Overclock.net › Forums › Intel › Intel CPUs › [Official] The Ivy Bridge Stable / Suicide Club **Guides, Voltages, Temps & BIOS Templates** Inc SPREADSHEET
New Posts  All Forums:Forum Nav:

[Official] The Ivy Bridge Stable / Suicide Club **Guides, Voltages, Temps & BIOS Templates** Inc SPREADSHEET - Page 80

post #791 of 3792
Quote:
Originally Posted by bebimbap View Post

They gave the Z77 less phases because the IB procs needed less than SB procs. But i heard even 4+2 is enough to achieve good oc.

Thanks. Looks like the LE suits.

Len
My System
(15 items)
 
  
CPUMotherboardGraphicsRAM
3770K ASUS p8z77-v GTX760Ti G.Skill F3-17000CL11D-8GBXL (24 G) 
Hard DriveOptical DriveCoolingOS
1x SSD, Seagate 4TB 2X Samsung SSD EVO 250G 1X ... Liteon DVD RW Corsair H100 Win10 X64 Pro 
MonitorKeyboardPowerCase
Asus PB28Q Saitek Antec 850W Corsair Carbide 400R 
Mouse
Logitech 
  hide details  
Reply
My System
(15 items)
 
  
CPUMotherboardGraphicsRAM
3770K ASUS p8z77-v GTX760Ti G.Skill F3-17000CL11D-8GBXL (24 G) 
Hard DriveOptical DriveCoolingOS
1x SSD, Seagate 4TB 2X Samsung SSD EVO 250G 1X ... Liteon DVD RW Corsair H100 Win10 X64 Pro 
MonitorKeyboardPowerCase
Asus PB28Q Saitek Antec 850W Corsair Carbide 400R 
Mouse
Logitech 
  hide details  
Reply
post #792 of 3792
Quote:
Originally Posted by punceh View Post

if your testing with 26.6 and its stable on 26.6. your chip is going to be stable for all non-avx purposes. since theres very little programs actually using AVX i dont see why people should put 0.1V extra trough a chip just to have it stable for programs you dont use? once AVX becomes of real world use for me, ill get mine stable on 27.7.

You don't need .1 more volts for avx code. It's closer to .01.

There are plenty of programs already using AVX and there are more every single day thanks in part to the latest Intel C++ compilers because they can compile some code to use AVX even if the programmer doesn't specifically implement it.

EDIT: I couldn't peacefully use my system knowing it could be unstable in the near future. If you crash 2 months from now, you're more likely to think it's from degradation rather than AVX. It just doesn't make sense to stabilize with 26.6 anymore.
Edited by Iketh - 5/27/12 at 6:29pm
My PC
(16 items)
 
Living Room
(3 items)
 
 
CPUMotherboardGraphicsRAM
Intel 3770K @ 4.5 x 4 or 4.6 x 3, 1.26v, de-lid... MAXIMUS V GENE Radeon HD 7970 GHz @ 1080/1525 2 x 4GB G.Skill @ 2200MHz, 9-10-10-29-2t, 1.60v 
Hard DriveHard DriveOptical DriveCooling
Samsung 840 250GB 2 x 3TB Seagate ST3000VX000 LG Blu-Ray Burner Cooler Master Hyper 212 Evo Lapped 
OSMonitorKeyboardPower
Windows 8 64-bit ASUS VG236HE 23" 2ms 120Hz Logitech 920-000914 Black 106 Normal Keys USB W... Kingwin 550W Platinum 
CaseMouseMouse PadAudio
Fractal Design Define R4 Logitech G9x Rocketfish Sound Blaster ZxR 
MotherboardRAMOS
TZ68A+ Mushkin  Windows 7 Home Premium 64-bit 
  hide details  
Reply
My PC
(16 items)
 
Living Room
(3 items)
 
 
CPUMotherboardGraphicsRAM
Intel 3770K @ 4.5 x 4 or 4.6 x 3, 1.26v, de-lid... MAXIMUS V GENE Radeon HD 7970 GHz @ 1080/1525 2 x 4GB G.Skill @ 2200MHz, 9-10-10-29-2t, 1.60v 
Hard DriveHard DriveOptical DriveCooling
Samsung 840 250GB 2 x 3TB Seagate ST3000VX000 LG Blu-Ray Burner Cooler Master Hyper 212 Evo Lapped 
OSMonitorKeyboardPower
Windows 8 64-bit ASUS VG236HE 23" 2ms 120Hz Logitech 920-000914 Black 106 Normal Keys USB W... Kingwin 550W Platinum 
CaseMouseMouse PadAudio
Fractal Design Define R4 Logitech G9x Rocketfish Sound Blaster ZxR 
MotherboardRAMOS
TZ68A+ Mushkin  Windows 7 Home Premium 64-bit 
  hide details  
Reply
post #793 of 3792
So I thought I would see if my chip would boot @ 5ghz and see where that would take me tomorrow, or get an approximation of where I needed to be tonight...

