Page 1 of 2

Core i5-750 system mild overclocking cool low power PC

Posted: Tue Jan 26, 2010 3:59 am
by Tyr Antilles
Core i5-750 system - moderate overclocking at 3.4GHz - cool and low power PC (last update 2010.04.28 - I will update this page if I will change anything)

This guide is for the ones who like to bring i5-750 closer to it's "true potential" without any risk for their components on the long therm. I am writing this based only on my personal experience with my components. After a lot of searching trial and testing, I managed to SAFELY overclock my system to 3.4GHz, keeping very low voltages and temperatures. This is a system designed to run on hot ambient up to 40ºC.

Why 3.4GHz? Because:
 3.4GHz over 2.66GHz default is a decent overclock of roughly 30% (27.82%) and it is more then enough for all my needs. I will see no benefits to run a 4GHz rocket, my games would not run much more faster and my work won't be done essentially quicker.
 I want to be completely on the safe side, I don't like fried CPU's.
 Keeping low voltages means less heat and less stress on components that leads to longer life; I intend to keep my computer a very long time.
 Power consumption is a LOT lower with all power saving features enabled. At 4 GHz you need to disable them.

The noise level is very low from the three 12cm fans that came with Lian-Li case, and the CPU fan don’t even counts. The fan from video card is audible only in 3D applications and the loudest components of the system are the three HDD’s. Overall the system is fairly quiet and under the desk where I keep it is totally not disturbing.


- HARDWARE:
o MB Gigabyte P55-UD3
Image
Image
Image
o CPU Intel Core i5-750
Image
o CPU Cooler Scythe NINJA MINI rev B (SCMNJ-1100)
Image
Image
o Therm paste Arctic Silver Ceramique (2.5 g)
o RAM Kingston DDR3 4 GB 1600 MHz KHX1600C8D3K2/4GX (CL8)
Image
o SVGA Sapphire Radeon HD 4670
Image
o HDD Western Digital WD1001FALS
o HDD Western Digital WD1001FALS
o HDD Western Digital WD1001FALS
Image
The three WD1001FALS are set on RAID0. As you can see access time and transfer speed rocks!
Image
Image
Image
o CD-R - none
o CD-RW Sony AD-5240S-0B
Image
o Sound card onboard
o Speakers QuickShot QS-835
Image
o LAN onboard
o LAN Adaper SiS900 PCI
o Bracket to connect any SATA HDD at eSATA + external SATA input
Image
o CASE Lian li PC-K7B
Image
o PSU Fortron Blue Storm II 500
Image
Image
o Keyboard Benq i150 (keyboard from kitt Key+Mouse)
Image
o Mouse A4Tech SWW-48 (PS2)
Image
o UPS APC Back-UPS CS, 650VA/400W (connected on a Bandrige Power Cable and protection)
Image
o Monitor Horizon 2206SW (Meditec) - 22" LCD - LG panel A+
Image
o Scanner Hewlett-Packard ScanJet 2400
Image
o OS Microsoft Windows 7 Home Premium 64 bit English OEM
Image

- The completed system:
Image
Image
Image
Image
Image

BIOS Setup: J36_170x20
It took me a lot to understand that I have to enter MANUALLY all default values on BIOS, because on Gigabyte motherboards [AUTO] means really weird things (for example Vtt voltage left on AUTO was set by BIOS to almost 1.4v !!! when the default value is 1.100v and maximum specified by Intel is 1.21v). I am not sure about the other manufacturers but I would recommend to all Gigabyte users to write the default values manually on BIOS to protect their components on the long therm.
UPDATE: Looks like with the latest bios update, namely F8C (motherboard_bios_ga-p55-ud3_f8c.exe) these issues were corrected by Gigabyte. However, I still prefer to set my options manually rather then trusting AUTO.

o BIOS update using QFLASH and 1Gb USB stick. I noticed USB stick of 4Gb (FAT32 same as stick of 1gb) can't be detected by QFlash as a valid drive.

o BIOS settings:
Image
Image
Image
Image
Image
Image
Image
Image
CPU Vcore lowered to = 1.09375 and Vtt = 1.170. All perfect stable.
Image
Image
Image
Image
Image
Image
Image

o TurboBoost deactivated from BIOS since bumping the multiplier to x24 at a frequency of 170x24=4080 MHz, is a sure crash at these low voltages, I don’t even want to test that. All power saving features are explicitly ENABLED.
However if you are willing to increase Vcore to make it stable, TurboBoost is possible but I do not recommended it.

