Intel C2Q Q9550 E0 @4.12Ghz, 485x8.5 :: Intel DX38BT (BIOS 2006) :: Swiftech Apogee GTZ CPU H2O Block :: 4 GB (2x2GB) Corsair TW3X4G1800C8DF @DDR3-1617 6-7-7-22-1T
Swiftech MCR220 Dual 120mm Fan Radiator :: Ghetto Rigged Swiftech MCW30 NB Water Block
DIAMOND Viper Radeon HD 2900XT 1GB GDDR4 OC'd 848 GPU/1140 Mem
2x150 GB WD Raptor X 10k in RAID 0 (OS) :: 3x320 GB WD Caviar SE16 SATA in RAID 0 :: 1x1 TB WD Caviar GP (SATA) in a ThermalTake MAX4 HDD Rack
Windows 7 Professional (64 Bit) :: Antec TruePower Quattro 850W :: 1 PATA DVD-ROM :: 1 PATA LG SuperMulti 20x DL & Lightscribe :: ThermalTake Shark case
I didn't do a lot of checking, but I've been working through a setup. When I had the multi set for 8 and EIST enabled. In Windows CPUz showed 6 at idle, then 8 under load. When I disabled EIST, in windows CPUz showed 6 at idle and 9 underload which is the max for my E8400.
Ok, if you know the max speed your CPU and FSB can run next you'll want to find the max memory speed. I don't have a system handy to look at the BIOS page right now so I'm going from memory on the names. This is where it starts getting more complicated but I'll give it a try writting up some steps.
I'd start by dropping your CPU multiplier down to 6 (maybe not necessary, but it helps to eliminate variables) and your host clock frequency down to 333.
I think you mentioned there was a Profile 1 listed for your memory that ran at 1333. I'd start by selecting that, then selecting manual so your starting out with the profile settings.
Reference Frequency is also sometimes called the bus select strap, BSEL, or just straps. Normally the CPU signals this to the MCH, your Q6600 signals 266 to the MCH so it knows what clock is coming in and what ratio to set to get the right memory speed. The Reference Frequency settings lets you override the BSEL into the MCH to force it to use a different ratio. It also sets up other internal things which can have significant effects on stability. It's best to select a Reference Frequency closest to the actual FSB frequency you will be running. So if your running at 400MHz Host Clock setting, I'd recommend selecting 400 for the Reference Frequency.
DDR Frequency (or is it called Memory Frequency) is the other part of the equation that makes the actual ratio clock ratio. GTJ has a nifty spread sheet that does the memory speed calculations. But if you select BSEL=400 and DDR=1600 it will run the memory at the host clock speed. If the host clock is 333, final memory speed will be 1333. If this works, then you can just start increasing the host clock again until you find where it's no longer stable.
If it doesn't run at Host Clock = 333, BSEL=400, DDR=1600 then you have some choices. You can try increasing the timings (bigger numbers) this will give a little lower performance but can increase stability. Or you can try setting DDR=1333. So at Host Clock = 333, BSEL=400, DDR=1333, the actual memory frequency will be less than 1333 (I don't know off hand what it will be). Now try booting with this setting, if it works, then start raising the Host Clock again until you find the point it's no longer stable.
Once you find the stability point you can again try raising the voltage to go faster. You could try increasing the timings to go faster, but often that results in lower overall performance. Setting the timings is really a lot of trial and error. I don't know a good method to step through them, other than just playing around with them. Maybe someone else does other than just playing
The first timing CAS Latency or perhaps its called tCL, I don't remember exactly, but it has a big effect on performance. A lot of other DDR timings get automatically configured are based off the CAS latency so it effects a lot of things.
Command Rate also has a big affect on performance. When set to Auto it will run 1T with 1 stick per channel and 2T with 2 sticks per channel. 1T is better performance, but it could be less stable.
Once you have settled on what seems to be the maximum memory speed, then you can adjust your CPU multiplier back up to try to get as close to the maximum CPU speed you found earlier.
Getting the best possible memory performance setting is tricky, it could be a trade off between Frequency, timings, and command rate. Everest by Lavalys is a good way to check DDR performance. Though even when measuring performance you can trade off between bandwidth and latency.
WHAT EXACTLY DO YOU WANT ME TO DO HERE???
SORRY I DONT UNDERSTAND WHAT YOU MEAN
SELECT THE PROFILE 1 AND USE THOSE SETTINGS WITH MANUAL OR
USE PROFILE 1
well anubi im using a northbridge small i believe 20mm cooler
that hooks into the small black clip that came with the mobo
im using the CNPS9700 LED zalman CPU COOLER
ARTIC SILVER 5.1 PASTE
AND A http://www.newegg.com/Product/ShowIm...e+Panel+Window
Tagan Aplus CS-EL Diablo Black 0.8mm SECC ATX Full Tower Computer Case - Retail
Extreme sports athletes have nothing on the Tagan’s CS-EL Diablo case when it comes to big air. The bigger the fan the more air it moves at a given speed. The slower a fan spins the quieter it is. When you put those facts together it leads you to one conclusion—big fans are better for keeping your computer cool. Tagan took this to heart and put a 250mm fan in the front of the case and a 360mm fan in the side panel. If you really think you need more air movement you can put a 120mm fan on the back panel.
THIS CASE IS AMAZING
IT BLOWS AIR LIKE NO OTHER I DID A LOT OF RESEARCH TO FIND THIS
AND ITS HUGE!!!!!!!
Hey Where Is The Mch You Are Referring To On The Motherboard
Never Mind Man I Got You I Dident Know I Could Do That
Thanks Bro
Well As 333 Fsb And Using Profile 1 then selecting manual
With Reference Frequency At 333
Result = Fail
Wont Post
im going to bed now have to be at work in 5 hours
ill do more testing tomorrow thanks for all your help
everyone who has helped me
see you tomorrow
Bookmarks