Welcome Guest ( Log In | Register )

Bump Topic Add Reply RSS Feed
5 Pages  1 2 3 > » Bottom

Outline · [ Standard ] · Linear+

 Overclocking E7XXX Series

views
     
JoBigShow
post Aug 24 2009, 04:02 AM

Ω Ξ Ψ Σ Д Ӫ Ж Π ✿
*****
Senior Member
969 posts

Joined: Jul 2009
From: ɹndɯnן ɐןɐnʞ
Using E7400 400 x 9 @???voltage (still in the process of finding the right stable volts)

Regarding Prime95, when using the torture test, which one should I choose to find the right stability for the cpu's voltage?:

1. Small FTT's
2. In-place Large FTT's
3. Blend

I'm pretty bump coz I got errors when running 'Blend' up to 1.42v, I'm informed that 1.3625v is recommended as maximum increments, and 1.4v+ is really pushing it. Is it just that my chip is bad that i gotta push above 1.45v+ to be stable?

In 'small FTT's' i havent encounter any error so far, is 'small FFt's' the one I should be running?

icon_question.gif


JoBigShow
post Aug 27 2009, 07:01 AM

Ω Ξ Ψ Σ Д Ӫ Ж Π ✿
*****
Senior Member
969 posts

Joined: Jul 2009
From: ɹndɯnן ɐןɐnʞ
QUOTE(dun_panic @ Aug 26 2009, 12:44 AM)
just run all the test to make sure ur system r rock stable, better if u run it 24hours
*
How do you know when u have enough voltage or too much voltage? at the moment I'm within 1.34v-1.3625v@ 3.6ghz, I hope that's adequate enough. Do you guys know whats the territory (in terms of voltage) for 4.0ghz for E7400? Hope its doesn't have to reach up to 1.41v...I've read somewhere some guy rev up his e7400 up to 1.5v+ @4.0ghz, that's just plain ridiculous...
JoBigShow
post Aug 27 2009, 06:13 PM

Ω Ξ Ψ Σ Д Ӫ Ж Π ✿
*****
Senior Member
969 posts

Joined: Jul 2009
From: ɹndɯnן ɐןɐnʞ
QUOTE(8tvt @ Aug 27 2009, 09:05 AM)
not ridiculous.. can boot on 4ghz between stable..is different..
and not all proc are same..
*
Even though its a RM400 chip doesn't mean we should blow it up. rclxms.gif
Different processors are unique which makes OC'ing fun i guess.
I think my chip aint that bad coz I dont need to juiced it up to 1.45v+ to get stable @ 4.0ghz. For now...

JoBigShow
post Aug 27 2009, 08:20 PM

Ω Ξ Ψ Σ Д Ӫ Ж Π ✿
*****
Senior Member
969 posts

Joined: Jul 2009
From: ɹndɯnן ɐןɐnʞ
QUOTE(kayQay @ Aug 27 2009, 11:18 AM)
I have same question.....which one to choose
1. Small FTT's
2. In-place Large FTT's
3. Blend

Any otai here can answer ? biggrin.gif
*
In 3 days of hard research on google, sweat.gif I learn that 'Blend' is the more suitable candidate for trying to find the stability in your system becoz its more practical. This is bcoz 'Blend' (stressed both cpu and ram) works how PC should work, by applying both usage of CPU and RAM simultaneously.

Small FFT's is a more theoretical approach if u want to know your proc more closely/intimately. (used a lot in OC contests)

Large FFT's is also theoratical if u wanna know the maximum level your whole system can endure, which stressed mostly everything in your system like PSU, CPU, RAM, capacitors of the mobo, etc. Beware, this one heats the pc like hell, my cpu temp reached up to 75c+ on load even though i have aftermarket cooler: Artic Cooling Freezer Xtreme. Large FFT's is the favourite of hardware manufactures to test their products to the max.

