WizaRd
I got the 2x 512MB ValueSelect Corsair yesterday and i installed vista.
I had kamen 2x 512MB ValueSelect Corsair ram installed before and these were exactly the same.
I put them all into my MSI K8N Neo4-F. They are PC3200 400MHz but according to MSI they should run at 333Mhz..
anyway,
Vista installed, loaded, updated and all, I am experiencing sometimes heik random reboots when i click masalan on something that should stimulate a pop-up or open a new directory etc . I disabled the automatic restart so i get to see a BSOD but when it locks up it just freezes (instead of the restart)
Now i took off the 2 new rams i installed and the system seems a little faster... ba3ed ma 7allo ya3moul restart (ive tried clicking some stuff but it is still fine)
So how does the system seem faster with 1GB instead of 2GB ? could there be anything to do with the timing of the ram ? (they are all identical.. all 4 of them :S )
Share some thoughts and i will report back if something happens with my 1gb now
WizaRd
OK i checked the voltage for the rams on the internet, it said 2.5V
In the motherboard it was set to 2.65V since 2 years... (even tho the existing rams page says 2.5V)
does lowering it to 2.5V could prove to be a solution for the slight stuttering and random reboots ?
Edit: After 2 hours of no reboot and lockup i switched to using the 2 new rams instead of the 2 old and i have the same performance.. it does not seem an issue with the rams themselves , maybe with the voltage ?
battikh
have u also tried changing the slots?
WizaRd
this is what i am doing now
rolf
a suggestion: get "ultimate bood cd" it contains, among others, some RAM testing utilities. use that to test your RAM configuration.
Alternatively, if you dont want to download UBCD, you can check out microsoft, they provide for free a ram testing utility (also in the form of a bootabe CD/floppy). Forgot the name though
WizaRd
i did as they instructed me on MSI official boards..
i tested with memtest86 (the most famous coz it is the only know i know lol) and for about 1-2 hours in test #5 no errors came up.
i tested with all 4 dimms inserted and with the 2 brand ones.. no errors.
I was told to put ram frequency to 166Mhz and Voltage to 2.7V.
It is working so far.. dunno when it decides to shit on itself...
LebaneseChiphead
Okay I do not want to confuse anyone or get too technical on you guys, but here is some information about what you were doing when you changed the voltage and the frequency. I will keep this information in general and I will not go into details of the specifications so I won't bore you to death.
Here is the relation between the voltage and the performance:
The ultimate goal of all of this is to have a square wave so you have maxium period and a lot faster speed where there is no Fall time and rise time from the Sine wave.
The slew rate of the signal had its rules too:
The slew rate = Voltage / time. ( voltage divided by time)
When Slew Rate is high:
If you drive it really hard You are going to get a square wave with
a ring more like a curved up sine wave at the top of the curve up
so you have more settling time for the load .
When Slew Rate is Low:
The lower the slew rate the more sinusoide the signal gets Too wavey like a bounsing sine wave
so in contrast the higher the slew rate the more square wave you get.
Lousy rise time will make it less stable and you will have jitter.
Now to the "electric current" issue AMPS.
The Higher slew rate (dv/dt) the higher the current draw in the capicator Load
So current draw is propertional to the slew rate(rise and fall time) & the capacitance of the Load(CL)
Now Localize current draw , creates a localized power drop which causes ground bounce
Ground bounce changes the ground reference, which forces the threshold to be variable in time, if a signal is driven low during
ground bounce it may not be observed a s a low as it did not pass through threshold.
Frequecy and reflections of the wave
they are two component for the frequency waves:
1- Voltage based dv/dt
2- Frequency based
In High frequency Domain we use the reflection to drive the signal from sine wave to square wave by using the 3rd and 5th and 7th harmonics(multiple of frequency period time)
In low frequency the reflections are attenuated or Damped/smaller
Frequency = 1/time
speed of light = frequency * Lamda
if Lamda is larger than the section ofthe wire, no reflections.
If lamda is shorter than the section of wire, reflections is due to the impedence mismatch.
I hope that helped you understand what was happening inside the Hardware when you changed your settings. :)
Chips IMR
WizaRd
and i thought all physics in high school were non-sense :D
