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

[skuldarin]

OH! i may be missing it... but where is the PCH voltage control? ever since i built both my towers, i constantly have to unplug and replug my mouse back in piriotically and i was going to try increasing I/O voltages but i dont seem to see it. it also seems odd that my BIOS seems to always register 3 keyboards and 2 mice... when the only I/O connections i have are the ONE keyboard and the ONE mouse... sound goes through HDMI on my setup so i'm curious as to why it sees all these I/O connections.

[Raja@ASUS]

OH! i may be missing it... but where is the PCH voltage control?

ICH voltage control.

[zoson]

Installed 1301 per Raja's comments about tRL and Memory Recheck. With 1301 and Memory Recheck enabled, my memory is detected correctly as if I manually set tRL to -1/-1/-1.

Now when are we going to see a bios with the new marvell firmware?

[CryptiK]

Just thought I'd comment re: my voltage measuring before of vdimm which I measured as high vs BIOS. Well after a new DMM battery, its actually reading spot on what I set in BIOS to the nearest 0.005v. Much better than I thought before! Also makes my new ram kits appear much better than I thought when I first tested too as I was only feeding them 1.61v

[Raja@ASUS]

Just thought I'd comment re: my voltage measuring before of vdimm which I measured as high vs BIOS. Well after a new DMM battery, its actually reading spot on what I set in BIOS to the nearest 0.005v. Much better than I thought before! Also makes my new ram kits appear much better than I thought when I first tested too as I was only feeding them 1.61v

Might be worth going back and editing that other post where you thought the voltages were high and it had something to do with reviews.

[zinsco]

OK, I did it. Got my 4.8. Was with bios 1301.

[Hondacity]

1.6v + 80C temps + full load = degradation(fast or slow depending on your torture...)

stability tests on these newer cpus don't guarantee it will survive ...

[Rasmus66]

1,6 volt is quite high for 4,8 ghz.

Cant you go any lower then that?

[zinsco]

1.6v + 80C temps + full load = degradation(fast or slow depending on your torture...)

Not a 24/7 run. 4.5 @ about 1.43v with temps in the 60c's + full load are for my daily use.

1,6 volt is quite high for 4,8 ghz.

Cant you go any lower then that?

Nope, can't pass linx without really high vcore.

Here's my settings:

Target CPU Frequency                  4800MHz
Target DRAM Frequency                 1539MHz

LN2 Mode                              [Disabled]***
QPI Loadline Calibration              [Enabled]***

Ai Overclock Tuner                    [Manual]
OC From CPU level Up                  [Auto]
2D Performance Booster                [Disabled]
CPU Ratio Setting                     [25.0]
CPU Turbo Power Limit                 [Disabled]

>CPU Configuration----------------------------
CPU Ratio Setting                     [25.0]
C1E Support                           [Disabled]
Hardware Prefetcher                   [Enabled]
Adjacent Cache Line Prefetch          [Enabled]
MPS and ACPI MADT ordering            [Modern ordering]
Intel(R) Virtualization Tech          [Enabled]
CPU TM Function                       [Disabled]
Execute Disable Bit                   [Enabled]
Intel(R) HT Technology                [Disabled]	
*Active Processor Cores               [All]
A20M                                  [Disabled]
*Intel(R) SpeedStep(TM) Tech          [Disabled]
Intel(R) C-STATE Tech                 [Disabled]
C State package limit setting         []
C1 Auto Demotion                      []
C3 Auto Demotion                      []
----------------------------------------------

BCLK Frequency                        [192]
PCIE Frequency                        [105]
DRAM Frequency                        [DDR3-1539MHz]
UCLK Frequency                        [3079MHz]
QPI Link Data Rate                    [Auto]
		
Memory Recheck				[Enabled]
Memory Configuration Protect          [Disabled]

>DRAM Timing Control--------------------------
CAS# Latency                          [ 7 DRAM Clock]
RAS# to CAS# Delay                    [ 7 DRAM Clock]
RAS# PRE Time                         [ 7 DRAM Clock]
RAS# ACT Time                         [22 DRAM Clock]
RAS# to RAS# Delay                    [Auto]
REF Cycle Time                        [Auto]
WRITE Recovery Time                   [Auto]
READ to PRE Time                      [Auto]
FOUR ACT WIN Time                     [Auto]
Back-To-BackCAS# Delay                [Auto]

Timing Mode                           [1N]
Round Trip Latency on CHA             [Auto]
Round Trip Latency on CHB             [Auto]
Round Trip Latency on CHC             [Auto]

WRITE To READ Delay(DD)               [Auto]
WRITE To READ Delay(DR)               [Auto]
WRITE To READ Delay(SR)               [Auto]
READ To WRITE Delay(DD)               [Auto]
READ To WRITE Delay(DR)               [Auto]
READ To WRITE Delay(SR)               [Auto]
READ To READ Delay(DD)                [Auto]
READ To READ Delay(DR)                [Auto]
READ To READ Delay(SR)                [Auto]
WRITE To WRITE Delay(DD)              [Auto]
WRITE To WRITE Delay(DR)              [Auto]
WRITE To WRITE Delay(SR)              [Auto]
----------------------------------------------