hope anyone else can verify me on this... nod.gif


JoBigShow
post Aug 28 2009, 08:33 AM

Ω Ξ Ψ Σ Д Ӫ Ж Π ✿
*****
Senior Member
969 posts

Joined: Jul 2009
From: ɹndɯnן ɐןɐnʞ
QUOTE(farscope @ Aug 28 2009, 12:21 AM)
how does this compare to the simple LinX?
*
LynX measures the CPU performance in Gflops. (It's something i havent heard of apart from the more familiar Lucas-Lehmer test.)
This pretty popular among russian overclockers.
LynX stressed the cpu way more than Orthos and Prime95.
It is claimed that LnyX saves more time bcoz 1 hour of LynX stressed test is equivalent to 6 hours of stressed test performed by prime95.
Bcoz of this, it is known that LynX raised up more higher temps than u get when running 'Large FTT's' on prime95. e.g 10c-15c more!
So u have to be careful when leaving u'r PC unattended when running this software.
(Large FTT's on prime95 already makes my cpu really hot, i will expect this software to raise my cpu temps up to 80c, even with my Artic Cooling Freezer Xtreme.) rclxub.gif

Here's a thread on Lynx from the original creator:

http://www.xtremesystems.org/forums/showthread.php?t=201670

I'm gonna try this 'linpack interface' soon...


JoBigShow
post Aug 28 2009, 06:58 PM

Ω Ξ Ψ Σ Д Ӫ Ж Π ✿
*****
Senior Member
969 posts

Joined: Jul 2009
From: ɹndɯnן ɐןɐnʞ
QUOTE(farscope @ Aug 28 2009, 10:55 AM)
wow.. is it? not aware of tat.. I din use prime95 cz i noe its time consuming and straight using LinX..

I stress my E7400 @3.2Ghz for about half an hour and hit the max temperature of 58°C at night..
*
Maybe those people who got their temps very high because:
1. using stock cooler.
2: OC more than 3.6ghz on C2D
3. OC'ing Quad proc up to 3.4ghz+
4. Set their Vcore more than 1.45v+ (Increase in temp relevantly due to increase in voltage/vcore.)

Your e7400@3.2ghz doesn't get really hot bcoz you still can use stock vcore @ 3.2ghz. My E7400 does.
It's when u OC up to 3.6ghz+ is where u'r in business. nod.gif
I used set my vcore on 'AUTO' when i was still OC'ing @ 3.5ghz (334 x 10.5) a month ago.
(now maintain around 1.3625v @ 3.6ghz~3.8ghz, can topup to 4.0ghz+ if i want, but somehow i luv my proc too much to push it that high)

Have you encounter any bugs/problem in LynX? Unless u have the latest version: LinX 0.6.2.1
(I will download LinX soon to try it out later)

Nywayz, there's a thread on LYN about LinX:

http://forum.lowyat.net/topic/892376
JoBigShow
post Aug 28 2009, 07:18 PM

Ω Ξ Ψ Σ Д Ӫ Ж Π ✿
*****
Senior Member
969 posts

Joined: Jul 2009
From: ɹndɯnן ɐןɐnʞ
QUOTE(8tvt @ Aug 28 2009, 09:13 AM)
i got 90C..
*
90c rclxub.gif rclxub.gif rclxub.gif
Buying a proc at the shop really like playing blackjack or lottery,
its all starts when the salesperson picks up your proc randomly among other processors on the shelf.
It seems in this game of "processor lottery", u looked unlucky bro...
Statistically-wise, some say the probability to get the 'unlucky' chip from Intel is less than 2/10.
I read some people(usually highly superstitious geeks) up north in mainland China will choose a proc by their serial number,
apparently they want the s/n which has a lot of '8's in it to avoid getting the 'unlucky' chip. You tell me...
JoBigShow
post Aug 28 2009, 08:11 PM

Ω Ξ Ψ Σ Д Ӫ Ж Π ✿
*****
Senior Member
969 posts

Joined: Jul 2009
From: ɹndɯnן ɐןɐnʞ
QUOTE(farscope @ Aug 28 2009, 07:04 PM)
yup.. i OC to 3.2Ghz and stop there last time bcz i was still using stock cooler.. and yes @ stock Vcore..

