Jump to content
Corsair Community

TWIN3X2048-1333C9DHX + Striker II Extreme


cje

Recommended Posts

Hey guys, sorry to bother you. I have searched, but I can't find a thread that is similar to my question.

 

On the side of my modules it states 9-9-9-24 1.5v rev 3.1.

 

I only noticed yesterday, that my motherboard, when setup defaults had been loaded with everything on auto at stock settings, was giving the RAM modules a reported 1.66/168 VDIMM!

 

So, I changed the option of voltages in BIOS down to 1.5v, and now due to the striker overvolting, it is giving it 1.58v. But my system is unstable. It passed some Intel Burn Tests and OCCT, but restarted when I told the pc to shut down. I am going to run memtest today and see what it will yield.

 

My question is, what voltage should I run the modules at? I have 4x 1gb sticks installed. If I set 1.6v in BIOS, BIOS will report that they are getting 1.66/1.68v, which I have been told is too much if they're rated for 1.5v?

 

Many thanks in advance.

Link to comment
Share on other sites

  • Replies 67
  • Created
  • Last Reply

If you populate all four DRAM slots, then you are advised to drop the DRAM bandwidth and raise the memory controller voltage.

 

So, you are advised to drop the bandwidth to 1066Mhz.

 

The only memory Corsair supports in a four DRAM slot population is their Quad kits. This does not mean two dual kits or even worse two dual kits of different versions will not work together, just that you have no recourse to RMA Corsair if they do not work together. If there are problems with the DRAM modules, they are still supported via lifetime warranty, but that is for physical defects and errors in a two stick kit mode. Matched Quad sets are highly binned modules for a four slot population. The Dual sets are tested as a kit, not a Quad and warranted as a dual kit, not a Quad setup. You need to Memtest them as Kits (2 modules, not as a four DRAM slot population). If they fail as a kit or singly, then you have RMA. Keep in mind that there is no support for mixing and matching different versions of DRAM either since different version are of different fabrications and/or manufacturers. They might work, they might not work, but you have no support if they do not work. Thus you are advised to buy all the DRAM at the same time and in this way be more assured of getting the same versions. If you buy online, be sure to specify same versions of the modules and do not expect even then that they will work and again, no support unless you purchase a quad kit.

 

For a four DRAM slot population of 800Mhz DRAM you are advised to drop the bandwidth to 1066Mhz and raise the NB Voltage +.2v. You can then attempt to recoup some lost bandwidth via the CPU <--> MCH <--> DRAM overclock. Once stability is found, you can then, if you wish, attempt to regain some of the lost bandwidth of the CPU <--> MCH <--> DRAM stream with an overclock. First stability is necessary as some motherboards, even though they state a four slot population, will not work with stability on a four slot population and need to be RMA'd. So, first stabilty, then overclock.

 

Please research this thread:

 

http://www.houseofhelp.com/v3/showthread.php?t=64360&highlight=Unpredictable+results+QUAD2X4096+Gigabyte+X38-DS5

 

There is a limitation of the On Motherboard (Intel) memory controller or On CPU (AMD) and a population of all four banks at the full speed of a two bank run. This limitation is across the board with both AMD and Intel chipset based boards, DDR, DDR2 and DDR3.

 

Keep in mind that often issues can arise in the longer term even though you have shorter term stability. The issues seldom damage the DRAM unless you raise the DRAM voltage too high. They damage the on CPU (AMD) memory controller which has a harder time keeping up with the demands of a four DRAM slot population being run as fast as a two DRAM slot population. This is why you are advised to perform the DRAM lowering and MCH Voltage raising.

 

If you need help with specific settings I can set a template for you.

 

Regarding DRAM Voltage. Those modules are warranted up to and including 2.0v and so 1.7v is a walk in the park for them.

Link to comment
Share on other sites

If you could post me a template that would be great. I just cannot get this board stable at 1600 fsb. I am unsure of advanced memory timings and such like. I am not home all day today, and my pc is running everything on auto, except VDIMM, which I've manually set to 1.5v. Ill see what it says this evening.
Link to comment
Share on other sites

You did not read what I posted or the link or you would not have asked that question. There is no use of my helping you when you demand all four modules to run at the speed of a two module kit.

 

Please CAREFULLY read both my post and the FULL link before posting in here again.

 

If you could post me a template that would be great. I just cannot get this board stable at 1600 fsb. I am unsure of advanced memory timings and such like. I am not home all day today, and my pc is running everything on auto, except VDIMM, which I've manually set to 1.5v. Ill see what it says this evening.
Link to comment
Share on other sites

You did not read what I posted or the link or you would not have asked that question. There is no use of my helping you when you demand all four modules to run at the speed of a two module kit.

