***Asus Rampage III Extreme Owners Thread***

[Tergon]

Out most of tomorrow will give it another run with all memory selected soon.

nice chip you got there, regardless it being only 2gb test under LinX.
~1.38Vcore is very nice @ +4.4Ghz, you need to retest under LinX and select all memory to fully test your hardware.
if you test again with all memory, please let us know

[zoson]

Out most of tomorrow will give it another run with all memory selected soon.

I'm also on the edge of my seat to see your all memory run. Please be sure it's using >9216mb memory (optimally >10gb).

[Tergon]

The system froze after 5.3 minutes, hard lock forced hardware reboot. Will have to tweak a few more settings any suggestions or if you need info to tell let me know. Had to enable Swap file to get the program to run, I hate swap files.

I'm also on the edge of my seat to see your all memory run. Please be sure it's using >9216mb memory (optimally >10gb).

[zoson]

The system froze after 5.3 minutes, hard lock forced hardware reboot. Will have to tweak a few more settings any suggestions or if you need info to tell let me know. Had to enable Swap file to get the program to run, I hate swap files.

You can probably get around the swap issue by enabling diagnostic startup using MSCONFIG.

I think Andreezy is spot on, you probably do not have enough vCore. 1.38v is very low for 4.4GHz... A bit more volts there would most likely do you good.

With 12GB Ram, 1.25v QPI also seems low to me... Generally the rule I've been following is set QPI to 1.35v and then see how high your UCLK can go.

For 143 BCLK, you don't need to increase PCIE speed, 100 should be fine.

Lastly, your IOH is high. I use 1.2v IOH for 200BCLK with 2x GTX560 Ti's in SLI.

[NathObeaN]

Quick question!

Does this motherboard support dynamic voltage when overclocking? I have overclocked my i7 930 to 4.2GHz, enabled SpeedStep, C-State and C1E but when the system idles, only the multi drops, the vCore stays at the same voltage? Isn't the voltage supposed to drop?

[Tergon]

There you go guys all that low voltage and stable still

You can probably get around the swap issue by enabling diagnostic startup using MSCONFIG.

I think Andreezy is spot on, you probably do not have enough vCore. 1.38v is very low for 4.4GHz... A bit more volts there would most likely do you good.

With 12GB Ram, 1.25v QPI also seems low to me... Generally the rule I've been following is set QPI to 1.35v and then see how high your UCLK can go.

For 143 BCLK, you don't need to increase PCIE speed, 100 should be fine.

Lastly, your IOH is high. I use 1.2v IOH for 200BCLK with 2x GTX560 Ti's in SLI.

[zoson]

There you go guys all that low voltage and stable still

It was stable because your CPU did not get properly loaded. 53 gflops is dismally low for 4.4gflops. With 9216 mem at 4.2ghz I have 83gflops.

[zoson]

Gflops mean nothing... it's like bandwith readouts with Everest boosted by C states... test the IMC with HyperPi32M 12 threads...

No, this is the issue me and Andreezy have been talking about. Except it's happening to him at 9216 instead of ~10gb.
I can clock my cpu up to 4.5 with 1.425vcore and it appears stable as long as I'm using >9216. The difference in gflops is 80 -> 55. His config is so similar to mine that he should have very similar results, except slightly faster.

It's easy to see the difference because the cpu cores won't get as hot as they normally would.
vs

[Leeghoofd]

So the issue is the Linx coding prolly ? (deleted my previous post , my bad )

[Andreezy]

So the issue is the Linx coding prolly ? (deleted my previous post , my bad )

there is a thread started by zozon, its not Linx Coding.
i have a few kits that worked @ full bandwidth 100% load under LinX used
to output 79/81gflops.

same test as of *today* gives a max Gflops output of 55/56gflops.
so its some bug/degradation going on.

Gulftown is Bugtown, look @990X above by Tergon 100% load and i bet his temps are 70's
seems like sweetspot, but its not....

i bet same test with 4GB will brings his cores +80c and or crash his cpu.
Leeghoofd you need a GT and test 12GB of ram to come into this problem,
so far its a hit and then miss, problem is killing us here

[Leeghoofd]

thta's why I didn't include a Linx screenshot in my 12GB Sniper review... took too long...

[Andreezy]

if you did the review with a gulftown and took too long then maybe you werent stable, when i was stable and full bandwidth the system flew through the test.
now that the gulftown is capped, my i7 950 gives it a run for its money on full load.


this is a screencap on my current system, i yanked the gulftown..
its never going to see the daylight for a while, unless i sell it..

[Tergon]

The last test posted was with 9216gb selected and my temps were only in the mid to high 60's to low 70's, made a few adjustments to settings QPI to 1.26 in BIOS and uncore value to 3254 up from min of 3011 or so. Another run through and at two tests done (will post screen shot if it all finished) the temps are running hotter at the lowest core at 72c and highest at 82c. The GFlops are now up to 83.9 and 84.2. system seems stable as I am posting at the same time on it. The only other thing I did was take the side off the case for a little breathing room. Hope it makes this run, no adjustment to vcore at 1.37 in BIOS. The test is fast with the higher GFLops each line before was taking 11 minutes and now down to 9 each calculation. Will post screen shot if it finishes without error.

No, this is the issue me and Andreezy have been talking about. Except it's happening to him at 9216 instead of ~10gb.
I can clock my cpu up to 4.5 with 1.425vcore and it appears stable as long as I'm using >9216. The difference in gflops is 80 -> 55. His config is so similar to mine that he should have very similar results, except slightly faster.

