Overclock.net › Forums › Cooling › Water Cooling › Need to shave off a few more C degress with water cooling upgrades, need opinions please :)
New Posts  All Forums:Forum Nav:

Need to shave off a few more C degress with water cooling upgrades, need opinions please :) - Page 3

post #21 of 61
This should be good for a few degrees.
post #22 of 61
Thread Starter 
Quote:
Originally Posted by Mindchi|l View Post

This should be good for a few degrees.

thanks, that is a great idea and an ingenious design! grats on this one smile.gif
I may get one in future if all my current efforts fail...
post #23 of 61
Thread Starter 
ok. the 24v pump controller installed and 3 extra fans (ap-15's) were added to the internal EX360 rad. unfortunately I am unable to install extra 2 fans on the external RX240 rad without draining the loop and disconnecting it - reason is the gutted shroud I used for mounting it is coming from a slightly skinnier Sctyhe Slipstream and there is maybe 1mm difference in thickness of it and the AP-15 shroud thickness ... so I can't add even a single fan over there (to the empty spot), this will have to wait until I drain the loop for maintenance in future.

anyhow, so far I noticed a drop in CPU temps under load at same settings, reaching maybe 2-3C degrees, however my current ambient room temp is higher by nearly 2C versus former screenshot, so it could be 4-5C lower in fact. time will tell if this was worth it smile.gif

maybe tomorrow I will turn the AC in morning to stabilize room temperature and re-do the same test as posted earlier, still please keep in mind that only one radiator (internal 360) runs the push-pull config as of now (plan was to convert both of them at same time).
Edited by feniks - 9/29/12 at 6:05pm
post #24 of 61
Thread Starter 
huh, seeing maybe 2C improvement in under load temps at high overclocks ... damn ivy bridge ... it's the second chip and runs even hotter than my former one ...
now trying a different approach with vcore offsets, but as far as I can see at 4.7GHz it gives me no thermal improvements either (when stable) ... perhaps I should just ditch Intel Burn Test 2.54 and start using Prime95 again ... that should beat some under load temps down surely haha!
post #25 of 61
Thread Starter 
succeed in a different approach path with vcore offsets! biggrin.gif

recent cooling system upgrades (stronger flow @ 24V pump power + more fans on one of rads now in push-pull config) and today's change in approach was able to get me to 4.9GHz with offset vcore of +0.245 and 25% LLC (and was running AC to drop ambient room temps to 24.0C instead of 27.4C) ... anyways, it worked under IBT! I got it fully stable with no WHEA warnings for 1 round of IBT in Maximum mode (or 10 rounds of Standard Test with 1GB used) using 2x4GB of RAM @ 1333 @ 1.50V! it also passed benchmarking software easily, however at nomal settings of memory running 4x4GB @ 2200MHz @ 1.65V biggrin.gif

my 3770 @ 4.9GHz took +0.245 vcore offset, resulting in 1.400V vcore under load (1.472v at idle at 49x multi without EIST).

mve49ghzoffset0245vcore.png
Uploaded with ImageShack.us

hottest core still spiked up to 105C, but it happened only two times and thermal throttling did NOT occur at that time (I watched it closely in CPU-Z), all good! biggrin.gif normally that hot core stayed at 102-103C for most of time.

3dmark11 P10453 http://www.3dmark.com/3dm11/4536572
vantage P37637 http://www.3dmark.com/3dmv/4320285
SuperPI (1.5 xmod) 6m 48.243s

biggrin.gif

my next goal is the 5GHz on 3770K, but for that I will need to wait for a specific time window ... namely when the outside temperatures drop suddenly to like 45F and our building central heating lags for a few days (this results in around 13-18C ambient room temperature he he) biggrin.gif

... shouldn't take long to come biggrin.gif
Edited by feniks - 10/1/12 at 9:37pm
post #26 of 61
Quote:
Originally Posted by feniks View Post

I got it fully stable with no WHEA warnings for 1 round of IBT in Maximum mode (or 10 rounds of Standard Test with 1GB used) using 2x4GB of RAM @ 1333 @ 1.50V! it also passed benchmarking software easily, however at nomal settings of memory running 4x4GB @ 2200MHz @ 1.65V biggrin.gif

You do know that IBT works by comparing the results from successive "rounds" (times to run). If they match, all is well. Without the comparison, all that is proven is that the computer did not BSOD.