 

Please CAREFULLY read both my post and the FULL link before posting in here again.

 

At the moment, I am running only 2 sticks. I apologise I did not mention this earlier, I somehow omitted it.

 

What I am saying is: I get instability with 2 sticks, when I set RAM to 1.5v as it says on the sticker,. BUT I ran memtest for hours today and it had 47 passes. I was giving them 1.6v (actual reported 1.66). That is how much the motherboard is giving it on auto by default.

 

All I want to know is: Am I OK running these sticks at 1.6v? You said in a previous post that they are OK up to 1.7v. So I guess that is a yes?

 

Thank you for your replies though. I did not have issues running 4 sticks, but I have taken them out for the time being as I was previously aware that they cause extra pressure on the memory controller.

 

In short: I think my cpu is a bad clocker at the moment, took out 2 sticks of ram to find where the cpu is stable.

 

But, as I'm sure you know, this board is a very hard board to learn. This is my first overclock with socket 775, and I didnt choose a beginner board for some stupid reason. :D

Link to comment
Share on other sites

You are warranted under Corsairs Lifetime warranty for up to and including 2.0v. That being said, I have not seen a great difference with any more than 1.8v and your modules.

 

What kind of overclock are you looking for? Approximately?

 

 

All I want to know is: Am I OK running these sticks at 1.6v? You said in a previous post that they are OK up to 1.7v. So I guess that is a yes?
Link to comment
Share on other sites

You are warranted under Corsairs Lifetime warranty for up to and including 2.0v. That being said, I have not seen a great difference with any more than 1.8v and your modules.

 

What kind of overclock are you looking for? Approximately?

 

I would be extremely happy with 1600 fsb (3.6ghz), but it is not stable even with a 1.52 vcore. I did think it was a RAM issue, as sometimes I am getting a "flash" of a blue screen with Parity error/ memory check, and then a restart. No matter how much NB and vcore I gave it the same happens over around 1520 fsb upwards.

 

Also, a bit more worrying is that ASUS PC Probe II is reporting a drop in memory voltage to 0.00v, and drop in SB voltage to 0.00v, and a drop in a fan speed to 0rpm. This happens intermittently, during stress testing. I know it cant be right, as the reported fan is still spinning happily, but then it recovers and shows correct value after 3 seconds.

Link to comment
Share on other sites

Only the very best Q6600's can make such an overclock. I would look closer to 3.2Ghz. Drop the multiplier to 8 and 400Mhz FSB = 3.2Ghz with 1600Mhz FSB. What cooler do you have? Do you still have the CPU box? If so, can you post the FPO/Batch number (on a white sticker on the box)?

 

Edit: The issues you are having with PC Probe show that you are clocking too highly with too high voltages. Electronic Migration is damaging components.

 

That's not a bad overclocker that can't make 3.6Ghz. If you are looking for such an overclock you need a newer revision CPU.

 

Download CoreTemp

Download CPU-z from
.

Please post a screenshot of Core Temp and CPU-z's CPU tab.

 

 

I would be extremely happy with 1600 fsb (3.6ghz), but it is not stable even with a 1.52 vcore. I did think it was a RAM issue, as sometimes I am getting a "flash" of a blue screen with Parity error/ memory check, and then a restart. No matter how much NB and vcore I gave it the same happens over around 1520 fsb upwards.

 

Also, a bit more worrying is that ASUS PC Probe II is reporting a drop in memory voltage to 0.00v, and drop in SB voltage to 0.00v, and a drop in a fan speed to 0rpm. This happens intermittently, during stress testing. I know it cant be right, as the reported fan is still spinning happily, but then it recovers and shows correct value after 3 seconds.

Link to comment
Share on other sites

Having checked, other software is not reporting this problem (HWmonitor, everest). It's run 1520 fsb (34200 mhz) prime95 small ffts for an hour. I ran memtest for 8 passes just fine. Temps are hitting 70 degrees on load, according to coretemp and realtemp.

 

Settings currently are:

 

Fsb: 1520

RAM: 1520

 

pcie:100

SPP<>MCP:200

 

RAM timing: 9-9-9-24-2T

 

Voltages:

Vcore: 1.45

Loadline Calibration: Enabled

VTT:1.32

VDIMM: 1.6

NB:1.44

SB:1.5

 

Cpu-z is reporting a vcore of 1.504 though. Due to loadline calibration being enabled, it doesnt droop on load anymore.

