|
![]() |
Overclock.net - Overclocking.net > Intel > Intel - General | |
Even the slightest OC gives me the BSD...
|
||
![]() |
|
|
LinkBack | Thread Tools |
|
|
#31 (permalink) | |||||||||||||
|
New to Overclock.net
|
Thanks to your guidance and tips, I have successfully reached 3658 Mhz stable (seems stable, at least).
It appears to me that being below the 400 Mhz FSB line has been the main anti-BSD here. Why, I don't know... (do you?) As some of you told me, the multiplier should be lower than 10x, but for that I must cross the 400 Mhz FSB barrier I am afraid to cross. Being things like they are now, what should be the next steps (if any)? Thank you all for your help so far! You have certainly changed my non-OC-but-very-BSD-rig into a decent-OC-non-BSD-little-monster!!!! ![]() ![]() ![]()
|
|||||||||||||
|
|
|
|
|
#32 (permalink) | ||||||||||||
|
News Fiend
![]()
Join Date: Mar 2009
Location: The Loan Star State
Posts: 1,651
Rep: 143
![]() ![]() Unique Rep: 96
Trader Rating: 0
|
If 400 fsb is giving you trouble, give it a bit more north bridge voltage. That's how the cpu and RAM communicate--the faster the speed and higher the fsb the more volts it needs. Try giving your NB 1.24 for now. You should not have to exceed 1.26v on the NB for 400fsb. If you can't get stable there (at 400 fsb) you are hitting a wall somewhere; either your chip or mobo has hit the fsb wall. Generally, you should be able to get to that 400.
__________________Now here's where volts come in to play. More volts = more heat. The maximum temp you want to see on the cpu when you are stress testing is 65c--that's my personal opinion. The thermal limit for that chip is 71.4c. You don't want to reach that (71.4c) or go above it, it may damage the chip and the heat will cause instability. Heat will kill that chip faster than volts will. Raise the vcore until you see 1.32v in CPU-Z. Until you learn more, that should be your limit. Vcore 1.32, NB 1.24/1.26 and go higher. You will eventually have to lower that cpu multi to at least 9 to hit 400 fsb. Do you know how to test for stability? And, by the way...Good work. It takes time but you'll get it.
|
||||||||||||
|
|
|
|
|
#33 (permalink) | |||||||||||||
|
4.0ghz
![]() |
Quote:
Some of those cut down chips are FSB limited (the E7000 series is a cut down wolfdale). I have seen others commenting on the 400 FSB limit of the E7000 series. I have an E5200 in an HTPC build which runs 3.2GHz OC, but as soon as I go up one notch from 266 to 267 FSB, the PC fails to boot. I cant say for certain this is your issue though. You can verify this by dropping your CPU multi to x6, and seeing if you can run 400 FSB. You might need more northbridge voltage with your board though. What voltage is it at now? Your next step is to run Prime 95 stress testing to see if it is stable. You can find it in the downloads section. Also make sure to get realtemp to monitor your temps.
__________________
eVGA 680i Overclocking guide .......... ASUS P5E/Maximus Thread.......... 4.4GHz CPU-Z
Last edited by ericeod : 2 Weeks Ago at 11:31 PM |
|||||||||||||
|
|
|
|
#34 (permalink) | ||||||||||||
|
News Fiend
![]()
Join Date: Mar 2009
Location: The Loan Star State
Posts: 1,651
Rep: 143
![]() ![]() Unique Rep: 96
Trader Rating: 0
|
Absolutely true. Mine won't even do 405 stable no matter what the voltage or how far I lower my multi.
__________________And even though running 1:1 is not always the best configuration, it would give him more options for the RAM oc when he's ready for it--if the bios dividers are the same as the Ultra Durables. He put in some good work today though.
|
||||||||||||
|
|
|
|
|
#35 (permalink) | |||||||||||||
|
4.0ghz
![]() |
Quote:
__________________
eVGA 680i Overclocking guide .......... ASUS P5E/Maximus Thread.......... 4.4GHz CPU-Z
|
|||||||||||||
|
|
|
|
#36 (permalink) | ||||||||||||||
|
New to Overclock.net
|
Quote:
|
||||||||||||||
|
|
|
|
|
#37 (permalink) | ||||||||||||||
|
New to Overclock.net
|
Quote:
By the way, this morning I switched on the PC but I didn't POST. I had to keep down the power button until the PSU shut down and then I ****ched on again, POSTing and logging to windows as if nothing ever happened. What do you think of that? Thanks!