o System Memory Multiplier set at 8. This allow rising BCLK to 170 while keeping the memory frequency at only 1360 MHz at 7-7-7-20 timings. This allow me to set Performance Enhance on TURBO without causing instability. This apparently have an effect on Round Trip Latency (RTL) parameter and in my case the difference was clearly measurable in all windows applications. I pondered a lot between setting RAM at 1600 MHz with 8-8-8-20 and Performance Enhance on STANDARD as many users on different forums recommend, and the above settings. I underline again: RAM at 1360 MHz at 7-7-7-20 timings with Performance Enhance on TURBO -> windows "feel" much faster and everything is moving smoothly compared with setting RAM at 1600 MHz with 8-8-8-20 and Performance Enhance on STANDARD.
More about RTL here: http://www.anandtech.com/show/2869/6

- CPU-Z and Gigabyte Easy Tune 6 reports (CPU on idle and full load with Prime95)
Image
Image
Image
Image
Image
Image
Image
Image
Image

- POWER CONSUMPTION (including monitor and entire system connected to the UPS)
o idle (170x9=1530MHz) = 116W (CPU=idle, GPU=idle)
Image
o full load (170x20=3400MHz) = 180W (CPU=full load, GPU=idle)
Image
o full load CPU+GPU(170x20=3400MHz) = 232W (CPU=full load, GPU=full load) (Priome95+Furmark in same time)
Image

- CPU TEMPERATURES (recorded after 23 minutes of INTEL BURNTEST - ambient temp = 22ºC)
o idle (170x9=1530MHz) = 31ºC (the highest core)
o full load (170x20=3400MHz) = 60ºC (60,58,58,58 for each core after 23 min. of IntelBurnTest at Maximum stress level)
o Note1: INTEL BURNTEST is the most stressful test for PC and in real programs you can’t get temps this high even if you are using CPU at full load for hours.
o Note2: With this computer at 3.8GHz the speed (Gigaflops) on IntelBurnTest was 53? GFlops but the CPU temperature hit over 80ºC and power consumption increased to 310W. This is not for me, thank you very much.


- TEST1 = Memtest86+ v4.00 (MS DOS) => 3 hours memory testing on MS-DOS mode, 5 time coverage => 0 errors, PASSED

- TEST2 = MEMTEST (Windows)
4 instance running 234% coverage => 0 errors, PASSED
Image

- TEST3 = PRIME 95 running for 3 hours => 0 errors PASSED
Image

- TEST4 = INTEL BURNTEST V.2.3
Standard 1024Mb, 3 times => 0 errors PASSED
Maximum 3461Mb, 10 times => 0 errors PASSED (1392 sec.)
Image

- TEST5 = 3DMark 2006 (tests: select all, settings: 1680x1050, noAA, run each test 10 times, all the rest at default) => 0 errors PASSED
3DMark score 7425, SM2.0 = 2574, HDR/SM3.0ï‚« = 2973, CPU = 4964

- TEST6 = FURMARK v.1.8.2
o Benchmarking: fullscreen, 1680x1050, noAA, 600000ms (10 min.) => 0 errors PASSED
Average fps = 21 (min=18, max=33)
Image
GPU temp = 77ºC (normal) -> 78ºC (max.) (22ºC ambient)
o Stability test: windowed, 1024x768: 600 sec => 0 errors PASSED
Average fps = 32
GPU temp = 71ºC (normal) -> 73ºC (max.) (22ºC ambient)
Image

- TEST6 = PRIME95 + FURMARK Stability test (running both on same time)
FURMARK Stability test: windowed, 1024x768: 600 sec => 0 errors PASSED
o Average fps = 32
o GPU temp = 76ºC (max.) (25ºC ambient)

- TEST7 = LONG TIME STABILITY TEST USING VARIOUS APLICATIONS AND MULTITASKING.
A long time stability test is needed for LGA 1156 platforms, as were seen computers fully stable that pass all stress tests flawlessly and then crash at random. This seems to happen because of a too low VTT - V Core voltages on overclocking. Since I tried to keep voltages (and temperatures) to a minimum for my modest overclock, this test is perhaps the most important indication that the computer is FULLY STABLE.
Playing different games for several hours on different days => no crash, all stable, 0 errors, PASSED

- TEST8 = Apex DC++ and uTorrent
Downloading/Uploading with high speed => no crash, all stable, 0 errors, PASSED
Image
Image

- TEST9 = VIDEO CONVERTING
Converting 104 video files (4 converter instances to use CPU 100%), 687minutes.; average 396sec per file (6.60min) =>0 errors, PASSED

~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~
Possibile frequency increase for i5-750 in percentage:
i5-750 CPU freq. % increase over default value
Default freq. 2660 default
Moderate overclock 3400 27.82 %
High overclock 3800 42.90 %
Extreme overclock 4000 50.40 %

Some of the common i5-750 overclocking multipliers. Voltages are only for reference and potentially dangerous; DRAM voltage to be applied depend of your RAM modules. If you experience crashes it means the timings of the RAM are too tight or voltages must be increased further (possible to 1.50000 CPU Vcore and 1.45 VTT = on your own risk! Be aware that high voltages applied for extended period of time lead to accelerated component degradation and even burning!). If you get memory errors you need to relax the timings. I used in the past 7-7-7-20 for 166BCLK x 20CPU Vcore = 3320MHz CPU clock, at memory frequency = 1328MHz, and 8-8-8-20, 8-8-8-24, 9-9-9-24 for others.
Image
Image
Image
CPU Vcore lowered to = 1.09375 and Vtt = 1.170. All perfect stable.