 

Are my voltages too high? Could they be causing electronic migration?

 

Again, thanks for your assistance, much appreciated.

Link to comment
Share on other sites

If other diagnostic softwares are not showing the voltage drop and spiking, then the issue is likely Probe.

 

A set of 3.4Ghz with 70C load on a G0 is certainly as high as I would go. I personally would not go higher than 65C but that's only a few C anyway. You are doing well already. If I knew your batch number I could tell better but if you need that kind of voltage it is not the LxxxBxxx stepping and those are mostly the ones that can more easily make the 3.6Ghz speed. They need far less voltage and have 1.2xxx VID (seen in Core Temp). If you have a higher VID, then you have the LxxxAxxx stepping and have made a very good overclock already. One hour Prime does not really show a good stability, but if you are not finding any crashes or bluescreens with your overclock, then it is fine.

 

If you were to move to an E8400, then you could move to 4.0Ghz fairly easily. Again, what cooler are you using?

 

Having checked, other software is not reporting this problem (HWmonitor, everest). It's run 1520 fsb (34200 mhz) prime95 small ffts for an hour. I ran memtest for 8 passes just fine. Temps are hitting 70 degrees on load, according to coretemp and realtemp.

 

Electronic migration is an amalgamation of temps and voltages. You are fine with the shown voltages, but your temps are certainly high enough that if you are using a poor cooler, you can damage the CPU.

 

Settings currently are:

 

Fsb: 1520

RAM: 1520

 

pcie:100

SPP<>MCP:200

 

RAM timing: 9-9-9-24-2T

 

Voltages:

Vcore: 1.45

Loadline Calibration: Enabled

VTT:1.32

VDIMM: 1.6

NB:1.44

SB:1.5

 

Cpu-z is reporting a vcore of 1.504 though. Due to loadline calibration being enabled, it doesnt droop on load anymore.

 

Are my voltages too high? Could they be causing electronic migration?

 

Again, thanks for your assistance, much appreciated.

Link to comment
Share on other sites

Sorry mate, Cooler is Noctua UH-12p with 2 fans in Push/Pull. Its in an Antec 1200 case, all fans on low. I've put them all on high now, and temps have dropp to 68.

 

I found the cpu box, batch no: L807A413. VID of 1.3v.

 

As for the problem with the varying voltages, I had reported it to ASUS long ago. Their reply was "If your overclock is stable, you should not worry".

 

I loaded defaults again, and it doesnt happen. So it must be electronic migration as you said. I lowered vcore from 1.425 to 1.4125, and it doesnt happen nearly half as much. Well, to be exact VTT was reported 3v for 1 second, and that has been the only episode in the last 10 minutes.

Link to comment
Share on other sites

Yes, that's an "A". With your system components I would set to 400Mzh FSB and an 8 X multiplier and 3.2Ghz with 1066Mhz memory. Then you are at Intel Spec.

 

Here's the thing. Intel SPec is 266MHz, 333Mhz and 400Mhz. Any value between the SPec is not a stable as the SPec. 267 - 332, 334 - 399 are out of SPec values. So you will do better with both voltages and temps with this value.

 

Test it out. It will still be very fast.

 

Sorry mate, Cooler is Noctua UH-12p with 2 fans in Push/Pull. Its in an Antec 1200 case, all fans on low. I've put them all on high now, and temps have dropp to 68.

 

I found the cpu box, batch no: L807A413. VID of 1.3v.

 

As for the problem with the varying voltages, I had reported it to ASUS long ago. Their reply was "If your overclock is stable, you should not worry".

 

I loaded defaults again, and it doesnt happen. So it must be electronic migration as you said. I lowered vcore from 1.425 to 1.4125, and it doesnt happen nearly half as much. Well, to be exact VTT was reported 3v for 1 second, and that has been the only episode in the last 10 minutes.

Link to comment
Share on other sites

What setting do you suggest? Should I leave RAM timings on auto? Or set 9-9-9-24 and 2t?

 

I have run 1600 x8 before, and that was what I was testing when my motherboard decided it did not recognize my OS installation. When it bluescreened before at 1600 fsb, when it tried to boot up I got the message "SYSTEM DISK ERROR, PLEASE INSERT SYSTEM DISK AND PRESS ENTER".

 

Its happened with 2 different hard drives and with 32 bit and 64 bit vista. I thought it had corrupted my hard drive, but if I set BIOS to boot with cd rom only, then let it try to boot, then reset, set it back to hard drive, it would boot as if nothing had happened.

 

Sorry if this seems stupid, I am just unsure of setting 1600 fsb again.