now i'm planning to push it to 3.6Ghz when i'm free.. not going for 4.0Ghz anyway..

can share yr settings for 3.6Ghz?

i just downloaded 0.6.2 today and will giv it a try later tonite..
I'm using 0.5.9 before this and never encounter any bugs so far.. (or i din notice it  sweat.gif )
*
Yes, @ 3.6ghz:
400 x 9 fsb; 1.336v~1.344v
IDLE : core0 = 31c core1 = 34c
LOAD: core0 = about 60c core1 = about 58c
No OC on Dram bcoz the increase is minor in terms of performance.
(but sometimes I increase my 800mhz by 20mhz and 400fsb by 10 to keep that 1:1 ratio, e.g 410 x 8.5 fsb @820mhz DDR2 =3.485ghz,
fairly stable depends on the mobo and cpu)

Make sure u disable most of these settings when OC'ing the cpu:
Intel® Virtualization tech, C1E, and more importantly disable Intel SpeedStep ™ tech

And of course I got aftermarket cooler: AC Freezer Xtreme

I recommend that 1.336v~1.344 is enough for 3.6ghz on a E7400 chip. If not, then just increase OR decrease in increments.
(Make sure it's 1.336v-1344v in CPU-Z, in BIOS u might need to set the vcore slightly above 1.336v)
Remember this is only my opinion based on my e7400. I hope your E7400 is not that different.

Addtional info:
@3.8ghz
400 x 9.5 fsb; 1.352v~1.3625v

@4ghz
400x10 fsb; 1.3600v ~ 1.3625v (i try not to go above 1.3625v since its not recommended by Intel when OC on 45nm chips)
Even at 1.3625v@4.0ghz its seems quite stable

This post has been edited by JoBigShow: Aug 29 2009, 04:08 AM
JoBigShow
post Aug 28 2009, 08:48 PM

Ω Ξ Ψ Σ Д Ӫ Ж Π ✿
*****
Senior Member
969 posts

Joined: Jul 2009
From: ɹndɯnן ɐןɐnʞ
QUOTE(nidzamputra @ Aug 28 2009, 08:13 PM)
stress about 5 minute..then auto shutdown...before stress idle 40'C

normally i'm using Prime95 to stress my proc...maximum 75'C
*
The Thor's Hammer is Xiggy's finest cooler, and u got auto shutdown from reaching TJ max?!!
Man I start to think LinX is somewhat of a monster.. shocking.gif

Probably Linx is equivalent to Crysis in comparison to gaming...in the form of cpu stress test hmm.gif
maybe i gotta plan for watercooling for this test brows.gif
JoBigShow
post Aug 28 2009, 09:54 PM

Ω Ξ Ψ Σ Д Ӫ Ж Π ✿
*****
Senior Member
969 posts

Joined: Jul 2009
From: ɹndɯnן ɐןɐnʞ
QUOTE(nidzamputra @ Aug 28 2009, 08:55 PM)
maybe my mobo bios setting...i enabled auto shutdown when cpu reach 90'C in bios..by default it's set 85'C

actually i wanna to try 2nd hand CoolIT Domino ALC...but no luck because the seller lost backplate for 775 socket... cry.gif
*
That CoolIT is awesome bcoz u place the entire system inside your casing, compare to other watercooling that u have to placed the reservoir outside the case. Plus it has lcd display, a bonus when u have see-through casing side panel. thumbup.gif

For those who dont know the cooler, check this out:

http://techgage.com/article/coolit_domino_alc_cpu_cooler/1
http://techgage.com/article/coolit_domino_alc_cpu_cooler/2

But one thing for sure this item is rare in M'sia, i dont know where they sell it here, sad.gif
Bcoz i wish to slay LinX with this brows.gif
JoBigShow
post Aug 29 2009, 01:51 AM