Posted: Tue Jan 26, 2010 4:41 am
by frenchie
Hi,

What fans did you use ?
How loud is the GPU ?


Note : nice monitor :D

Fans

Posted: Tue Jan 26, 2010 5:04 am
by Tyr Antilles
For case cooling i used 3 12cm fans that came with Lian Li case (+ one 12 cm fan from PSU if you want to count that).
For CPU I used the default 8 cm fan of the Scythe NINJA MINI (exit) + one AMD fan from an older system of 7 cm or maybe less (in).

One problem arose from the RAM modules that blocked the 8Cm fan to be installed in front of the heatsink, so, I installed it on the rear while in front right over a RAM module I placed the smaller AMD fan (in). Both fans have variable speed so, the noise is not too loud. I am sure these two can be easily replaced with one or two quiet fans.

The GPU is indeed louder when is used at 100% but on idle the noise is imperceptible. Some noise can be clearly heard from HDD's when they are heavy used, so the system overall is acceptable loud for me.

The monitor having a TN panel, have a very poor viewing angle from the bottom, but besides that is very good overall. PRAD have a very good review for it here http://www.prad.de/en/monitore/review/2 ... troduction and I can say mine is exactly like they describe it.
P.S. I decided to buy this one because of the girls :P

Re: Fans

Posted: Tue Jan 26, 2010 12:56 pm
by hybrid2d4x4
Tyr Antilles wrote: P.S. I decided to buy this one because of the girls :P
I knew it! lol

Nice system. How much power does the monitor draw?

Posted: Tue Jan 26, 2010 2:34 pm
by CA_Steve
....................
- CPU TEMPERATURES (ambient temp = 23ºC)
o idle (166x9=1494MHz) = 20ºC (20, 16, 16, 16 for each core)
o full load (166x20=3320MHz) = 53ºC (53, 53, 51, 53 for each core)
....................

Nice build. Might want to review the calibration of your temp monitoring s/w as it's impossible for the cpu idle temps to be below ambient. Your actual load temps are probably 7-10C higher than shown.

Posted: Tue Jan 26, 2010 5:58 pm
by flyingsherpa
I can't believe you posted that many pics, yet didn't include any of the completed system :lol:

Re: System

Posted: Tue Jan 26, 2010 6:32 pm
by Tyr Antilles
hybrid2d4x4 wrote: Nice system. How much power does the monitor draw?
I didn't tested yet but should be less then 40 W at the brightness/contrast settings I use it.
CA_Steve wrote: Nice build. Might want to review the calibration of your temp monitoring s/w as it's impossible for the cpu idle temps to be below ambient. Your actual load temps are probably 7-10C higher than shown.
You have a point Steve, but the temperatures are calibrated right, I used for calibration HW Monitor program and Easy Tune 6 (that came with Gigabyte board) and they both report same values. To be sure I used in a higher overclock the beeping at 70 gr C option from BIOS, and right when my sensors from Windows reached 70 gr C, the overheating alarm set in BIOS started to beep, so I guess they are calibrated right. Only Coretemp utility had to be calibrated exactly at -10 gr C to match the other reported values of HWMonitor and Easy Tune 6.
flyingsherpa wrote: I can't believe you posted that many pics, yet didn't include any of the completed system
Haha, you are absolutely right. I shall take some as soon I will have a lil time :D

Posted: Tue Jan 26, 2010 8:58 pm
by jtcb
I see you use Kingston ram. Is there a particular reason you use that brand? I am also considering Kingston because they are the 'brand' for memory. I don't plan to overclock my i7 860. I wondering which of these are good for me? What is XMP?

http://www.kingston.com/hyperx/corei5/

memory choice

Posted: Tue Jan 26, 2010 10:39 pm
by Tyr Antilles
jtcb

I decided to use Kingston mem. because I read some favorable reviews about them on different forums and because it was the cheapest good brand memory i found.

XMP (Xtrem Memory Profile) is a predefined overclocked profile embedded on the memory. When you set XMP on BIOS the memory is automatically overclocked at settings recommended by the producer. I choose to not use it even if my memory support it because I decided that is better to overclock the CPU with no risk at all and have better timings for memory (currently 7-7-7-20). The performance increase is a lot higher then if I only used XMP.

Good memory for you depends of your preference and budget. I would recommend some on the 1600MHz range, maybe you will want to overclock it a little later, this was my choice and I am happy with it. If you plan to overclock (CPU+MEM) like mad, go for the highest 2133MHz, but since you already said you are not into the overclocking, it really don't matter, you can just pick up the cheapest.

