F2-6400PHU2-2GBHZ on ASUS P5B Deluxe

[theviper]

Hey there,

i've got three questions about running 2x GBHZ 2GB Kits (4GB) on an ASUS P5B Deluxe / Intel C2D E8500 Machine.


First: Differences?

The two Kits have significantly different age (one has a 75xx and the other one a 83xx Serial Number & the older ones have black / the newer ones a green PCB).

My question is right now:

Is there a significant difference except the PCB Color?
E.G. different memory chips?

Second: Dual Channel Setup

The Board has four memory Slots named A1 / A2 and B1/B2
What would be the best Setup regarding Dual Channel:

Putting the New Ones in Channel B1 and B2 (The old ones in a1 a2) which is my setup right now.
Or installing them this way:

Old A1
Old B1

New A2
New B2


Third: Timings, Clock & Temperature

I'm actually trying to overclock that Machine.

First tried to run them as follows:

CL4 (4-4-4-12-4-42-3-9-5-9) timings
DRAM Frequency: 437MHZ
Memory Voltage: 2.1V
FSB: 350MHz
Multiplier: 9.5
CPU Voltage: 1.30
_____________

Result is 2x3.3 GHz (instead of 2x3.17) which is basically just a first step. (aim: 2x3.5)

Received Prime errors. When i changed timings to CL5 everything worked fine.

When changing timings to CL5:

Do those values (4-42-3-9-5-9 as seen above) except the first ones (from 4-4-4-12 to. 5-5-5-12) need to be changed as well?


What would be a point on the temperature scale that should not be passed?

Right now they have some 44?C under Full Load after a 6hrs run.

Thanks in advance. Have a nice day =)

[genetix]

Values seems a bit high meaning latencies, but that doesn't explain the errors. Check our Performance level on there. add it +1t and check the errors again. I own these blocks on 2GB and 1GB sticks here. I know one of the sticks broke down on me, but well they lasted like forever.

Anyway, when you test memory and Board/CPU stability the Prime95 doesn't do any good for you. Prime95 calculates. and first failures comes from CPU vCore &/or NB voltages. Your memory you are testing first test them starting with 'MemTest86+ in MS-DOS (USB stick's quite nice with HP format tool).

When you get through tests 4,6 & 7 you should then test in Prime95+ as then you know (99% certainty) that the memory is stable and you can advance on your board settings.

Hope that helps.

-edit-

to answer your question about memory difference no there is no difference. however, I'd still suggest that you use 1-0-1-0 on different memory modules just in case. Where 1(A1-B1) = first kit 0(A2-B2) = second kit. What goes to Latencies I think you can even drop few, but more on that later.

[GSKILL TECH]

There is no significant/altering difference between memory as long as the specifications are the same.

Order does not make much of a difference, but you can place the new ones in A1 and B1 and the older ones in A2 and B2.

For timing values, set them to AUTO and let the computer generate those for you.

For temp, just compare an hour to 6 hours to give you a good idea on whether or not it is overheating it.

Keep us posted on your progress.

GSKILL SUPPORT

[Tradesman]

Theviper,

I'd run memtest on each set in slots 1-3 (A1-B1). You definitely want the sets separated as suggested above one set (I'd suggest the older set ) in A1-B1, the newer in A2-B2. The reason for doing this is for you system to operate in dual channel mode at peak efficiency you need identical or as near the same as possible in each channel Channel 1 is A and Channel 2 is B, if you only run two sticks say in B1-B2 you will be running in single channel mode.
Disagree with GSkill to just set the timings to AUTO, with the P5B it may well set them to an under performing value, if you set it to auto and the timings are higher (looser) that the advertised specs, I'd set them manually to spec. If that is stable, you could then try tightening them a little to increase performance.


TO GSKILL ,

The viper asked a pertinent question which deserves an answer, do the two sets have the same chips and specs? Gskill has been known to change chips on different RAM series's in the past, if these use different chips i.e. 1 set using chips that run hotter than the other can affect decisions in where you load them.
Think the key here is if the customer comes to you asking for info (which is the purpose of these forums) you should provide the info, not just ignore them.

[GSKILL TECH]

Running timing on AUTO simply helps people attain a benchmark point. With that value, they can tweak the numbers slightly and have their system running stable.

Regarding the chips, I don't have them here with me, I did not make them either, so I wouldn't be able to tell if there is a major difference between the two. All I know is between those two serial numbers, there was no major change in chipsets to my knowledge. I did not ignore the question, I gave him the best answer I could provide with the information he provided me.

GSKILL SUPPORT