Jump to content
Corsair Community

2sets of TWIN2X4096-6400C5DHX (8GB) on Asus Striker Extreme


b1z

Recommended Posts

Beer is the only solution..

 

ROFLMAO

 

Also, keep in mind that with Intel chipsets, you very likely will have to drop a speed bin if you populate all four banks. The 800Mhz will very likely have to be run at 667Mhz. There is no method to unlink the memory from the CPU on Intel chipsets so that is another issue to take into account.

 

Systems are optimized at this time for two banks and full DRAM throughput. Four banks is problematic even on Intel.

 

Sorry to be such a spoilsport.

Link to comment
Share on other sites

  • Replies 63
  • Created
  • Last Reply

You're actually answering so fast..

I wrote the last post so you don't miss it if I had edited my previous post :D:

 

Since I'm no native english speaker I had no clue what the hell a "spoilsport" is...

Just until I've looked it up in a dictionary.

I've read the translation and just deeply thought - damn he is so damn right. :D:

 

Just began to think about selling the Striker Ext. and buying a good Intel based.. and then that....

 

One thing is for sure: If I'm buying another board, I'll do a lot of research before...

 

Thats enough for today.

Lets see if someone else answers within the next days.

You're too realistic and I'm a goddamn fanatic :p:

 

But you really convinced me of a MB switch as the next step.

Thanks..

Link to comment
Share on other sites

What the hell is this now...

Loaded setup defaults, set Vdimm and timings according the specs,

pulled out 2 modules to test each pair on its own again which worked fine earlier.

(Just to be sure the pairs are OK so i can go and switch the board and be sure the RAM is ok.... since memtest results and errors say nothing about the memory itself on a crappy board - I wanted to test them on the old and then on the new to be sure)

 

And what happens now?

Random freezes within memtest and within the BIOS itself, now even with just one pair installed.

Can someone just explain me what has happened to my board within the 5-6 days it worked all like a charm.. now nothing works anymore. Nothing.

 

This whole thread is more like a board related problem than memory related.

I'm sorry for that but I wasn't sure at first so I've asked here.

But now... I'm not even sure anymore if it's really just the board - or if both

components are defunt now.

 

:confused::confused::confused:

Link to comment
Share on other sites

I would turn the system off. Remove the power plug. Remove the CMOS battery. Remove the DRAM. Remove the Video Card. Let stand for a few days if you can. I doubt if you have damaged the components yet and this should allow the DRAM and all other CMOS substrates to fully release the discharged electronic resonance.
Link to comment
Share on other sites

I've pulled everything out yesterday, I'll let it sit 3 days completely unplugged :sigh!:

 

In the meantime.. I'm doing some more reading.

 

 

Ramguy? I've just found one of the first quotes by you with a tested 4 dimm setup on 680i. (Striker Extreme)

 

But we have tested this MB with 4 6400 Modules

at 800 MHz with an 1066 FSB CPU, with these settings:

SLi ready: Disabled

Memory Voltage: 2.0v

NB Core Voltage: 1.4v

TCL: 5

TRCD: 5

TRP: 5

TRAS: 15

Command Per Clock: 2

TRC: 28

 

http://www.asktheramguy.com/v3/showthread.php?p=311256#post311256

 

So, seems like now there is some lab experience with 4 banks populated and 680i! :biggrin:

 

I do have a similar setup, however to make it short I have a native 1333 Mhz CPU fsb (4*333) and the 2GB sticks. (PC6400C5)

 

I don't know how you came to tRC: 28 but I think the baseline for this is tRAS+tRP and you've loosened it a bit?

 

I'll try these next, any suggestions now with your lab experience?

 

SLi ready: Disabled

Memory Voltage: Auto (with new 1305 bios sets to 2,1V if 4 banks populated!)

NB Core Voltage: 1.4v

TCL: 5

TRCD: 5

TRP: 5

TRAS: 18

Command Per Clock: 2

TRC: 28

 

Just loosened the tRAS a bit, but whatever. Perhaps you have another idea. :o:

Link to comment
Share on other sites

What BIOS version are you using and what is the O.S. you are using and what are the exact settings you have set for Both memory and CPU?

 

BIOS is 1305 (2007/08/27)

OS is GNU/Linux 2.6.23.1 x86_64 (64bit)

 

 

The exact BIOS Settings are.. well, since it got unstable I was yet unable to find alternative ones which would have worked. You can read in post 21 about what I've tried so far, including setting the RAM to 667mhz.

Perhaps I've just missed the right combination.

 

However, as requested here are the settings I've got memtest86+ v1.70 run through for totaling around 45 hours without errors.

Note that the Striker Extreme overvolts a bit so the 1.85V can be seen as 1,9V on others..

 

Asus Striker Extreme + TWIN2X4096-6400C5DHX + Intel QX6850

Loaded Setup defaults, set only these manually.

The remaining Auto Settings could be of interest...