Posted: Tue Jan 26, 2010 11:29 pm
by jtcb
Thanks Tyr. Can you check if this is right for me?

http://www.ncix.com/products/?sku=46053

OR can you choose one that best for me under DDR3?

http://www.ncix.com/search/?categoryid= ... ton+HyperX

mem

Posted: Wed Jan 27, 2010 12:38 am
by Tyr Antilles
jtcb

Sorry, I can't choose, i have KHX1600C8D3K2/4GX a CL8 rated mem. http://www.ncix.com/products/?sku=44921 ... e=Kingston; the one you pointed is OK but is only CL9

On the other side this one looks really good:
http://www.ncix.com/products/?sku=45327 ... e=Kingston
is slightly cheaper and faster, but Input Voltage = 1.9V. :? I don't know if is not too much, but probably you can run it at 1600 with CL8, 7 or even lower without increasing voltages that much.

Good luck with your choice.

Re: System

Posted: Wed Jan 27, 2010 9:12 am
by CA_Steve
Tyr Antilles wrote:
CA_Steve wrote: Nice build. Might want to review the calibration of your temp monitoring s/w as it's impossible for the cpu idle temps to be below ambient. Your actual load temps are probably 7-10C higher than shown.
You have a point Steve, but the temperatures are calibrated right, I used for calibration HW Monitor program and Easy Tune 6 (that came with Gigabyte board) and they both report same values. To be sure I used in a higher overclock the beeping at 70 gr C option from BIOS, and right when my sensors from Windows reached 70 gr C, the overheating alarm set in BIOS started to beep, so I guess they are calibrated right. Only Coretemp utility had to be calibrated exactly at -10 gr C to match the other reported values of HWMonitor and Easy Tune 6.
Unless the laws of physics have been altered in your home, something has to give. Because, if your numbers hold true, when you turn your PC from idle to off, the core temps will go UP to meet ambient temp. :D

So, maybe what you are reporting as ambient temp is really some other monitored temp, like the mobo temp and not your room temp?

By definition, unless you are using refrigeration, ambient has to be the lowest reported temperature. Otherwise, you are absorbing heat and not generating heat.

temps

Posted: Wed Jan 27, 2010 9:53 am
by Tyr Antilles
You are right again Steve. I was looking on the BIOS at startup and the reported temps (on idle of course) are around 25 - 30 C, then in Windows are reported the same WHEN THE CPU IS SET AT X20 MULTIPLIER. I can only think that entering in a power save state like X9 multiplier, mess somehow the temp sensors, I don't understand how, since the temps in full load matches the ones in BIOS.

You can see in this screenshot the temps monitored with Core Temp (on taskbar) and with Easy Tune 6 (the utility provided by Gigabyte with the mobo), when at that time the ambient temperature was 24 C. Also the system temp is reported as 32 by the same Easy Tune 6... strange.
Image

Posted: Wed Jan 27, 2010 11:13 am
by ntavlas
This is a system designed to run on hot ambient up to 40ºC
I didn`t know it got that hot in London :p

I would give realtemp a try, it is one of the most trusted programs in reporting cpu temps.

It would be great if you could upload some pictures when you have the time.
<---Big fan of Lian Li cases and ninja heatsinks.

Posted: Wed Jan 27, 2010 3:04 pm
by Cistron
ntavlas wrote:
This is a system designed to run on hot ambient up to 40ºC
I didn`t know it got that hot in London :p
Every morning I pray to the weathergods, but then it's once again grey cloud. I somewhat think I'm caught up in Groundhog Day.

location and temps

Posted: Wed Jan 27, 2010 7:41 pm
by Tyr Antilles
Haha, you got me, I am not in London. This remind me of that old joke...
A Frenchman and an Englishman meet on a tavern. After two months in London the French complains furious: "I have had enough, mist, mist and mist again, in this country is only mist!" The Englishman hurt in it's patriotic pride reply annoyed: "Of course not! When is raining is never mist..."

I will try to get those system photos a.s.a.p. but that means I need to shut down the comp. and I can't do that now... is running stright of almost 5 days already...

Thanks for the realtemp ideea, I will try it right away ....
...... several minutes later .....
I guess you were right guys, realtemp reported values 10 G higher :cry:
I think I shall belive it instead of EasyTune and HWMonitor ... still is hard to believe that Gigabyte provided a tool that shows temperatures 10 G lower then the real ones :evil:
I adjusted core temp offset to 0 and now it reports the same values as realtemp.

Idle
Image

Load
Image

After 45 mins. of Prime95 the temperature stabilized at 62G Thank you very much.

Re: Fans

Posted: Wed Jan 27, 2010 10:19 pm
by Spirit02
Tyr Antilles wrote: P.S. I decided to buy this one because of the girls :P
Ah Adriana Lima....

