flawlssdistortn
Senior member
- Sep 21, 2004
- 680
- 0
- 0
lol, clocking the RAM down to 333mhz 1T is about the same as 400mhz at 2T. And yes, both are significantly large performance hits. Thx GTTeamKiller i'll try my luck with more voltage.
Originally posted by: Brickster
Some info I found:
Whether or not it is possible to run the Command Rate at 1T depends on a variety of factors, though.
- Memory bus frequency
- Number of chips on the DIMM (the more chips, the longer it will take the controller to sort out which specific chip is the one that needs to be selected)
- Quality of the DIMMs used
- Number of DIMMs running in the board (if there are more DIMMs, the total umber of chips will naturally increase)
- Position of any given DIMM (move the DIMM to the farthest DIMM slot and it will fail)
Source: LINK
originally posted by jaiinquisa in www.houseofhelp.com/forums
Recently built a computer system comprising of an A64 3500 and a GA-K8NSNXP-939 (initially the F3 BIOS installed) with Corsair 3200 Twin XMS in dual channel mode (DIMM slots 1 and 2).
My initial Windows XP Home installation undertook without error. In addition installtion of Doom3 and UT2k4 occured without error. Doom3 runs perfect, however, UT2k4 runs perfect at first but then the Blue Screen appears providing me with the myriad code 0x0000008E.
I updated all relevant drivers and updated the BIOS to F4k but still the Blue Screen appeared that became more and more consistent in duraton of time.
I concluded that it is a RAM error. (It is a RAM error code). I believed it was faulty RAM; thus I returned the RAM to Overclockers.co.uk to expidite an exchange.
Installed the new memory in DIMM slots 1 and 2 and ran UT2k4. Argh! Blue Screen again with the same code 0x0000008E therefore I searched the Corsair Support Forum and found this thread. I read some of the thread and became, well, depressed.
UT2k4 runs perfectly in single channel mode. I can even run the RAM in DIMM slots 1 and 3 or 2 and 4.
It seamed to me that the Blue Screen only appeared in UT2k4 - the game I play the most. Disheartened I considered obtaining a MSI Neo2 Platinum via exchange.
But...
I installed the F4 BIOS released recently. Yay! The HDD light now functions in conjunction with SATA HDD's (Pfft - Big Deal!). I became brave and decided to remove a RAM module from DIMM slot 3 and insert the RAM into DIMM slot 2 - now in dual channel mode.
Heart beating fast and praying (for the first time) I ran UT2k4. Played it. Played it again and again. Without a single Blue Screen initially. However, now a Blue Screen may appear intermittently.
There have been many stories behind the plethora of myriad Blue Screens. One thing to note is that the Blue Screens remain Processor - RAM - Motherboard related but seem to provide different error messages.
Why? You may already know. It may have been posted already. Heres the explanation anyway. I believe it is better to obtain an explanation than remain in the shroud.
I shall state as simply as possible.
RAM is divided via memory addresses and each memory address may or may not have a variable assigned to it. The variable stores data relevant to a particular program and another variale stores data relevant to another program. Variables are created and initialised or assigned to in due course when a program is executed/within execution.
RAM functions like a stack of plates; the first plate placed in the pile is the last plate to be used. In the case of RAM the first data stored (temporarily) is the last data to be obtained. However, if the processor requires the first data stored (temporarily) but other data has been written to variables assigned to subsequent memory addresses ('placed on top') the stack must be 'popped' - (See section regarding instruction and stack pointer).
Heard of D-Step Code?
The code is in the BIOS and is updated in the F4 BIOS. The D-Step code is a component in the control of the dual channel memory controller (Ironic, a controller being controlled?)
The D-Step Code functions in conjunction with the instruction and stack pointer.
The instruction and stack pointer.
The instruction and stack pointer are in the CPU or Microprocessor register and remain reliant on the quality of the dual channel memory controller D-Step coding.
Pay close attention.
Example.
In memory address 100, 101 and 102 three variables are assigned and initialised with data utilised by my beloved UT2k4. Within the execution of UT2k4, UT2k4 requires the data in the variable assigned to memory address 101 but 102 also stores UT2k4 data within it's assigned variable. Thus, 102 must be popped in order to obtain the required UT2k4 data within the variable assigned to memory address 101.
Therefore the CPU utilises the relevant 'pointers' to pop the data and maintain track of the data required.
However, what is likely to happen is the stack pointer 'misses' (much like an untuned car) the variable. Due to the fact D-Step Code and 400Mhz RAM module(s) work as well together as me jumping up and down with nitro-glycerin.
If the computer system could speak, it would be saying something like, 'Where the heck has that memory address gone to with that variable thingy assigned to it and thus prefers the easy option - THE BLUE SCREEN.