1.33 - got start screen, but never finished
1.34 - start screend, then flashed like it was going into win, then reset
1.35 - got past SS, but then blue screen (more about that in a bit)
1.36 - same (so i decided to drop multi)
1.35 @ 49 - same BS code
1.35 @ 48 - same BS code
1.35 @ 47 - same BS code

Code was - "driver IRQL not less or equal" and checking the dmp I see "ntoskrnl.exe+25a40"

Back at stock I boot fine, and am looking into the code now... As far as I have gotten so far is a long list of driver issues it could be (most of those posts I found online were very old) or a memory issue.

Im going to run memtest overnight, but is there anything else specific I should look into, or a way to determine which driver could be causing this?
    
CPUMotherboardGraphicsRAM
i5 3570k Batch# 3207B712 Asus P8Z77-V PRO Asus 6950 DCii (2gb) 8Gb Gskill Ares 1866 
Hard DriveOptical DriveCoolingOS
1x120Gb ssd, 1x750Gb, 2x1.5Tb raid0, 2x3Tb Asus 12x BD-RW AC Cuplex Kryos (water) Windows 7 Ultimate x64 
MonitorKeyboardPowerCase
73" Mitsubishi DLP (73738) MX 5500 Corsair AX750 NZXT Switch 810 
MouseMouse Pad
MX Revolution Xtrac Ripper 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
i5 3570k Batch# 3207B712 Asus P8Z77-V PRO Asus 6950 DCii (2gb) 8Gb Gskill Ares 1866 
Hard DriveOptical DriveCoolingOS
1x120Gb ssd, 1x750Gb, 2x1.5Tb raid0, 2x3Tb Asus 12x BD-RW AC Cuplex Kryos (water) Windows 7 Ultimate x64 
MonitorKeyboardPowerCase
73" Mitsubishi DLP (73738) MX 5500 Corsair AX750 NZXT Switch 810 
MouseMouse Pad
MX Revolution Xtrac Ripper 
  hide details  
Reply
post #794 of 3792
1.35 maybe low for 5 Ghz. Try higher, maybe around 1.4V, though I doubt that will work as well. I think 1.45V is average for 5 ghz right now.
E IV
(16 items)
 
  
CPUMotherboardGraphicsRAM
Intel i5 3570K ASRock Z77 OC Formula Nvidia Zotac GTX 770 Gskill Trident 2400 Mhz 
Hard DriveHard DriveHard DriveOS
Samsung 840  OCZ Vertex 2 Seagate 7200 Windows 8 
MonitorKeyboardPowerCase
22" Benq Logitech Antec HCP 850W Caselabs SM8 
MouseMouse PadAudioAudio
Logitech G400 Razer Asus Xonar DGX Microlabs C Series 
  hide details  
Reply
E IV
(16 items)
 
  
CPUMotherboardGraphicsRAM
Intel i5 3570K ASRock Z77 OC Formula Nvidia Zotac GTX 770 Gskill Trident 2400 Mhz 
Hard DriveHard DriveHard DriveOS
Samsung 840  OCZ Vertex 2 Seagate 7200 Windows 8 
MonitorKeyboardPowerCase
22" Benq Logitech Antec HCP 850W Caselabs SM8 
MouseMouse PadAudioAudio
Logitech G400 Razer Asus Xonar DGX Microlabs C Series 
  hide details  
Reply
post #795 of 3792
Quote:
Originally Posted by mav2000 View Post

1.35 maybe low for 5 Ghz. Try higher, maybe around 1.4V, though I doubt that will work as well. I think 1.45V is average for 5 ghz right now.

Agreed... At this point I wasnt trying to stabilize, just see what would get me into windows... The BS error made me want to drop multi down to see if it persisted and the fact I couldnt even get into win with 1.35 @ 4.7 was getting me a bit worried.

As far as running stability for 5 that will more than likely not happen unless I de-lid this chip (which I only plan on doing if it seems to be a decent chip, hence the seeing what voltage just to get into win) I planned on seeing what max multi would be, but driver IRQL BS has put a hold on that.