Ω Ξ Ψ Σ Д Ӫ Ж Π ✿
*****
Senior Member
969 posts

Joined: Jul 2009
From: ɹndɯnן ɐןɐnʞ
QUOTE(kiLLua the babap @ Aug 29 2009, 12:40 AM)
ModernoPC got stock if im corect

btw how about Corsair H50 ?
*
For the same price as CoolIT, (in the US) the Corsairs outperformed the CoolIT
in terms of 'self contain watercooling system' from most reviews i've read.
The Corsair probably dedicate its function 100% on cooling, and the CoolIT 65% on cooling; 35% on the temp sensors and lcd displays and looks..
I like the CoolIT but i'll go for the Corsair bcoz results matters even though it look industrious and vulgar.
But then again, where to buy these products in KL? i havent seen any of them in Low Yat Plaza... sad.gif


Attached thumbnail(s)
Attached Image Attached Image Attached Image
JoBigShow
post Aug 29 2009, 03:51 AM

Ω Ξ Ψ Σ Д Ӫ Ж Π ✿
*****
Senior Member
969 posts

Joined: Jul 2009
From: ɹndɯnן ɐןɐnʞ
QUOTE(campsol2k @ Aug 29 2009, 02:31 AM)
Don't ever2 disable CPU TM function... unless you want to fry your procs...
*
I almost 4got what's its for. sweat.gif (cpu overheating protection function)
But it's only enable under the setting of:
1. Frequency>= 3.6Ghz FSB800
2. Frequency>= 2.8Ghz FSB533
But as a safety net, i should have enable it doh.gif

This post has been edited by JoBigShow: Aug 29 2009, 04:07 AM
JoBigShow
post Aug 29 2009, 07:31 PM

Ω Ξ Ψ Σ Д Ӫ Ж Π ✿
*****
Senior Member
969 posts

Joined: Jul 2009
From: ɹndɯnן ɐןɐnʞ
QUOTE(farscope @ Aug 29 2009, 01:26 PM)
i think 4.0Ghz will be too much heat for my RS to handle.. i'm happy to stay it at 3.6Ghz if my temperature does not exceed 70°C

btw, using 360x10 and 400x9 will have any difference in temperature?

i think higher FSB = higher voltage = higher temps rite? no? hmm.gif

Added on August 29, 2009, 2:13 pmhave a question here..

i set my Vcore to manual and set it to 1.26V but after i boot up, cpu-z onli shows 1.232V.. y huh? hmm.gif

Added on August 29, 2009, 2:54 pmlatest setting in bios i pump it to 1.28725V cz not stable in low voltage..

but cpu-z still show voltage reading between 1.216V~1.232V  rclxub.gif
*
I use to OC'ed my chip to 3.5ghz (334 x 10.5) when i was using my old 667mhz DDR2 to obtain that 1:1 ratio.
I set my vcore to 'AUTO' and running stable.

360x10 and 400x9?
If u'r concerned about temps u should opt for the 360fsb, but if u have DDR2800mhz and u want that 1:1 ratio, u opt for 400fsb.

i think higher FSB = higher voltage = higher temps rite? no?
That's why Intel created Core 2 Extreme procs, they have multipliers unlocked up to 14,
this way you can maintain a low~mid range FSB but still attain high Ghz, meaning normal temps but high performance/ghz.

i set my Vcore to manual and set it to 1.26V but after i boot up, cpu-z onli shows 1.232V.. y huh?
It's 'v droop'. (It's an Intel specification designed to protect your CPU from overshooting voltage), i think u should trust your BIOS nod.gif

latest setting in bios i pump it to 1.28725V cz not stable in low voltage..but cpu-z still show voltage reading between 1.216V~1.232V
Man that some big ass v droop! if u Google around, there's alot of threads about this issue. usually corcerned with enable/disable LLC. Well I'm not familiar with EP45-UD3L.... sweat.gif

JoBigShow
post Aug 29 2009, 08:07 PM

Ω Ξ Ψ Σ Д Ӫ Ж Π ✿
*****
Senior Member
969 posts

Joined: Jul 2009
From: ɹndɯnן ɐןɐnʞ
QUOTE(Raikkonen @ Aug 29 2009, 07:43 PM)
Mine E7300@ stock speed.

