 2.06 v RAM.

----------
UPDATE
----------

Now i've set 1.66 v RAM and it's running again, so far that is. I wonder, how long. This time i've left all the max stability voltage values that i've discovered before gaining stability from lower v RAM.

Could any body explain what happened in the first update of this post?

----------
UPDATE
----------

It crashed after nearly 30 mins of stressing on 1.66 v RAM and didn't boot again. Increased to 1.70 v - didn't boot, increased back to 2.06 v - booted, decreased back to 1.66 v - booted, now trying out 1.60 v - so far, so good. I really do not understand.

----------
UPDATE
----------

I'm excited lol. It managed orthos (small FFTs) priority 5 stressing with the apps running mentioned in the first post for 1h 33 min. Settings were as listed in first post except RAM voltage lowered to 1.62 v, NB - 1.65 v, SB - 1.2 v. Trying out 1.6 v RAM now. I hope it's getting sufficient voltage and it will bring total stability to this OC otherwise i would hit a wall. Btw "x62" and "x61" GTL multipliers yeld most stability for some mysterious reason.

Btw if it crashes under lower RAM voltage, and i let the BSOD process (memory dump?) be completed, it boots, otherwise it requires more v RAM to boot (and after that it can boot with low v RAM again).

----------
UPDATE
----------

So the 1.62 v RAM is an ultaimate number. Can't really increase or decrease v Core, v PLL or USB VTT. Now running: NB - 1.65 v, SB - 1.05 v. Hopefuly, lower SB will give less stress to Mosfet chips thus resulting in more stable power supply.
Edited by DEADmetal - 10/29/11 at 4:04am
CPPC4_OVERRIDE
(15 items)
 
  
CPUMotherboardGraphicsRAM
e8400 P5E3 Deluxe Asus EAH6950 1GB Corsair DOMINATOR 2X 2GB DDR3 1600 MHz 
Hard DriveCoolingOSMonitor
WD 36GB 10k RPM, WD 1TB 7.2k RPM Gigabyte liquid Windows 7 Ultimate 64 bit Dell E198WFP 
KeyboardPowerCaseMouse
Logitech UltraX Media slim Corsair HX850W Antec Skeleton Cyborg R.A.T. 9 wireless 6400 dpi 
Mouse PadAudio
Leather Creative I-Trigue 3500 
  hide details  
Reply
CPPC4_OVERRIDE
(15 items)
 
  
CPUMotherboardGraphicsRAM
e8400 P5E3 Deluxe Asus EAH6950 1GB Corsair DOMINATOR 2X 2GB DDR3 1600 MHz 
Hard DriveCoolingOSMonitor
WD 36GB 10k RPM, WD 1TB 7.2k RPM Gigabyte liquid Windows 7 Ultimate 64 bit Dell E198WFP 
KeyboardPowerCaseMouse
Logitech UltraX Media slim Corsair HX850W Antec Skeleton Cyborg R.A.T. 9 wireless 6400 dpi 
Mouse PadAudio
Leather Creative I-Trigue 3500 
  hide details  
Reply
post #4 of 4
Thread Starter 
Ok, it seems i've really hit a wall. I don't want to switch my RAM to single channel, if that would even help.

It's good for gaming, no matter how much i load it, but Orthos test fails sooner or later.

Anyone would happen to have any more advices?

-----
edit

I've just booted windows with 517 MHz FSB with same settings for sake of curiosity, of course, it BSOD'ed even before Orthos was launched, however it fully loaded windows.
Edited by DEADmetal - 10/30/11 at 9:51am
CPPC4_OVERRIDE
(15 items)
 
  
CPUMotherboardGraphicsRAM
e8400 P5E3 Deluxe Asus EAH6950 1GB Corsair DOMINATOR 2X 2GB DDR3 1600 MHz 
Hard DriveCoolingOSMonitor
WD 36GB 10k RPM, WD 1TB 7.2k RPM Gigabyte liquid Windows 7 Ultimate 64 bit Dell E198WFP 
KeyboardPowerCaseMouse
Logitech UltraX Media slim Corsair HX850W Antec Skeleton Cyborg R.A.T. 9 wireless 6400 dpi 
Mouse PadAudio
Leather Creative I-Trigue 3500 
  hide details  
Reply
CPPC4_OVERRIDE
(15 items)
 
  
CPUMotherboardGraphicsRAM
e8400 P5E3 Deluxe Asus EAH6950 1GB Corsair DOMINATOR 2X 2GB DDR3 1600 MHz 
Hard DriveCoolingOSMonitor
WD 36GB 10k RPM, WD 1TB 7.2k RPM Gigabyte liquid Windows 7 Ultimate 64 bit Dell E198WFP 
KeyboardPowerCaseMouse
Logitech UltraX Media slim Corsair HX850W Antec Skeleton Cyborg R.A.T. 9 wireless 6400 dpi 
Mouse PadAudio
Leather Creative I-Trigue 3500 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Intel - General
Overclock.net › Forums › Intel › Intel - General › Squeezing E8400 with X38 and stuck - weird stuff happening