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

[Official] The Sandy Stable Club **Guides, Voltages, Temps & BIOS Templates** Inc SPREADSHEET - Page 806

post #8051 of 10697
Quote:
Originally Posted by ryuji View Post

it shows on asus motherboards, i have little experience with other vendors however
you REALLY need to sit there and stare at the vcore for a long time tho, as it updates somewhat slow, there will be a really low/really high comparatively voltage that will show up once in a while
what gets you in trouble is when the voltage is 'low' RIGHT when the cpu needs it wink.gif
when trying for lower llc values, focus on what you see as that minimum voltage, try and match them up. i think you will notice lower average voltage needed for stability. of course test 12 hrs as usual once you 'find' something

Makes sense. At higher clock, though, lowering LLC would mean sacrificing idle voltage. but i guess that doesn't really matter much if the max load voltage can be reduced.
Quote:
Originally Posted by Stu-Crossfire View Post

Thanks very much you two guys, this afternoon has been a great education for me, please keep it up. smile.gif

Lol same here.
Quote:
Originally Posted by AeroZ View Post

I don't think that that's an issue at all because I can use offset voltage to set my idle voltage and additional turbo voltage to set load voltage. Basically any LLC level is ok. I can't figure out if it's better to keep vcore more stable under load with 75% LLC or let it jump more with 50% or even lower level of LLC.

From what ryuji says it's better to have it more stable so you don't get caught "off-guard" when the vcore decided to drop when you need it. However it appears vcore should be more stable at 50% than 75%, and not the other way around.

edit: was about to go to bed when a thought struck me: tongue.gif Obviously for stress testing and benching then max load vcore stability is better, but wouldn't it be better for the chip in terms of longevity to have high LLC to minimise idle voltage, since a chip is almost definitely going to spend more time idle than being at max load throughout the course of its life?
Edited by Sashimi - 3/28/12 at 8:23am
Nebula
(33 items)
 
  
CPUCPUMotherboardGraphics
Intel Core i7 2700K Intrel Core i7 2700K Asus p8z68-v Pro Gen3 MSI N580GTX Lightning Xtreme 3GB Edition 
GraphicsRAMHard DriveHard Drive
MSI N580GTX Lightning Xtreme 3GB Edition G.Skill RipjawX F3-17000CL9D-8GBXM Western Digital Caviar Black 2TB 7200RPM Seagate Barracuda LP 2TB 5900RPM 
Optical DriveCoolingCoolingCooling
Pioneer BDR-206BK 12X Blu-Ray Writer Drive Koolance RP-452X2 Dual 5.25in Reservoir for 1-2... Koolance 450s Pump Koolance 450s Pump 
CoolingCoolingCoolingCooling
EK Supreme HF Bitspower Black Freezer Acrylic Top With Clear ... Bitspower Black Freezer Acrylic Top With Clear ... Black Ice GTX 240 Radiator 
CoolingCoolingCoolingCooling
Black Ice GTX 480 Radiator Koolance CTR-SPD24X2 Dual 24V Pump Controller Nidec Servo Scythe Gentle Typhoon 120mm Case Fa... Xigmatek Crystal Red LED 140mm Case Fan 1000RPM 
CoolingCoolingOSMonitor
Bitfenix Spectre PRO Red LED 230mm Case Fan 900RPM Silverstone Air-Penetrator UV-Blue 120mm Case F... Windows 7 Home Premium 64 Bits Benq XL2410T 120hz 
MonitorKeyboardPowerCase
Samsung SyncMaster 740N 17" Azio Levetron Mech 5 Mechanical Gaming Keyboard Cooler Master Silent Pro Hybrid 1300W Aerocool X-Predator Evil Black Edition 
MouseAudioOtherOther
Logitech G9X Gaming Mouse Auzentech X-Meridian 7.1 2G Lamptron FC5 V2 Fan Controller Bitspower Dual 20cm Cold Cathode Kit UV 
Other
Bitspower Dual 30cm Cold Cathode Kit UV 
  hide details  
Reply
Nebula
(33 items)
 