Can OC anytime but need to make sure got extra money just in case current system kaput  tongue.gif
*
What a rare occasion to see someone who actually insuring their processor when thinking about OC'ing,
usually alot of peeps blatantly take risks without thinking the consequences if the proc failed from OC'ing...maybe its the testosterone...maybe its just the lifestyle of OC'ers,...maybe its just the rush of not blowing the hell out u'r processors!...such mystery it is.
A smart man you are notworthy.gif
JoBigShow
post Aug 30 2009, 05:24 AM

Ω Ξ Ψ Σ Д Ӫ Ж Π ✿
*****
Senior Member
969 posts

Joined: Jul 2009
From: ɹndɯnן ɐןɐnʞ
OC Trial #1: 29/08/09 - 30/08/09
E7400 2.8ghz; 400 x 9.5 = 3.8Ghz @ 1.360v

Run time on prime95: 7 hours

Rig:
E7400 with Artic Cooling Freezer Xtreme
Asus P5QL-PRO
800mhz DDR2 Kingston HyperX
FSP EPSILON 80Plus 600w
» Click to show Spoiler - click again to hide... «


My hypothesis is that I might have overkill in terms of vcore resulting in high max.temp.
(but the max temp of 70c was just a temporary instant spike; on average the temps is about 60c)
Therefore for 'OC Trial #2' I will dramatically decrease the vcore in the range of: 1.32v< x <1.344v,
if sucess, i will continue decrease the vcore to point of failure in my following trials.

Something I find about fine tuning vcore is that, its easy to just randomly use high vcore and settled with it than to find the lowest possible vcore that the procs can be stable on.
(In the 21st century its all about low power consumption, reliable and energy savings @ reasonable performance, thumbup.gif like a Toyota Prius to a Lotus Hybrid or a Lexus GS 450h Hybrid)
I'll post my result for 'OC Trial #2' the next time i'm free.



JoBigShow
post Aug 30 2009, 08:16 AM

Ω Ξ Ψ Σ Д Ӫ Ж Π ✿
*****
Senior Member
969 posts

Joined: Jul 2009
From: ɹndɯnן ɐןɐnʞ
OC TRIAL #2 30/08/09

E7400 400 x 9.5 = 3.8Ghz @ 1.30v(BIOS) 1.296v(CPU-Z)

Run time on LinX 0.6.2: 28m 49s

Rig:
E7400 with Artic Cooling Freezer Xtreme
Asus P5QL-PRO
2G 800mhz DDR2 Kingston HyperX
FSP EPSILON 80Plus 600w

» Click to show Spoiler - click again to hide... «

» Click to show Spoiler - click again to hide... «

This time I felt pretty ambitious and drop significantly from my previous vcore(1.362v) to 1.30v. Suprisingly, no errors. thumbup.gif
The max temp is predictably high when running LinX, but what add insult to injuries is that I'm running thin out of thermal paste,
probably might improve by deducting at least up to 5c to my max temp if i get new paste applied. (anyone can recommend any good thermal paste??)
I didn't run 'OC TRIAL #1' with LinX because my max temp nearly reached 90c @ 13625v,
hence opted me to choose to run a much slower(boring) rclxub.gif stress test, prime95.