I have that wallpaper ;)

EasyTune 6

Posted: Thu Jan 28, 2010 12:59 am
by Tyr Antilles
After I did some research on net I think I need to edit the first post again. It looks like RealTemp and CoreTemp are right and EasyTune6 is indeed wrong and this problem seems really common with this utility.

wazza300 Senior Member on some other forum said: et6 takes temps from the motherboard sensors and the heatsink, realtemp takes temps from the cpu core sensors inside the chip itself hence core temp, always go off the coretemp as its the part that gets hottest before it gets cooled by the heatsink and fan

here are some links I discovered with infos about this problem:
http://forums.tweaktown.com/f69/easy-tu ... res-37259/
http://www.tomshardware.co.uk/forum/246 ... ifferences
http://www.hardforum.com/showthread.php?t=1469948
http://www.xbitlabs.com/discussion/5694.html
http://forums.anandtech.com/archive/ind ... 76020.html
http://forums.bit-tech.net/showthread.php?p=2195106
http://techreport.com/forums/viewtopic.php?f=30&t=65528

And I was so happy with my low temps, now 62 C on full load is too much.

Posted: Thu Jan 28, 2010 1:00 am
by Tyr Antilles
[Edit] strange that was posted twice.

Posted: Thu Jan 28, 2010 8:00 am
by CA_Steve
-->And I was so happy with my low temps, now 62 C on full load is too much.<---

Naw, it's just more in line with what's expected. Remember, a lot of the tests you were running are stress tests - conditions that you'll never see in real life apps.

Posted: Thu Jan 28, 2010 10:04 am
by diver
Rather than a static overclock, you might try one with all of the turbo, speed step and C states enabled. Just set Vcore to "Normal". that will give you stock voltage which is sure to work at a bclk of 146. You probably should make sure LLC is disabled for this one. For more performance, try 160. Use a +.0125 DVID offset, if you need it. This latter setting will do 3660 Mhz on 4 cores and 3480 Mhz on one or two cores. These should be compatible with quiet cooling, although worst case core temperatures will be +45C over ambient with the Minja. I have also had good luck with a static overclock of 173 bclk and Vcore of 1.16875 (bios) with LLC enabled. All of this on a Gigabyte GA-P55-UD2.

CPU load

Posted: Fri Jan 29, 2010 12:22 am
by Tyr Antilles
CA_Steve wrote:Naw, it's just more in line with what's expected. Remember, a lot of the tests you were running are stress tests - conditions that you'll never see in real life apps.
Unfortunately for video encoding and other image and video rendering stuff I am working with, the CPU will need to do just that, 100% load sometimes for several days in a row.
diver wrote:Rather than a static overclock, you might try one with all of the turbo, speed step and C states enabled. Just set Vcore to "Normal". that will give you stock voltage which is sure to work at a bclk of 146. You probably should make sure LLC is disabled for this one. For more performance, try 160.
I already tried that, is working but the power consumption and especially the temperatures are rising to a level I don't feel comfortable. I will rather stay with a cool and quiet machine overclocked moderately for everyday use in hot environment.

----------
Also, I lowered the memory voltage a little bit and the temperature in full load decreased with 2C on the second test to 62C. Memory was still fully stable, maybe I can try to lower it even more.
I am planning to take today some photos of the system to post them here.

Posted: Fri Jan 29, 2010 8:26 am
by CA_Steve
Nearly 40C rise does seem a tad high for the Minja in this config. Seems like it should be in the 30C range...or less. Perhaps the Minja isn't mounted well.

Are you using the stock push pin mounting or did you get a bolt-thru kit for the Minja? Sometimes, the push pin mounts can give less than stellar performance.

Have you tried running the video apps for an hour and see what your actual temps will be?

What kind of stress test temps do you get when you run the CPU at stock speeds?


Regarding the GPU, you should be able to use Catalyst Control Center to set the fan speed to a lower rpm - just monitor temps when you try it.

Ninja not mounted well

Posted: Fri Jan 29, 2010 11:18 am
by Tyr Antilles
CA_Steve wrote:Perhaps the Minja isn't mounted well.
I started to think at this myself, I had difficulty to mount ninja because of the Intel type pins that kept slipping. Those push pins are a pain. Maybe i should remove it and try to mount it again.

Bolt-thru kit for the Ninja? I didn't know that such thing exist, i just used what was in the package.

Yes, I did run Prime95 for one night (6h) and other night Memtest in 4 instances in same time to bring CPU at 100%. The max temps were the ones reported - from idle at 32C - to max full load 64C (after the new measurements with RealTemp).

About Catalyst, yes I use it, when I don't use video acceleration the GPU is silent, when I do I don't care about the noise anyway. I usually want the cooler on AUTO.