  
CPUCPUMotherboardGraphics
Intel Core i7 2700K Intrel Core i7 2700K Asus p8z68-v Pro Gen3 MSI N580GTX Lightning Xtreme 3GB Edition 
GraphicsRAMHard DriveHard Drive
MSI N580GTX Lightning Xtreme 3GB Edition G.Skill RipjawX F3-17000CL9D-8GBXM Western Digital Caviar Black 2TB 7200RPM Seagate Barracuda LP 2TB 5900RPM 
Optical DriveCoolingCoolingCooling
Pioneer BDR-206BK 12X Blu-Ray Writer Drive Koolance RP-452X2 Dual 5.25in Reservoir for 1-2... Koolance 450s Pump Koolance 450s Pump 
CoolingCoolingCoolingCooling
EK Supreme HF Bitspower Black Freezer Acrylic Top With Clear ... Bitspower Black Freezer Acrylic Top With Clear ... Black Ice GTX 240 Radiator 
CoolingCoolingCoolingCooling
Black Ice GTX 480 Radiator Koolance CTR-SPD24X2 Dual 24V Pump Controller Nidec Servo Scythe Gentle Typhoon 120mm Case Fa... Xigmatek Crystal Red LED 140mm Case Fan 1000RPM 
CoolingCoolingOSMonitor
Bitfenix Spectre PRO Red LED 230mm Case Fan 900RPM Silverstone Air-Penetrator UV-Blue 120mm Case F... Windows 7 Home Premium 64 Bits Benq XL2410T 120hz 
MonitorKeyboardPowerCase
Samsung SyncMaster 740N 17" Azio Levetron Mech 5 Mechanical Gaming Keyboard Cooler Master Silent Pro Hybrid 1300W Aerocool X-Predator Evil Black Edition 
MouseAudioOtherOther
Logitech G9X Gaming Mouse Auzentech X-Meridian 7.1 2G Lamptron FC5 V2 Fan Controller Bitspower Dual 20cm Cold Cathode Kit UV 
Other
Bitspower Dual 30cm Cold Cathode Kit UV 
  hide details  
Reply
post #8052 of 10697
Quote:
Originally Posted by Sashimi View Post

edit: was about to go to bed when a thought struck me: tongue.gif Obviously for stress testing and benching then max load vcore stability is better, but wouldn't it be better for the chip in terms of longevity to have high LLC to minimise idle voltage, since a chip is almost definitely going to spend more time idle than being at max load throughout the course of its life?

I wouldn't be concerned. electro-migration is a function of voltage, yes, but temperature is a large factor as well as frequency. all three are at a minimum when idle on SB, temperature drastically so. If degradation/damage is going to happen its from transients and full load operational parameters. not to say less voltage isnt better for life, but like many other things its a balancing game of variables
Desktop
(12 items)
 
Router/Server
(6 items)
 
 
CPUMotherboardGraphicsRAM
2600k@4.7 Ghz Asus Sabertooth P67 Sapphire HD5870 4x4gb Corsair Vengeance 
Hard DriveOptical DriveCoolingMonitor
120 GB OCZ Vertex 3 Bluray RW drive Silent dual loop water cooling gpu+cpu 2x ASUS VE247H (23.6 inch LED) 
PowerCaseMouseAudio
Antec Truepower Quattro 1KW Thermaltake Armor Razer Naga Epic Asus Xonar D2X, 5.1 channel tube amp 
CPUMotherboardRAMHard Drive
E8400 Rampage Formula Mushkin DDR2 3x WD Caviar Black 1TB 64mb cache 
CoolingOS
Silent water cooling 1x120mm radiator, chipset+... Gentoo Linux 
  hide details  
Reply
Desktop
(12 items)
 
Router/Server
(6 items)
 
 
CPUMotherboardGraphicsRAM
2600k@4.7 Ghz Asus Sabertooth P67 Sapphire HD5870 4x4gb Corsair Vengeance 
Hard DriveOptical DriveCoolingMonitor
120 GB OCZ Vertex 3 Bluray RW drive Silent dual loop water cooling gpu+cpu 2x ASUS VE247H (23.6 inch LED) 
PowerCaseMouseAudio
Antec Truepower Quattro 1KW Thermaltake Armor Razer Naga Epic Asus Xonar D2X, 5.1 channel tube amp 
CPUMotherboardRAMHard Drive
E8400 Rampage Formula Mushkin DDR2 3x WD Caviar Black 1TB 64mb cache 
CoolingOS
Silent water cooling 1x120mm radiator, chipset+... Gentoo Linux 
  hide details  
Reply
post #8053 of 10697
Does this help at all guys?

Vcore behavior with different LLC settings (2600K, graphs inside)

BTW, no reason at all to watch CPU-Z to determine peak min and max Vcore values, for starters it's way too slow imho...

All you have to do is use an app that actually logs per second or even faster (i.e. AIDA and HWInfo are two apps that can do this, there are a many more) and analyze the raw data after the measurement session...just don't forget to set it to log values per second or faster.