It's easy to see the difference because the cpu cores won't get as hot as they normally would.
vs

[Tergon]

I thought 10 loops would be good to go, new found stablity, now have to shoot for a higher overclock with this chip. Check screen shot max memory selected 10gigs 4.48GHz swap file enabled.

The last test posted was with 9216gb selected and my temps were only in the mid to high 60's to low 70's, made a few adjustments to settings QPI to 1.26 in BIOS and uncore value to 3254 up from min of 3011 or so. Another run through and at two tests done (will post screen shot if it all finished) the temps are running hotter at the lowest core at 72c and highest at 82c. The GFlops are now up to 83.9 and 84.2. system seems stable as I am posting at the same time on it. The only other thing I did was take the side off the case for a little breathing room. Hope it makes this run, no adjustment to vcore at 1.37 in BIOS. The test is fast with the higher GFLops each line before was taking 11 minutes and now down to 9 each calculation. Will post screen shot if it finishes without error.

[Andreezy]

good going, those are good outputs.
let us know what and if anything like a cap limit pops up.

[Brama]

Quick question!

Does this motherboard support dynamic voltage when overclocking? I have overclocked my i7 930 to 4.2GHz, enabled SpeedStep, C-State and C1E but when the system idles, only the multi drops, the vCore stays at the same voltage? Isn't the voltage supposed to drop?

I asked to Asus VIP forum and I was told that with the hybrid VRM circuitry of R3E it is impossible.
In R2E was possible, using the "relative" voltage option in BIOS e setting properly the offset.

[zalbard]

The last test posted was with 9216gb selected and my temps were only in the mid to high 60's to low 70's, made a few adjustments to settings QPI to 1.26 in BIOS and uncore value to 3254 up from min of 3011 or so. Another run through and at two tests done (will post screen shot if it all finished) the temps are running hotter at the lowest core at 72c and highest at 82c. The GFlops are now up to 83.9 and 84.2. system seems stable as I am posting at the same time on it. The only other thing I did was take the side off the case for a little breathing room. Hope it makes this run, no adjustment to vcore at 1.37 in BIOS. The test is fast with the higher GFLops each line before was taking 11 minutes and now down to 9 each calculation. Will post screen shot if it finishes without error.

Nice improvement, looks decent!
Now to go over 4.5GHz.

[Tergon]

I have tried many things to get a faster overclock stable, as so as I go over 4.5 the system craps out with no matter what the settings are. Any hints would be helpful.
I have tried the following,
Higher Vcore
Higher QPI
Higher IOH
Higher BLCK Vs Higher Multi for same clock rate
How much does CPU PLL have to play into it? I found 1.81 (board default to high) dropped to 1.65 is this to low. Should I try to up the CPU PLL!! So many bloody variables to take into account, fun tweaking but gets tedious at times, you just want it to work. Also contemplating going to water cooling, but will this just lower the temps, doesn't neccessary mean an instantly better or higher overclock. Lots of decisions to try.

[Andreezy]

Nice improvement, looks decent!
Now to go over 4.5GHz.

dont listen to zalbard he is evil

j/k

summertime is here so everyone is kind of detuning their systems.
you have a nice chip/vcore going on, dont ruin it
optimize your memory and voltages and as you learn your hardware
you will figure out what it takes.

do a bios template and post it up, it will help you in the future for
reference if anything goes.

[zoson]

I have tried many things to get a faster overclock stable, as so as I go over 4.5 the system craps out with no matter what the settings are. Any hints would be helpful.
I have tried the following,
Higher Vcore
Higher QPI
Higher IOH
Higher BLCK Vs Higher Multi for same clock rate
How much does CPU PLL have to play into it? I found 1.81 (board default to high) dropped to 1.65 is this to low. Should I try to up the CPU PLL!! So many bloody variables to take into account, fun tweaking but gets tedious at times, you just want it to work. Also contemplating going to water cooling, but will this just lower the temps, doesn't neccessary mean an instantly better or higher overclock. Lots of decisions to try.

Different CPU's have different needs for CPU PLL. My CPU likes 1.25875v CPU PLL. I've seen many others saying all sorts of values. 1.35, 1.4, 1.45, 1.55, etc...
With higher BCLK and Uncore speed you *will* need to increase your QPI/DRAM voltage.

[tistou77]

except setting "trl" for 0003, it requires less tension, more stable, etc... than the 1301 or not? (I leave "trl" on AUTO)

thanks

[gotendbz1]

I seem to have a weird problem. after reappling tim on my cpu(took out cpu and reseated) my memory channels are all messed up. I had 6x2gb installed before, but now it seems that the only way i can get it to work is only have ram in channels a1/b1. If i have ram in a1/b1/c1(red one's) slot the system will only see 2gb as installed.

I can't get it to boot with any other memory combo beside channels a1/b1, but once it boots up it work fine, as i can run all app/play games and currently writing this on said machine.

what could the problem be? or is a rma needed.

[CryptiK]

Try reseating your CPU.

[zoson]

Also, when you reseat your cpu, check your LGA for bent pins.

[gotendbz1]

i had to remove everything, took out mobo check for bent pins, cleaned and reinstalled and it working now.

also what score should i be getting from a stock 975 with 12gb of ram with IBT standard bench?