4.8 will more than likely be my run for 24/7 use though...
Edited by Jesse D - 5/28/12 at 12:13am
    
CPUMotherboardGraphicsRAM
i5 3570k Batch# 3207B712 Asus P8Z77-V PRO Asus 6950 DCii (2gb) 8Gb Gskill Ares 1866 
Hard DriveOptical DriveCoolingOS
1x120Gb ssd, 1x750Gb, 2x1.5Tb raid0, 2x3Tb Asus 12x BD-RW AC Cuplex Kryos (water) Windows 7 Ultimate x64 
MonitorKeyboardPowerCase
73" Mitsubishi DLP (73738) MX 5500 Corsair AX750 NZXT Switch 810 
MouseMouse Pad
MX Revolution Xtrac Ripper 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
i5 3570k Batch# 3207B712 Asus P8Z77-V PRO Asus 6950 DCii (2gb) 8Gb Gskill Ares 1866 
Hard DriveOptical DriveCoolingOS
1x120Gb ssd, 1x750Gb, 2x1.5Tb raid0, 2x3Tb Asus 12x BD-RW AC Cuplex Kryos (water) Windows 7 Ultimate x64 
MonitorKeyboardPowerCase
73" Mitsubishi DLP (73738) MX 5500 Corsair AX750 NZXT Switch 810 
MouseMouse Pad
MX Revolution Xtrac Ripper 
  hide details  
Reply
post #796 of 3792
Quote:
Originally Posted by Iketh View Post

You don't need .1 more volts for avx code. It's closer to .01.
There are plenty of programs already using AVX and there are more every single day thanks in part to the latest Intel C++ compilers because they can compile some code to use AVX even if the programmer doesn't specifically implement it.
EDIT: I couldn't peacefully use my system knowing it could be unstable in the near future. If you crash 2 months from now, you're more likely to think it's from degradation rather than AVX. It just doesn't make sense to stabilize with 26.6 anymore.
Agreed with your edit remarks.Software is a constantly evolving ever-changing medium and better to future proof, IMO by being prepared. If you look around at how many people think their systems are stable after running Prime 95 for endless hours only to crash unexpectedly.....
post #797 of 3792
Quote:
Originally Posted by mandrix View Post

Agreed with your edit remarks.Software is a constantly evolving ever-changing medium and better to future proof, IMO by being prepared. If you look around at how many people think their systems are stable after running Prime 95 for endless hours only to crash unexpectedly.....


The issue wasn't that 26.6 is most likely to represent software that is available.
The issue is that 27.7 isn't a full release version. Just like 27.4, 27.5, 27.6 it is a beta and might not be stable, so you might get errors that aren't related at all to your OC.

Now 27.7 might end up being the next release version, so you might be getting it ahead of the release date, but again, what ever you are comfortable with. thumb.gif
post #798 of 3792
Quote:
Originally Posted by bebimbap View Post

The issue wasn't that 26.6 is most likely to represent software that is available.
The issue is that 27.7 isn't a full release version. Just like 27.4, 27.5, 27.6 it is a beta and might not be stable, so you might get errors that aren't related at all to your OC.
Now 27.7 might end up being the next release version, so you might be getting it ahead of the release date, but again, what ever you are comfortable with. thumb.gif

Prime 27.7 has already been "released" as the newest stable version. go to home page, it isnt beta anymore...
http://www.mersenne.org/freesoft/
http://www.mersenne.org/
My System
(15 items)
 
  
CPUMotherboardGraphicsRAM
6900K 4.4ghz  MSI X99A Krait Titan XP Gskill 32GB 3200 C14 
Hard DriveHard DriveCoolingOS
Samsung 960 Pro 1TB OS Samsung 850 pro 1TB games D5, XT45 280 rad, Nemesis 280 rad,  Windows 10 64bit 
MonitorPowerCase
Acer Predator XB271HU 144hz Seasonic Prime Titanium 850 Fractal Design R5 
  hide details  
Reply
My System
(15 items)
 
  
CPUMotherboardGraphicsRAM
6900K 4.4ghz  MSI X99A Krait Titan XP Gskill 32GB 3200 C14 
Hard DriveHard DriveCoolingOS
Samsung 960 Pro 1TB OS Samsung 850 pro 1TB games D5, XT45 280 rad, Nemesis 280 rad,  Windows 10 64bit 
MonitorPowerCase
Acer Predator XB271HU 144hz Seasonic Prime Titanium 850 Fractal Design R5 
  hide details  
Reply
post #799 of 3792
Quote:
Originally Posted by bebimbap View Post


So anyway, if you use LLC just know you have been warned you might be sending spikes of voltage much higher than you anticipated.

What I simply fail to understand with those like you preaching against using LLC is this...