I have also found that I need to run many rounds of IBT, 8 hours of Prime, and various 3D tests before my computer is stable, as they all stress in different ways.
Dad's Beast
(18 items)
 
  
CPUMotherboardGraphicsRAM
i7 3770K ASROCK Z77 Extreme4 Asus GTX 590 Patriot Viper Black Mamba 
Hard DriveHard DriveHard DriveHard Drive
2 X Corsair Force 3 in RAID 0 2 X Seagate Barracuda in RAID 1 Seagate Barracuda Seagate Barracuda 
Optical DriveCoolingOSMonitor
LG Blu Ray player DVD Burner Thermochill PA120.2 GT AP15 in push, DDC 3.25/X... Windows 7 Professional 64bit Optoma HD66 (3D) 
KeyboardPowerCaseMouse
Logitech G15 Rev 1 ad Rev 2 (Dual Locations) OCZ ZX 1250 HAF X Logitech M570 and Microsoft Trackball Explorere... 
AudioOther
Logitech Z5500 with Asus Xonar GS 4 Sets of Nvidia 3D Glasses 
  hide details  
Reply
Dad's Beast
(18 items)
 
  
CPUMotherboardGraphicsRAM
i7 3770K ASROCK Z77 Extreme4 Asus GTX 590 Patriot Viper Black Mamba 
Hard DriveHard DriveHard DriveHard Drive
2 X Corsair Force 3 in RAID 0 2 X Seagate Barracuda in RAID 1 Seagate Barracuda Seagate Barracuda 
Optical DriveCoolingOSMonitor
LG Blu Ray player DVD Burner Thermochill PA120.2 GT AP15 in push, DDC 3.25/X... Windows 7 Professional 64bit Optoma HD66 (3D) 
KeyboardPowerCaseMouse
Logitech G15 Rev 1 ad Rev 2 (Dual Locations) OCZ ZX 1250 HAF X Logitech M570 and Microsoft Trackball Explorere... 
AudioOther
Logitech Z5500 with Asus Xonar GS 4 Sets of Nvidia 3D Glasses 
  hide details  
Reply
post #27 of 61
Thread Starter 
Quote:
Originally Posted by Capt Proton View Post

You do know that IBT works by comparing the results from successive "rounds" (times to run). If they match, all is well. Without the comparison, all that is proven is that the computer did not BSOD.
I have also found that I need to run many rounds of IBT, 8 hours of Prime, and various 3D tests before my computer is stable, as they all stress in different ways.

yes, I do smile.gif and that stands true up to Sandy Bridge. However with Ivy Bridge there is a new simple pointer, a warning in Event Logs which states that CPU was unstable (WHEA warning = Windows Hardware Error Architecture) even though it passed a stress test.

What I do now with IVB chips is to simply run 10 rounds of Standard mode of IBT2.54 and adjust vcore up by 1 notch at a time until I fully get rid off WHEA warnings, it takes about 30 seconds per round in Standard Mode (used 1GB of RAM). Once there I run 1-5 rounds of IBT in Maximum Mode (using nearly all available memory), and most of the time it proves stable or sometimes call for 1 more notch of vcore if WHEA warning return (usually only 1 at this point). That's what I did at 4.7GHz and 4.8GHz and after it proved stable even after I left my computer running IBT for half of the night successfully passing 20 rounds of IBT2.54 in Maximum Stress mode using 16GB of RAM.

with 4.9GHz however I had no thermal headroom to run IBT in Max Stress mode with 8GB of RAM for any longer, would need to drop the ambient room temps deeper down. anyways, there was no WHEA warnings, so it was stable and passed benchmarking. I bet it was stable enough even for gaming (might need a notch or two higher on vcore though for rock solid stability.)

I do not use Prime95, it simply takes ages to prove stable/unstable, usually I opt in for IBT, it's much quicker and unless temps are a problem then I revert to Prime95 (runs cooler by a few C degrees) or OCCT.