I will run 'OC TRIAL #3' with a more ambitious vcore setting; 1.28v < x < 1.292v @3.8ghz
I have the feeling I'm getting closer to fine-tuning it to the closest minimum stable vcore laugh.gif

I can hear "4ghz" is calling me to its domain. I want to join the exclusive 4ghz+ club BUT...lets take care of business one at a time.
I will post my next OC trial soon biggrin.gif

This post has been edited by JoBigShow: Aug 30 2009, 08:23 AM
JoBigShow
post Aug 30 2009, 11:49 PM

Ω Ξ Ψ Σ Д Ӫ Ж Π ✿
*****
Senior Member
969 posts

Joined: Jul 2009
From: ɹndɯnן ɐןɐnʞ
QUOTE(farscope @ Aug 30 2009, 11:56 AM)
nice OC there.. i might try follow yr settings n see how it goes.. laugh.gif

anyway, did u enable LLC?

Added on August 30, 2009, 12:11 pmTP.. u can try tuniq tx-2 pr tx-3.. thumbup.gif
*
I have the LLC (Load line calibration) Enable bcoz its improves v droop directly. If disabled, it will follow Intel specs, and knowing Intel® is paranoid when people want to OC, (due to warranty breach and lawsuit) they will usually have their settings with very big gaps in v droop to overprotect the chip. If u still have problems with this huge v droop dilemma, i guess u might want to start a new thread to isolate this matter, probably a lot of people might have the same problem as you. The worst case scenario about irregular v droops is that it might involve manipulating BIOS, etc. I'm not a pro with this kinda things sweat.gif

tq for the TP recommendation, I hope they got the stuff in L.Y.Plaza, i used to use Xigmatek's TP, they somehow become solid now, not a good TP. rclxub.gif
JoBigShow
post Aug 31 2009, 02:34 AM

Ω Ξ Ψ Σ Д Ӫ Ж Π ✿
*****
Senior Member
969 posts

Joined: Jul 2009
From: ɹndɯnן ɐןɐnʞ
QUOTE(OC4/3 @ Aug 31 2009, 01:05 AM)
Eeehhh just to clear up some misinformation.
Intel specify the vdroop spec is for the sake of protecting VRM circuit and processor smile.gif That the main reason for the specify spec wink.gif
*
Thanks 4 the correction btw, i was just being sarcastic about Intel, just a capitalist joke tongue.gif
JoBigShow
post Aug 31 2009, 07:35 PM

Ω Ξ Ψ Σ Д Ӫ Ж Π ✿
*****
Senior Member
969 posts

Joined: Jul 2009
From: ɹndɯnן ɐןɐnʞ
I remember I used to encounter something like this years ago, i did everything troublelshoot possible, i thought the mobo was done for. But then for no reason i decide to replace the mobo battery,(that cell battery to boot up BIOS)...the mobo was a live again! dunno this would help u....
JoBigShow
post Sep 1 2009, 07:27 AM

Ω Ξ Ψ Σ Д Ӫ Ж Π ✿
*****
Senior Member
969 posts

Joined: Jul 2009
From: ɹndɯnן ɐןɐnʞ
QUOTE(nidzamputra @ Aug 31 2009, 09:46 PM)
i'm just tried already..but no chance to bring 'him' back...already change cmos battery with new one but still refused to boot... cry.gif  cry.gif  cry.gif

maybe i need replacement mobo.. rclxub.gif
*
Damn, that was some of my best solution for fixing mobo but it failed to revive yours. doh.gif

One last thing, did u try using the PSU from u'r first rig, this is a tedious(leceh) process, my bro used to have mobo probs
(LAN light is ON, HDD ON, everything is ON but not BOOTing), he change psu and comes the famiiliar Boot beeping sounds and everything's alive again.

I dont want to give you false hope bro, I'm just saying if u'r up for it.........
The DFI is such a good board to be given up on nod.gif
« Next Oldest · Overclockers United · Next Newest »
 

5 Pages  1 2 3 > » Top
Add ReplyOptions
 

Change to:
| Lo-Fi Version
 0.0205sec    1.15    7 queries    GZIP Disabled
Time is now: 20th December 2025 - 09:14 AM
All Rights Reserved © 2002- 2025 Vijandren Ramadass (~unite against racism~)
Removal Request
Powered by Invision Power Board © 2025  IPS, Inc.