|
||||||||||||||
|
|
|
|
|
#38 (permalink) | ||||||||||||
|
News Fiend
![]()
Join Date: Mar 2009
Location: The Loan Star State
Posts: 1,651
Rep: 143
![]() ![]() Unique Rep: 96
Trader Rating: 0
|
Let me just give you a heads up before someone else does: try not to double post. If something new comes to mind, a question, or suggestion and no new posts have been made in your thread, hit the "Edit" button of your last post and and the new info to that post like this
__________________Edit: Then add the new stuff here. Keeps the threads cleaner. __________________________________________________ _______________ As for the miss post in the morning, the oc might not have been stable. Just because you boot into Windows does not mean it's stable. This is the boring work part now; stability testing. #1 Open CPU-Z and make sure the bios settings you used yesterday are still in place. When a computer goes through a boot loop because of instability it will usually revert itself back to the previous "Default" bios settings before powering up to prevent damage to itself--gotta love technology. #2 Google and download these programs: Real Temp, Core Temp, Orthos and memtest86+ #3 Your RAM has heat spreaders right? Maybe give them an extra .02 volts (to 1.82v) in bios before starting. #4 Open Real Temp, then open Orthos. Set Orthos to run the test called Small FFTs. This is a Linpack test that will check the cpu for stability. Your goal is to pass this test for at the least 10 hours. *Disable your screen saver before testing. Do not use the computer while testing. Keep an eye on those temps while testing.* You'll usually see your max temps after about 40 minutes. You DO NOT want to get near 70c on any of your cores. In my personal opinion, 65c should be your limit. 60c would be ideal for a high with your cooling solution. #5 If you fail the Small FFT test that usually means you need more vcore on the cpu. But, with your cooler (you kinda need a better one for a monster oc) I am not going to recommend that. Instead I personally would recommend you back the clocks down a bit, leave the voltages alone and test again. If you want to let the cpu rest a bit between tests, 20 min or so, that would be cool. Test, adjust, until you get Small FFT stable. #6 Once you get Small FFT Stable, use Orthos and set the test to "Blend" this time. The goal for this test is 8 hours. This test checks the North Bridge mostly and how data moves between the cpu and RAM. Failing this test means you need a bit more volts on the North Bridge. But, I personally feel that 1.24/1.26 should be more than enough for the fsb you are running. You shouldn't have to exceed that. Once you get Blend stable, we'll see what we can do about your RAM. Just remember, that's your hardware...you paid money for it. If you are not an enthusiast willing to shell out big bills for new stuff, treat it accordingly; adjust your overclock to reasonable levels to maintain it's longevity. If you want to get those clocks higher, you'll need a better cooler and you will need some good thermal grease...in overclocking, every degree counts and even the investments should be made wisely. Good luck bro. You're doing good. Read through some of the guides in the Intel Motherboard section for boards by the same maker and you'll get a grasp of this in no time. Edit: Terminology: Post = Computer successfully reaching the bios screen and passing it. Boot = Reaching Windows.
|
||||||||||||
|
|
|
|
|
#39 (permalink) | |||||||||||||
|
New to Overclock.net
|
I don't understand (technically) why is it better to OC the FSB more than the multiplier. My actual setup is working seemingly perfect @366FSBx10Multi=3660MHz (besides that wierd no-POST issue at the mornings). Dumb question for the seasoned OC'er: is my actual config too amateur to leave it like it is?
__________________
|
|||||||||||||
|
|
|
|
|
#40 (permalink) | ||||||||||||
|
News Fiend
![]()
Join Date: Mar 2009
Location: The Loan Star State
Posts: 1,651
Rep: 143
![]() ![]() Unique Rep: 96
Trader Rating: 0
|
That's a good question. I'm no expert...learned the little I know here--from the smarter folks...but here's how I see it in my head (attachment):
The "wider" the fsb, the wider the bandwidth, the more/faster information goes between the cpu and RAM. I could be wrong but that's what I gathered. My chip walls at 400 fsb so I work with that for my oc's. Edit: also, if you decide to try and squeeze more juice out of that RAM, you'll notice that as your fsb goes up, so do the options for oc'ing (using the dividers). Gives you more options for higer clocks.
|
||||||||||||
|
|
|
![]() |
| Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
| Thread Tools | |
|
|