RAM is dynamic and volatile and the reason for the mulitude of Blue Screens.
For example memory addresses 100, 101 and 102 may be utilised for UT2k4 within my computer system but Doom3 may utilise 100, 101 and 102 within Mr.A's computer system so Doom3 may fail.
Of course, the stack pointer 'missing' may happen at any time, at any location in memory.
Memory tests test the memory addresses individually - highly unlikely to provide you with a Blue Screen - the memory is not at peak and the stack pointer controlled by the D-Step Code is not 'firing' all about the place.
Prime95 tests the memory addresses in all circumstances - peak load, high load, med low, low load - different combinations of memory addresses. IRQ_NOT_LESS_THAN_EQUAL_TO error message will occur. (Prime95 = Rape of RAM)
The GA-K8NSNXP provided by Gigabyte, providing a motherboard that 'supports' 400Mhz RAM modules but really does not because it utilises D-Step Code is possible to redeem. (E-Step Coding is utilised for 400Mhz RAM modules).
Change the RAM timings - this is not my department - not an overclocker. Essentially time the RAM to function with the D-Step Code. (Manual tuning).
or...
Wait until Gigabyte updates the Step Coding within the BIOS. Hopefully another few updates shall be produced in order to perfect the Step Coding.
Conclusion.
I hope the above outlines and specifies the 'problem'. Obviously, the above is an in a nutshell view and thus specifics are alot more complicated. The annoying part, the problem(s) would have been avoided if AMD had said to Gigabyte, 'The A64 Socket 939's functions with 333MHz RAM modules with D-Step Coding. E-Step Coding is ustilised with 400Mhz RAM modules.
I feel two companies need a slap with a large trout.
Best of luck. Jai (CCIE) - England.
Originally posted by: Pohemi420
This is getting rediculous...in fact...it's recockulous. I was having the same memory issues as most of you with Corsair TwinX1024-3200XL modules. Assuming the on-die memory controller was causing at least SOME of the problem with the low latency Samsung chips, I decided to try the Kingston 1gb KHX4000 modules. These modules have latency timings of 3-4-4-8, and are tested at 2.65 volts. I figured the more relaxed latency timings would agree with the AMD memory controller better....
I was wrong. I've had WAY more problems with the Kingston memory than I ever had with the Corsair. The Mobo keeps resetting itself to run the memory timings at 3-4-3-8, which I'm not even sure the memory is capable of. I've upped the DIMM voltage to 2.7, and even 2.8, and still CAN NOT get this system stable in dual-channel at DDR200. I'm fed up. Gigabyte(after weeks of hearing nothing) got back to me and says,"Oh, well...we tested this board here in our labs with the same memory and CPU you are using, and found no problems whatsoever". They continue to try to blame it on my software(WHAT!!!) Needless to say, I can't RMA my Gigashit board now. You know what? I DON'T CARE. I'm about to throw it in the dumpster. I've decided to order an Asus board, and will update you guys when I get it. Good luck to the rest of you paying insane amounts of money to have a system that runs sub-par when it should be SMOKIN!
Originally posted by: feilenkopf
I would like to find out more data but i dont know how to get into the advanced Bios setting .I read about the F1 + ctrl Key but i tryed that during boot up and in the normal dos menu, it will not come up ,i am sure i do something wrong but what? Please help.Also is The program sandra free or do i have to pay for it,and where can i get it.And why want the 2 buttons in Easy tune 4 not work , where you can change the voltage and all the other stuff,Help.
I know, a lot of qustions, but if i dont ask, i will not learn.
Originally posted by: swNYC
Here's the new F5 BIOS for the GA-K8NSNXP-939 (originally posted by davey at PCPerspective)
Download Link
Originally posted by: flawlssdistortn
Guys would it be dangerous to bump my memory to 2.7V even though it's rated at 2.5?
Originally posted by: Brickster
For all you folks having issues with 4-DIMMS... here is some good information from Today's review on teh new Epox board on Anantech...
"Tests with all four DIMM slots populated on the Epox board were something of a disappointment, however. Most 939 boards that we have tested require 2T Command Rate with 4 DIMMs in two dual channels. Not surprisingly, 2T was also required with 4 dimms on the 9NDA3+. However, this Epox is the only 939 board we have tested that absolutely forces DDR333 when four DIMMs are installed. Some other boards that we have tested also try to force DDR333 with 4 DIMMs, but this can normally be corrected to DDR400 by selecting that speed in BIOS. "
I believe our board allow us to run at 400 on 2T, is this correct?
I cannot confirm, I only use two DIMMS.
Thanks,
Originally posted by: flawlssdistortn
I can't even get stability with 2 DIMMs at DDR400 unless command rate is 2. And you know what? That performance is about the same as 1T at DDR333.