 

EDIT- Also, just noticed that it says 1.35v max. on the side of the cpu box. But it says 1.5v on intel's site?

 

EDIT 2- I just booted up with 1600x8, and you're right. I set vcore to 1.35 in BIOS, with loadline calibration disabled, and its passed 5 passes on intel burn test just now.

 

Current settings:

fsb:1600

ram:1066

 

Ram timings: 9-9-9-24-2T

 

Voltages:

Vcore: 1.35 (1.408 idle, 1.360 LOAD)

VDIMM: 1.6(1.66)

VTT:AUTO (1.42)

Link to comment
Share on other sites

Test for a while.

 

What setting do you suggest? Should I leave RAM timings on auto? Or set 9-9-9-24 and 2t?

 

I have run 1600 x8 before, and that was what I was testing when my motherboard decided it did not recognize my OS installation. When it bluescreened before at 1600 fsb, when it tried to boot up I got the message "SYSTEM DISK ERROR, PLEASE INSERT SYSTEM DISK AND PRESS ENTER".

 

Its happened with 2 different hard drives and with 32 bit and 64 bit vista. I thought it had corrupted my hard drive, but if I set BIOS to boot with cd rom only, then let it try to boot, then reset, set it back to hard drive, it would boot as if nothing had happened.

 

Sorry if this seems stupid, I am just unsure of setting 1600 fsb again.

 

EDIT- Also, just noticed that it says 1.35v max. on the side of the cpu box. But it says 1.5v on intel's site?

Link to comment
Share on other sites

Current settings:

fsb:1600

ram:1066

 

Ram timings: 9-9-9-24-2T

 

Voltages:

Vcore: 1.35 (1.408 idle, 1.360 LOAD)

VDIMM: 1.6(1.66)

VTT:AUTO (1.42)

 

It was running Prime95 SmallFFTs happily, but then I emptied the recycle bin, it froze, and gave me a blue screen Stop 0124.

 

more vcore?

 

EDIT - just did the same thing while browsing this site with 1.37 vcore in BIOS. froze, and blusescreened.

Link to comment
Share on other sites

I enabled Loadline, 1.35vcore. It blue screened 0124 again just after booting windows. On reboot with 1.375 vcore it blue screened again, this time saying 09C MACHINE_CHECK_EXCEPTION.

 

I then changed all RAM timings from 9-9-9-24-2T to AUTO, giving it 8-8-8-20-1T and its running fine.

 

What to do? Thanks.

Link to comment
Share on other sites

Play around with it like all of us enthusiasts do. You know your machine. There's no cut and dried settings. Just do it. :):

 

I enabled Loadline, 1.35vcore. It blue screened 0124 again just after booting windows. On reboot with 1.375 vcore it blue screened again, this time saying 09C MACHINE_CHECK_EXCEPTION.

 

I then changed all RAM timings from 9-9-9-24-2T to AUTO, giving it 8-8-8-20-1T and its running fine.

 

What to do? Thanks.

Link to comment
Share on other sites

Ok mate. Thanks for the help so far. With timings on AUTO it has been very happy for 30 minutes now. I will set prime95 to run tomorrow. I cant leave it overnight as its in my bedroom, and an Antec 1200 with lights and 9 120mm fans....

 

A question I wanted to ask is, is it ok to run blend on prime95 tomorrow? But, I wont be at home, I know temps will be OK, but what if it bsods around 10AM, I wont be home until 5PM. So it would be hanging like that for 7 hours. What do you think I should do? Is it OK to risk it?

Link to comment
Share on other sites

I have two of them. If it BSOD's it BSOD's. There no problem with it stuck on BSOD. :) :D:

 

Ok mate. Thanks for the help so far. With timings on AUTO it has been very happy for 30 minutes now. I will set prime95 to run tomorrow. I cant leave it overnight as its in my bedroom, and an Antec 1200 with lights and 9 120mm fans....

 

A question I wanted to ask is, is it ok to run blend on prime95 tomorrow? But, I wont be at home, I know temps will be OK, but what if it bsods around 10AM, I wont be home until 5PM. So it would be hanging like that for 7 hours. What do you think I should do? Is it OK to risk it?

Link to comment
Share on other sites

With 20 passes of burn test you should be fine. You can run either blend or small ffts. You've memtested and burn tested though.

 

1 more question before I go to bed. :D Should I run Blend or Small FFTs tomorrow? I have just run 20 passes Intel Burn Test, and 30 minutes of smalFFTs no problems.
Link to comment
Share on other sites

Archived

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


×
×
  • Create New...