Here are finally some photos with my system:
[EDIT on 2010.02.01: I CHANGED THE CPU COOLER FAN, LOOK ON THE FIRST POST FOR THE UPDATED PHOTOS]
Image
Image
Image
Image
Image
Image
Image

Re: Ninja not mounted well

Posted: Fri Jan 29, 2010 1:30 pm
by CA_Steve
Tyr Antilles wrote: Bolt-thru kit for the Ninja? I didn't know that such thing exist, i just used what was in the package.

Yes, I did run Prime95 for one night (6h) and other night Memtest in 4 instances in same time to bring CPU at 100%. The max temps were the ones reported - from idle at 32C - to max full load 64C (after the new measurements with RealTemp).
I used a backplate retention kit with my Ninja on a socket 775. Removed the push pins and replaced them with the spring loaded bolts in the kit + the backplate. It improved my core temps by 5C. Perhaps one of the socket 1156 retention kitswould work with your Mini Ninja.

Regarding apps and temps. I think we are missing each others points. I was wondering two separate things:

1) Have you tried testing the CPU temps as above while running stock speeds to see how much impact the OC'ing is having on temps?

2) Have you tried measuring the temps when running the actual video encoding/etc. apps instead of these stress tests? (At stock and at OC) Very few apps ever approach the same level of CPU stressing as running Prime95, Orthos, etc. Your apps may run 10C cooler making all of this worry moot.

Posted: Fri Jan 29, 2010 4:24 pm
by ntavlas
Have you tried further reducing cpu voltage? Or are you out of headroom?

A bolt through kit should help too. A 5 degree improvement is nothing to sniff at though I can`t confirm it first hand (never used the push pins on my ninja to compare).

Having said that, I don`t think 62 is unsafe even if it`s sustained for long periods of time.

Posted: Fri Jan 29, 2010 6:15 pm
by CA_Steve
ntavlas wrote:Having said that, I don`t think 62 is unsafe even if it`s sustained for long periods of time.
It's when his ambient goes from 24C to 40C and the subsequent cpu rise from 62 to 78C that might be a concern.

Re: Ninja not mounted well

Posted: Fri Jan 29, 2010 6:38 pm
by Tyr Antilles
CA_Steve wrote: socket 1156 retention kits

1) Have you tried testing the CPU temps as above while running stock speeds to see how much impact the OC'ing is having on temps?

2) Have you tried measuring the temps when running the actual video encoding/etc. apps?
About a bolt through kit - I will definitely look for something like this, thank you guys.

1. No, I never tested it at stock speeds because I had some issues with the PSU crackling noise when running at stock. But I did test it at a higher overclock of 3800MHz and the temperatures were a disaster, take a look here I stopped it because the temps kept rising:
Image

2. Yes, the temperature on video encoding and rendering when the CPU is used 100% was one C lower then the Prime95 temp, so, no big difference at all.
ntavlas wrote: 1. "Have you tried further reducing cpu voltage? Or are you out of headroom?"

2. "I don`t think 62 is unsafe even if it`s sustained for long periods of time"
1. Out of headroom, when I try to lower it even more it looks stable at first, but after a random amount of time, sometimes after a day, it gets a windows crash even if tests are successful. However crash never happen with CPU Vcore = 1.13125V, so I guess this is the limit.

2. It is too hot for an ambient temp of 25C, sometimes the ambient can rise as high as 35 - 38C... so, if we are adding 10-13C...

----------------------------------------------------------
Other thing: I decreased DRAM voltage to 1.500V and the system seems still fully stable. The temp decreased 1C as well. Looks like this Kingston memory needs the specified voltage of 1.65 only for 1600MHz clock speed or higher, I noticed that compared to [email protected], at [email protected] the temperature of the mem. heat-sinks lowered dramatically, they are almost cold now.
----------------------------------------------------------
I am very grateful for your advices, guys, please keep posting if you can :)

Reinstalled the CPU cooler

Posted: Sun Jan 31, 2010 11:51 pm
by Tyr Antilles
IMPORTANT! I changed the first post of this thread in order to reflect the latest hardware and configuration changes I made to my system, including the updated tests, voltages and temperature readings.

Concerned about the high temps I decided to reinstall the CPU cooler. So, I removed the motherboard and I discovered that one of the Intel plastic push-pins jumped from it’s place. Grrr. I reapplied a very even and thin layer of compound Arctic Silver Ceramique and carefully reinstalled the cooler. This time everything went smoothly and I changed the order of the fans too, removing the AMD fan. Now is much better, temperatures never go over 60C at a 25C ambient.

OLD post from 2010.02.01. First post is updated!