I hope it explains what I did and why I claim it was stable. IVB chips are different from any other before and those WHEA warnings they send are pretty useful.
Edited by feniks - 10/2/12 at 10:58am
post #28 of 61
OK, that works for you. I wouldn't consider my computer stable based only on that. At least seeral hours of intense gaming are required before I would declare my computer stable. I know many others are not as fussy.
Dad's Beast
(18 items)
 
  
CPUMotherboardGraphicsRAM
i7 3770K ASROCK Z77 Extreme4 Asus GTX 590 Patriot Viper Black Mamba 
Hard DriveHard DriveHard DriveHard Drive
2 X Corsair Force 3 in RAID 0 2 X Seagate Barracuda in RAID 1 Seagate Barracuda Seagate Barracuda 
Optical DriveCoolingOSMonitor
LG Blu Ray player DVD Burner Thermochill PA120.2 GT AP15 in push, DDC 3.25/X... Windows 7 Professional 64bit Optoma HD66 (3D) 
KeyboardPowerCaseMouse
Logitech G15 Rev 1 ad Rev 2 (Dual Locations) OCZ ZX 1250 HAF X Logitech M570 and Microsoft Trackball Explorere... 
AudioOther
Logitech Z5500 with Asus Xonar GS 4 Sets of Nvidia 3D Glasses 
  hide details  
Reply
Dad's Beast
(18 items)
 
  
CPUMotherboardGraphicsRAM
i7 3770K ASROCK Z77 Extreme4 Asus GTX 590 Patriot Viper Black Mamba 
Hard DriveHard DriveHard DriveHard Drive
2 X Corsair Force 3 in RAID 0 2 X Seagate Barracuda in RAID 1 Seagate Barracuda Seagate Barracuda 
Optical DriveCoolingOSMonitor
LG Blu Ray player DVD Burner Thermochill PA120.2 GT AP15 in push, DDC 3.25/X... Windows 7 Professional 64bit Optoma HD66 (3D) 
KeyboardPowerCaseMouse
Logitech G15 Rev 1 ad Rev 2 (Dual Locations) OCZ ZX 1250 HAF X Logitech M570 and Microsoft Trackball Explorere... 
AudioOther
Logitech Z5500 with Asus Xonar GS 4 Sets of Nvidia 3D Glasses 
  hide details  
Reply
post #29 of 61
Thread Starter 
Quote:
Originally Posted by Capt Proton View Post

OK, that works for you. I wouldn't consider my computer stable based only on that. At least seeral hours of intense gaming are required before I would declare my computer stable. I know many others are not as fussy.

yup, and my OC's at 4.7-4.8GHz have been proven stable under several hours of gaming (plus 20 rounds of IBT 2.54 in MAXIMUM mode using 16GB of RAM)... just not at 4.9GHz, because I cannot maintain low enough ambient room temps nor I can afford a phase change CPU cooler ... 4.9GHz is purely for benchmarking and it served its purpose. think what you will, you always have a right to post your opinion and I don't have to agree with it wink.gif

I know it was stable enough (for benchmarking or a short IBT run) and couldn't be further than 1-2 notches of vcore from rock solid stable (aka stable under everything), because that's how it works for me following the steps and stress testing methodology I described above, YMMV, I base my opinion on my own experience with my systems.
Edited by feniks - 10/2/12 at 4:14pm
post #30 of 61
Thread Starter 
Actually I figured I will take a hit on 5GHz OC this Saturday ... not that it became cold outside suddenly, no not yet wink.gif ... but I have an evil plan of placing my rig so the air intake (and my internal 360 rad) is oriented right in front of my AC cold air outlet when running at full frosty blast biggrin.gif
I might be able to drop down my CPU temps by good 5-10C with this approach if all works out as I think it would - sort of a computer Cold Air Intake LOL! biggrin.gif

Of course the ambient room temps will be cooled down to approx 22C if time allows (or max 24C if not much time).
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Water Cooling
Overclock.net › Forums › Cooling › Water Cooling › Need to shave off a few more C degress with water cooling upgrades, need opinions please :)