The longer the session (time-wise) the more valid the conclusions will be, personally when i was exploring my cpu's OC capabilities i logged everything and in fact i have kept all the logs just in case i want to extract the raw data and re-analyze them at some point...

Good thing is that anybody can experiment, log and see what's what and how "everything" works in the real world... wink.gif
post #8054 of 10697
i use cpuz because i trust it, it pulls that vcore voltage low-level from the cpu itself not the motherboard, if there are other utilities that do it the same way then thats cool with me but for the longest time intel only shared the low-level info with the cpuz developer. i'd rather go measure it myself.

Your charts look pretty nice tho, keep in mind that this will vary depending on cpu/mobo tho so each person will probably have to do there own testing

from the looks of your first graph it seems like high llc is the best result.

A better test would be comparing LLC with offset adjusted to obtain same average voltage output. Results could be completely different.
Edited by ryuji - 3/28/12 at 9:22am
Desktop
(12 items)
 
Router/Server
(6 items)
 
 
CPUMotherboardGraphicsRAM
2600k@4.7 Ghz Asus Sabertooth P67 Sapphire HD5870 4x4gb Corsair Vengeance 
Hard DriveOptical DriveCoolingMonitor
120 GB OCZ Vertex 3 Bluray RW drive Silent dual loop water cooling gpu+cpu 2x ASUS VE247H (23.6 inch LED) 
PowerCaseMouseAudio
Antec Truepower Quattro 1KW Thermaltake Armor Razer Naga Epic Asus Xonar D2X, 5.1 channel tube amp 
CPUMotherboardRAMHard Drive
E8400 Rampage Formula Mushkin DDR2 3x WD Caviar Black 1TB 64mb cache 
CoolingOS
Silent water cooling 1x120mm radiator, chipset+... Gentoo Linux 
  hide details  
Reply
Desktop
(12 items)
 
Router/Server
(6 items)
 
 
CPUMotherboardGraphicsRAM
2600k@4.7 Ghz Asus Sabertooth P67 Sapphire HD5870 4x4gb Corsair Vengeance 
Hard DriveOptical DriveCoolingMonitor
120 GB OCZ Vertex 3 Bluray RW drive Silent dual loop water cooling gpu+cpu 2x ASUS VE247H (23.6 inch LED) 
PowerCaseMouseAudio
Antec Truepower Quattro 1KW Thermaltake Armor Razer Naga Epic Asus Xonar D2X, 5.1 channel tube amp 
CPUMotherboardRAMHard Drive
E8400 Rampage Formula Mushkin DDR2 3x WD Caviar Black 1TB 64mb cache 
CoolingOS
Silent water cooling 1x120mm radiator, chipset+... Gentoo Linux 
  hide details  
Reply
post #8055 of 10697
Again imho and in my experience, it's way too slow and if you compare the values with let's say HWinfo/AIDA64 (and many other progs) you will probably realize that they read the same sensor/s (at least in my case)...the difference is that i.e if you set the AIDA to report sensor values per second the values will jump up and down more than CPU-Zs. They show the same thing, but the CPU-Z is slower...

Also CPU-Z is not the only one that shows the true Vcore, it really depends on the motherboard as well...for instance CPU-Z shows invalid Vcore for my Gigabyte H61N-USB3, the figures just don't make sense at all...HWInfo for some reason shows the "correct" ones (didn't bother with AIDA64 this time)...

Always imho.
post #8056 of 10697
Quote:
Originally Posted by fommof View Post

Again imho and in my experience, it's way too slow and if you compare the values with let's say HWinfo/AIDA64 (and many other progs) you will probably realize that they read the same sensor/s (at least in my case)...the difference is that i.e if you set the AIDA to report sensor values per second the values will jump up and down more than CPU-Zs. They show the same thing, but the CPU-Z is slower...
Also CPU-Z is not the only one that shows the true Vcore, it really depends on the motherboard as well...for instance CPU-Z shows invalid Vcore for my Gigabyte H61N-USB3, the figures just don't make sense at all...HWInfo for some reason shows the "correct" ones (didn't bother with AIDA64 this time)...
Always imho.