====================================================

Timing:

 

tCL: 5T

tRCD: 5T

tRP: 5T

tRAS: 18T

Command Per Clock (CMD): 2 clock

tRRD: Auto

tRC: Auto

tWR: Auto

tWTR: Auto

tREF: Auto

Async Latency: Auto

 

SLI Ready Memory: Disabled

 

Voltage:

 

VCore Voltage: Auto

Memory Voltage: 1.85 V

1.2V HT Voltage: Auto

NB Core Voltage: Auto

SB Core Voltage: Auto

CPU VTT Voltage: Auto

DDR2 Controller Ref Voltage: Auto

DDR2 Channel A Ref Voltage: Auto

DDR2 Channel B Ref Voltage: Auto

 

 

Other (perhaps you want to know?):

 

ECC stuff: Disabled (since no ECC memory is detected)

 

CPU Spread Spectrum: Disabled

PCIE Spread Spectrum: Disabled

MCP PCIE Spread Spectrum: Disabled

SATA Spread Spectrum: Disabled

LDT Spread Spectrum: Disabled

 

CPU Internal Thermal Control: Disabled

Limit CPUID MaxVal: Disabled

Enhanced C1 (C1E): Disabled

Execute Disable Bit: Enabled

Virtualization Technology: Enabled

Enhanced Intel Speedstep Tech: Disabled

LDT Frequency: 5x

 

PCIEX16_1 Frequency (Mhz): Auto

PCIEX16_2 Frequency (Mhz): Auto

PCIEX16_3 Frequency (Mhz): Auto

SPP<->MCP Ref Clock, Mhz: Auto

 

FSB - Memory Clock Mode: Auto

FSB - Memory Ratio: Auto

FSB (QDR), Mhz: 1333

MEM (DDR), Mhz: 800

 

USB Legacy Support: Disabled

 

AI Tuning: Auto

====================================================

 

My CPU is a QX6850 so I was using the native 1333Mhz (4*333) CPU FSB.

 

:):

Link to comment
Share on other sites

I would turn the system off. Remove the power plug. Remove the CMOS battery. Remove the DRAM. Remove the Video Card. Let stand for a few days if you can. I doubt if you have damaged the components yet and this should allow the DRAM and all other CMOS substrates to fully release the discharged electronic resonance.

DerekT? Any new ideas yet?

I'm going to try it again tomorrow with the mentioned settings..

The whole time unplugged is 84 hours - should be enough?!

 

I'm going to use these as the new starting point:

SLi ready: Disabled

Memory Voltage: 2.0v

NB Core Voltage: 1.4v

TCL: 5

TRCD: 5

TRP: 5

TRAS: 18

Command Per Clock: 2

DRAM FSB: 667Mhz

 

:):

Link to comment
Share on other sites

Find your stability first. Use memtest and be certain that your system, (with the above settings) is stable.

 

I can NOT emphasize this enough. Find your stability first. Be certain... Take the time. Run both Memtest for 3 -4 passes. Run Prime95 v2.55 for at least 8 hours.

 

Is your system fully stable?

 

Here's a screenshot of 3.8GHz with a P5B (Commando) with 2GB of DRAM. It will NOT make 3.6GHz with 8GB of DRAM.

 

http://i191.photobucket.com/albums/z225/DaveCim/stability.jpg

Link to comment
Share on other sites

Thats what I want to do, if the settings are unstable,

could it help to even set all sub timings to the slowest settings possible and work up from there?

If not, what about inrceasing the "big four"? Eg. using 6-6-6-20 at 2T?

 

Just to have some ideas in which direction to go if the above are still unstable :sigh!:

Link to comment
Share on other sites

Thats what I want to do, if the settings are unstable,

could it help to even set all sub timings to the slowest settings possible and work up from there?

If not, what about inrceasing the "big four"? Eg. using 6-6-6-20 at 2T?

 

You can try it but keep in mind that having the strobes closed for too long a period is as useless as having them open for too short a period. You must keep in mind that there is a definite possibility that you will NOT achieve stability with that DRAM setup. But we are both ahead of ourselves now. Just do the testing.

 

Edit: Rethinking your latest question. You just might get by with 5-6-6-20. I do doubt that you can allow the main strobe to stay open at a Column Access Strobe Latency (CL) of 6. You might with a CL of 5 though. Always run your system at a 2T Command Rate. If you have been testing with a Command Rate of 1, retest with a Command Rate of 2T.

Link to comment
Share on other sites

Damn I don't get it pass memtest anymore.

 

Tried everything metioned above at 800 Mhz and 667 Mhz.

 

The common error I have seen along all errors regardless the settings I've tried is this:

 

For Example the "random pattern inversion" test:

 

efefefef is written into the memory.

 

Then the inverse pattern is written over and

memtest reads and checks if the first pattern was succesfully overwritten by the inversed one.

 

This is where the error appears:

 

afefefef is read out.

 

This does not match the inverse pattern (afafafaf ?) so there is only the first bit / first two? written.

 

Seems like the auto subtimings are too tight anywhere because the write operation is corrupted / choked?

Something is closing too fast for the memory controller to address and write the data?

 

Any ideas? tWR?

Link to comment
Share on other sites

Yes tested them each one on their own and get no errors after 3 passes for each stick and each pair.

 

This was on everything set to auto, spd timings 5-5-5-18-2T and 1,85V.

 

 

The test # varies extremely depending the voltage and timing settings.

 

Using the suggested settings of 5-5-5-18-2T at 2.0V and everything else on auto gives the errors in test #7. (the best result this far since no errors before this)

 

Going to 5-6-6-20-2T at 2.0V gives errors as soon as test #2.

 

Using 5-5-5-18-2T at 2.0V plus setting the slowest settings for all other subtimings gives errors always in test #3.

Link to comment
Share on other sites

When I change the NB Voltage from 1,2V to 1,25V (even tested up to 1,4V) I get the errors very soon in test #2.

(rest is all auto via SPD 5-5-5-18-2T and 2,0V)

 

I still can't understand if this is a memory fault since the errors always appear to be similar (its always the first of the 8 bits which is right, the other 7 bit are wrong - in nearly all tests) but appear in totally random memory ranges like sometimes at 234M sometimes 1473M, 7291M... whatever.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.


×
×
  • Create New...