Intel specifies a "peak to peak" transient response, lets say for example 110mv. Transient spikes or undershoots and overshoots for example sake are each 75mv at max vid/max load, and occur over a 2-4 microsecond interval (oscilloscope with refresh rate of 100-4000 nanoseconds needed).

There are multiple ways of corralling the would be 150mv (75mv overshoot + 75 mv undershoot) into the specified 110mv.
Option 1 is add more output capacitors, ie more expensive motherboard, and such exists.
Option 2 is use active voltage positioning (vdroop) and fewer output capacitors.
http://cds.linear.com/docs/Design%20Note/dn224f.pdf link with actual oscilloscope and mathematically predicted both, and I used values from this actual oscilloscope to avoid confusion.

There are cheap motherboards with cheap vrm circuitry and relatively few output caps that need vdrop/vdroop at max to contain the peak to peak within specs at stock settings.
Then there are more expensive motherboards that have several times the caps, beefier vrm, that would maintain similar peak to peak spec even with LLC on modest levels, even when OCed.

For example, LLC high (which still allows .04v vdrop/vdroop on my UD5h), may well maintain a 110mv peak to peak at even modest overclock, and much better than spec 110mv at stock settings.
Whereas a cheap mobo with poor vrm/ much fewer caps, may have a 150mv peak to peak despite LLC is off, during a modest overclock, since barely meet specs at stock.

So when you tell someone with a mobo with a beefy vrm circuitry to turn off LLC because its bad, yet you say nothing to someone with cheap mobo overclocking, you are likely saying something like 110mv peak to peak is bad since your using LLC on high. But 150mv peak to peak is just fine, since you turned LLC off on your cheaper mobo.

My point isnt to say cheap motherboards are bad, it is just to say, there are two ways to maintain peak to peak within tolerance, 1)better vrm or 2)cheap vrm and more vdroop. I didnt buy a more expensive mobo to just be able to use higher LLC, yet maintain same specs as a cheaper mobo with LLC off, but since I do have better vrm, I am going to take advantage of it and use LLC.

And, you mention the downside of LLC, but dont mention the upside. With unchecked high vdroop, every time your computer encounters any new software that cause a higher wattage draw on your cpu than you have encountered before, your vcore drops much more, and you risk crash. Intel stock vcore is much higher than necessary for running prime at stock for more than one reason. With LLC on moderate to high levels, still allowing some vdroop but much less, you are much less likely to crash if your cpu consumes 10 more watts, since vcore drops much less compared to unchecked vdroop.
Edited by opt33 - 5/28/12 at 10:40am
My System
(15 items)
 
  
CPUMotherboardGraphicsRAM
6900K 4.4ghz  MSI X99A Krait Titan XP Gskill 32GB 3200 C14 
Hard DriveHard DriveCoolingOS
Samsung 960 Pro 1TB OS Samsung 850 pro 1TB games D5, XT45 280 rad, Nemesis 280 rad,  Windows 10 64bit 
MonitorPowerCase
Acer Predator XB271HU 144hz Seasonic Prime Titanium 850 Fractal Design R5 
  hide details  
Reply
My System
(15 items)
 
  
CPUMotherboardGraphicsRAM
6900K 4.4ghz  MSI X99A Krait Titan XP Gskill 32GB 3200 C14 
Hard DriveHard DriveCoolingOS
Samsung 960 Pro 1TB OS Samsung 850 pro 1TB games D5, XT45 280 rad, Nemesis 280 rad,  Windows 10 64bit 
MonitorPowerCase
Acer Predator XB271HU 144hz Seasonic Prime Titanium 850 Fractal Design R5 
  hide details  
Reply
post #800 of 3792
ah 27.7 went release on May 15 so the OP's reasoning behind not using 27.7 is no longer valid
I was just repeating the reasoning behind the original decision.

and I never said LLC was bad, all i said were there could be consequences to using it such as unwanted overshoot.
I didn't include the pros because everyone knows what they are anyway, such as more consistent voltage across all loads.

and although negative overshoot might crash your computer the unwanted overshoot could possibly fry your cpu which is the worse outcome.
And not everyone will use a mobo with "beefy" VRMs and then assume OCing has no risks because your selected voltage is well below the tolerance.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Intel CPUs
  • [Official] The Ivy Bridge Stable / Suicide Club **Guides, Voltages, Temps & BIOS Templates** Inc SPREADSHEET
Overclock.net › Forums › Intel › Intel CPUs › [Official] The Ivy Bridge Stable / Suicide Club **Guides, Voltages, Temps & BIOS Templates** Inc SPREADSHEET