Posted: Tue May 04, 2010 5:27 am
by Tyr Antilles
[This is the old thread posted on 2010.02.01. I updated the first thread of this post in order to reflect the latest configuration changes. Please read that one. I kept this as reference only. Replaced thumbnails with links for faster loading]


Core i5-750 system - mild overclocking - cool and low power PC

I just built an i5-750 based system and after a lot of searching I managed to overclock it a little without increasing voltages and keeping temperatures as low I can. This is a system designed to run on hot ambient up to 40ºC. Since it was done anyway I thought to share it with everyone, maybe someone will find these indications useful. I am not an experienced system builder, so if you find anything out of place with this built or if you have any comments please feel free to post replays.

The noise level is very low, from the three 12cm fans that came with Lian-Li case, and the CPU fan don’t even counts. The fan from video card is audible only in 3D applications and the loudest components of the system are the three HDD’s. Overall the system is fairly quiet and under the desk where I keep it is totally not disturbing.


- HARDWARE:
o MB Gigabyte P55-UD3
http://www.postimage.org/image.php?v=gxyADzi
http://www.postimage.org/image.php?v=gxyAG3r
http://www.postimage.org/image.php?v=gxyAIyA
o CPU Intel Core i5-750
http://www.postimage.org/image.php?v=gxyAL2J
o CPU Cooler Scythe NINJA MINI rev B (SCMNJ-1100)
http://www.postimage.org/image.php?v=gxyANxS
http://www.postimage.org/image.php?v=gxyAQ20
o Therm paste Arctic Silver Ceramique (2.5 g)
o RAM Kingston DDR3 4 GB 1600 MHz KHX1600C8D3K2/4GX (CL8)
http://www.postimage.org/image.php?v=gxyAXwr
o SVGA Sapphire Radeon HD 4670
http://www.postimage.org/image.php?v=gxyA_0A
o HDD boot Western Digital WD1001FALS
http://www.postimage.org/image.php?v=gxyB1vJ
o HDD Western Digital WD1001FALS
o HDD Western Digital WD1001FALS
o CD-R -
o CD-RW Sony AD-5240S-0B
http://www.postimage.org/image.php?v=gxyB3_S
o Sound card onboard
o Speakers QuickShot QS-835
http://www.postimage.org/image.php?v=gxyB6v0
o LAN onboard
o CASE Lian li PC-K7B
http://www.postimage.org/image.php?v=gxyB8_9
o PSU Fortron Blue Storm II 500
http://www.postimage.org/image.php?v=gxyBbui
o Keyboard Benq i150 (keyboard from kitt Key+Mouse)
http://www.postimage.org/image.php?v=gxyBdZr
o Mouse A4Tech SWW-48 (PS2)
http://www.postimage.org/image.php?v=gxyBgtA
o UPS APC Back-UPS CS, 650VA/400W (connected on a Bandrige Power Cable and protection)
http://www.postimage.org/image.php?v=gxyBiYJ
o Monitor Horizon 2206SW - 22" LCD
http://www.postimage.org/image.php?v=gxyBlsS
o Scanner Hewlett-Packard ScanJet 2400
http://www.postimage.org/image.php?v=gxyBnY0
o OS Microsoft Windows 7 Home Premium 64 bit English OEM
http://www.postimage.org/image.php?v=gxyBqs9

- The completed system:
http://www.postimage.org/image.php?v=aVfMXt0
http://www.postimage.org/image.php?v=aVfMZY9
http://www.postimage.org/image.php?v=aVfN1si
http://www.postimage.org/image.php?v=aVfN3Xr
http://www.postimage.org/image.php?v=aVfN6rA
http://www.postimage.org/image.php?v=aVfN8WJ
http://www.postimage.org/image.php?v=aVfNbqS

- BIOS Setup:
http://www.postimage.org/image.php?v=aVcHKMi
http://www.postimage.org/image.php?v=aVcHNgr
http://www.postimage.org/image.php?v=aVcHPLA
http://www.postimage.org/image.php?v=aVcHSfJ
http://www.postimage.org/image.php?v=aVcHXf0
http://www.postimage.org/image.php?v=aVcI1ei
http://www.postimage.org/image.php?v=aVcI3Jr
http://www.postimage.org/image.php?v=aVcI6dA
http://www.postimage.org/image.php?v=aVcI8IJ
http://www.postimage.org/image.php?v=gxyC3mr

- CPU-Z
(CPU on full load with Prime95)
http://www.postimage.org/image.php?v=gxyC5RA
(CPU at idle)
http://www.postimage.org/image.php?v=gxyC8lJ
http://www.postimage.org/image.php?v=gxyCaQS
http://www.postimage.org/image.php?v=gxyCdl0
http://www.postimage.org/image.php?v=gxyCfQ9

- POWER CONSUMPTION (including monitor and entire system connected to the UPS)
o idle (166x9=1494MHz) = 116W (CPU=idle, GPU=idle)
http://www.postimage.org/image.php?v=gxyCiki

o full load (166x20=3320MHz) = 204 - 212W (CPU=full load, GPU=idle)
http://www.postimage.org/image.php?v=gxyCkPr

o full load CPU+GPU(166x20=3320MHz) = 260 - 264W (CPU=full load, GPU=full load)
http://www.postimage.org/image.php?v=gxyCnjA

o turbo boost deactivated from BIOS as induces instability with these low voltages.