I wasnt saying your recording method was wrong. In fact i thought your data was good thumb.gif

i myself complained that cpuz was slow, but last time i checked out other hw monitors they all read mobo sensor readings, doesnt seem to be true anymore smile.gif
Desktop
(12 items)
 
Router/Server
(6 items)
 
 
CPUMotherboardGraphicsRAM
2600k@4.7 Ghz Asus Sabertooth P67 Sapphire HD5870 4x4gb Corsair Vengeance 
Hard DriveOptical DriveCoolingMonitor
120 GB OCZ Vertex 3 Bluray RW drive Silent dual loop water cooling gpu+cpu 2x ASUS VE247H (23.6 inch LED) 
PowerCaseMouseAudio
Antec Truepower Quattro 1KW Thermaltake Armor Razer Naga Epic Asus Xonar D2X, 5.1 channel tube amp 
CPUMotherboardRAMHard Drive
E8400 Rampage Formula Mushkin DDR2 3x WD Caviar Black 1TB 64mb cache 
CoolingOS
Silent water cooling 1x120mm radiator, chipset+... Gentoo Linux 
  hide details  
Reply
Desktop
(12 items)
 
Router/Server
(6 items)
 
 
CPUMotherboardGraphicsRAM
2600k@4.7 Ghz Asus Sabertooth P67 Sapphire HD5870 4x4gb Corsair Vengeance 
Hard DriveOptical DriveCoolingMonitor
120 GB OCZ Vertex 3 Bluray RW drive Silent dual loop water cooling gpu+cpu 2x ASUS VE247H (23.6 inch LED) 
PowerCaseMouseAudio
Antec Truepower Quattro 1KW Thermaltake Armor Razer Naga Epic Asus Xonar D2X, 5.1 channel tube amp 
CPUMotherboardRAMHard Drive
E8400 Rampage Formula Mushkin DDR2 3x WD Caviar Black 1TB 64mb cache 
CoolingOS
Silent water cooling 1x120mm radiator, chipset+... Gentoo Linux 
  hide details  
Reply
post #8057 of 10697
Quote:
Originally Posted by ryuji View Post

Your charts look pretty nice tho, keep in mind that this will vary depending on cpu/mobo tho so each person will probably have to do there own testing
from the looks of your first graph it seems like high llc is the best result.

I have already mentioned that best thing is to anybody get their own measurements and see how everything works...everybody can do it, if somebody wants to get some decent measurements he should run way more extended sessions, that experiment was just for me initially...
Quote:
Originally Posted by ryuji View Post

A better test would be comparing LLC with offset adjusted to obtain same average voltage output. Results could be completely different.

Yes but that would not show what i wanted to show (initially to my self): how llc affects Vcore while all the other parameters stay the same (actually only when idle, i knew exactly what's going on when on full load)... wink.gif

Anyway, as i wrote before, these kind of measurements can be done by anybody as long as he/she has the time...no reason at all to "trust" my raw data... wink.gif
Edited by fommof - 3/28/12 at 9:37am
post #8058 of 10697
Quote:
Originally Posted by ryuji View Post

I wasnt saying your recording method was wrong. In fact i thought your data was good thumb.gif
i myself complained that cpuz was slow, but last time i checked out other hw monitors they all read mobo sensor readings, doesnt seem to be true anymore smile.gif

It's all good my brother, btw thanks a lot for the graph at post #8041...nice one and rep given. thumb.gif
Edited by fommof - 3/28/12 at 9:36am
post #8059 of 10697
Here is my stable OC smile.gif

338
post #8060 of 10697
.
Petitmort
(14 items)
 
  
CPUMotherboardGraphicsHard Drive
i9-7900x Asus Rampage VI Extreme EVGA GTX 1080 TI FTW3 iCX Hydro Copper Samsung 960 Pro NVMe M.2 1TB 
CoolingOSMonitorMonitor
EK Water Loop Win 10 Dell 30" Atron 27" 
MonitorKeyboardPowerCase
Gateway 24" Razer Blackwidow X Chroma Seasonic 1000w Caselabs S8 
MouseMouse Pad
Razer Naga Chroma I use 6 mousepads, lol 
  hide details  
Reply
Petitmort
(14 items)
 
  
CPUMotherboardGraphicsHard Drive
i9-7900x Asus Rampage VI Extreme EVGA GTX 1080 TI FTW3 iCX Hydro Copper Samsung 960 Pro NVMe M.2 1TB 
CoolingOSMonitorMonitor
EK Water Loop Win 10 Dell 30" Atron 27" 
MonitorKeyboardPowerCase
Gateway 24" Razer Blackwidow X Chroma Seasonic 1000w Caselabs S8 
MouseMouse Pad
Razer Naga Chroma I use 6 mousepads, lol 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Intel CPUs
  • [Official] The Sandy Stable Club **Guides, Voltages, Temps & BIOS Templates** Inc SPREADSHEET
Overclock.net › Forums › Intel › Intel CPUs › [Official] The Sandy Stable Club **Guides, Voltages, Temps & BIOS Templates** Inc SPREADSHEET