- CPU TEMPERATURES (ambient temp = 26ºC)
o idle (166x9=1494MHz) = 28ºC (28, 25, 25, 25 for each core)
o full load (166x20=3320MHz) = 61ºC (59, 61, 59, 59 for each core after 22 min. of IntelBurn Test at Maximum stress level)


- TEST1 = MEMTEST
4 instance running 1900% coverage (5 hours) => 0 errors, PASSED
http://www.postimage.org/image.php?v=gxyCsiS

- TEST2 = PRIME 95 running for 1.5 hours => 0 errors PASSED
http://www.postimage.org/image.php?v=gxyCuO0

- TEST3 = INTEL BURNTEST V.2.3
ï‚« Standard 1024Mb, 3 times => 0 errors PASSED
ï‚« Maximum 3259Mb, 10 times => 0 errors PASSED (1323 sec.)
http://www.postimage.org/image.php?v=Ts12RtEi
http://www.postimage.org/image.php?v=Ts12Rw8r
o TEMPERATURES (ambient 26ºC): max. temp. on full load = 61ºC (59, 61, 59, 59 for each core after 22 min. of IntelBurn Test at Maximum stress level).
o Note1: that this is the most stressful test for PC and in real programs you can’t get temps this high even if you are using CPU at full load for hours.
o Note2: With this computer at 3.8GHz the speed (Gigaflops) on IntelBurnTest was 53? GFlops but the CPU temperature hit over 80ºC and power consumption increased to 310W. This is not for me, thank you very much.

- TEST4 = 3DMark 2006 (tests: select all, settings: 1680x1050, noAA, run each test 10 times, all the rest at default) => 0 errors PASSED
ï‚« 3DMark score 7425, SM2.0 = 2574, HDR/SM3.0 = 2973, CPU = 4964
http://www.postimage.org/image.php?v=gxyCEMA

- TEST5 = FURMARK
ï‚« Benchmarking: fullscreen, 1680x1050, noAA, 300000ms (5 min.) => 0 errors PASSED
http://www.postimage.org/image.php?v=gxyCHgJ
ï‚« Stability test: windowed, 1024x768: 600 sec (166x20 = 3320MHz) => 0 errors PASSED
o Average fps = 39
o GPU temp = 45ºC (min.) -> 77ºC (max.)
http://www.postimage.org/image.php?v=gxyCJLS

- TEST6 = PRIME95 + FURMARK Stability test (running both on same time)
ï‚« FURMARK Stability test: windowed, 1024x768: 600 sec (166x20 = 3320MHz) => 0 errors PASSED
o Average fps = 39
o GPU temp = 78ºC (max.)
http://www.postimage.org/image.php?v=gxyCTKr

- TEST7 = LONG TIME STABILITY TEST USING VARIOUS APLICATIONS AND MULTITASKING (a long time stability test is needed for LGA 1156 platforms, as were seen computers fully stable that pass all stress tests flawlessly and then crash at random. This seems to happen because of a too low VTT - V Core voltages on overclocking. Since I tried to keep voltages (and temperatures) to a minimum for my modest overclock, this test is perhaps the most important indication that the computer is FULLY STABLE.
ï‚« Playing CIV4 for several hours on different days => no crash, all stable, 0 errors, PASSED
http://www.postimage.org/image.php?v=gxyCYJJ

- TEST8 = Apex DC++
ï‚« Downloading/Uploading with high speed => no crash, all stable, 0 errors, PASSED
http://www.postimage.org/image.php?v=gxyD0dS

- TEST9 = VIDEO CONVERTING
ï‚« Converting 79 video files (4 converter instances to use CPU 100%), 7h 50min.; average 357sec per file =>0 errors, PASSED

~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~
Some of the common I5-750 overclocking multipliers. Voltages are only for reference and potentially dangerous; DRAM voltage depend of RAM modules. If you experience crashes it means the voltages must be increased further (possible to 1.50000 CPU Vcore and 1.45 VTT = on your own risk!). If you get memory errors you need to relax the timings. I have 7-7-7-20 for 166BCLK x 20CPU Vcore = 3320MHz CPU clock, at memory frequency = 1328MHz. I used 8-8-8-20, 8-8-8-24, 9-9-9-24 for the others.

http://www.postimage.org/image.php?v=aVfNdW0
http://www.postimage.org/image.php?v=aVfNgq9
http://www.postimage.org/image.php?v=aVfNiVi
(with light blue = default settings for Core i5-750 CPU; with light green, the settings I found suitable for my purpose)
~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~

[This was the old thread posted on 2010.02.01. I updated the first thread of this post in order to reflect the latest configuration changes. Please read that one.]