CPU Differential Amplitude            [Auto]
CPU Clock Skew                        [Auto]
IOH Clock Skew                        [Auto]
------------ Extreme Engine Digi+ ------------
Digi+ PWR Mode                        [X-Power]
PWM Volt. Control                     [Auto]
Load-Line Calibration                 [Full Calibration]
CPU Voltage OCP                       [Enabled]
CPU PWM Frequency                     [Auto]

Extreme OV                            [Disabled]
Extreme OC                            [Auto]

CPU Voltage                           [1.59375]
CPU PLL Voltage                       [1.40450]		
QPI/DRAM Core Voltage                 [1.36250]	
DRAM Bus Voltage                      [Auto]	

>DRAM REF Voltages----------------------------
DRAM DATA REF Voltage on CHA          [Auto]
DRAM CTRL REF Voltage on CHA          [Auto]
DRAM DATA REF Voltage on CHB          [Auto]
DRAM CTRL REF Voltage on CHB          [Auto]
DRAM DATA REF Voltage on CHC          [Auto]
DRAM CTRL REF Voltage on CHC          [Auto]
----------------------------------------------

IOH Voltage                           [1.2]
IOH PCIE Voltage                      [Auto]

ICH Voltage                           [Auto]
ICH PCIE Voltage                      [Auto]

---------- Spread Spectrum Control -----------
CPU Spread Spectrum                   [Disabled]
PCIE Spread Spectrum                  [Disabled]



***on-board jumper dependant

[zinsco]

So am I reading this right? 2000 @ 8-8-8 is faster than 1800 @ 7-7-7.

[Hondacity]

yep

[sprayingmango]

*** EDIT: I found this thread on another forum, do you think this is safe?

http://www.bios-mods.com/forum/Threa...01-Mod-request




Is there a way to manually update to the newest Marvell BIOS firmware?

[zoson]

*** EDIT: I found this thread on another forum, do you think this is safe?

http://www.bios-mods.com/forum/Threa...01-Mod-request




Is there a way to manually update to the newest Marvell BIOS firmware?

Check the asus bios thread. MrToad has been doing such mods and people are reporting good things.

[CryptiK]

Might be worth going back and editing that other post where you thought the voltages were high and it had something to do with reviews.

Done, removed erroneous results and apologized for the insinuation. Seen it happen on older boards (P965/P45 era) but this one is spot on.

[CryptiK]

Was your uncore the same for both runs? If not, this difference may be mostly attributed to it being higher for the DDR3-2010MHz CL8 run. You have to have the same uncore to really compare ram clocks and their effect on bandwidth and SPI32M etc.

This is DDR3-1600 6-6-6-18 | 1T | tRFC 60 | 3600MHz uncore (my 24/7 settings) - this has better bandwidth than your screen of DDR3-1800 7-7-7-22 1T and is getting close to your bandwidth at DDR3-2010MHz 8-8-8-24 1T.

So am I reading this right? 2000 @ 8-8-8 is faster than 1800 @ 7-7-7.

http://dl.dropbox.com/u/22074005/Xtr...em%20bench.PNG

[zoson]

Was your uncore the same for both runs? If not, this difference may be mostly attributed to it being higher for the DDR3-2010MHz CL8 run. You have to have the same uncore to really compare ram clocks and their effect on bandwidth and SPI32M etc.

This is DDR3-1600 6-6-6-18 | 1T | tRFC 60 | 3600MHz uncore (my 24/7 settings) - this has better bandwidth than your screen of DDR3-1800 7-7-7-22 1T and is getting close to your bandwidth at DDR3-2010MHz 8-8-8-24 1T.

In his earlier screenshots we can see that he has been keeping Uncore at 2x. Maybe he's like me and can't get 4ghz uncore, but regardless... His bandwidth is very low for both of those settings.

I would guess that his subtimings are untuned.
tWR, tRFC, tRRD, tFAW, and tRTP are big ones! If I leave them on Auto, instead of 21k/19k/20k read/write/copy I get 18k/16k/17k with 2000MHz ram 3.6ghz Uncore.

[Leeghoofd]

AIDA , Everest readouts also depend a lot on how you configure the CPU bios settings ( C states and co ) Same when comparing different versions..; sometimes bandwith readouts are over 2K different

And strangely sometimes the gained boost by tweaking, hardly reflect in eg SuperPi 32...

[zinsco]

Was your uncore the same for both runs? If not, this difference may be mostly
attributed to it being higher for the DDR3-2010MHz CL8 run. You have to have the same uncore to really
compare ram clocks and their effect on bandwidth and SPI32M etc.

Do you mean multiple, like 2x or 1.5x? Changing the bclk to get the diff. ram speed changes the uncore.
What am I missing here? Can you just key in the speed you want and I just never tried, thinking you had
to chose from what is available in the drop down menu? I can't go see at the moment cause my daughter
is using it for an online class.

In his earlier screenshots we can see that he has been keeping Uncore at 2x.
Maybe he's like me and can't get 4ghz uncore, but regardless... His bandwidth is very low for both of those settings.

I would guess that his subtimings are untuned.
tWR, tRFC, tRRD, tFAW, and tRTP are big ones! If I leave them on Auto, instead of 21k/19k/20k read/write/copy
I get 18k/16k/17k with 2000MHz ram 3.6ghz Uncore.

Not quite 2x. I've been keeping it between 1.5x and 2x. Should I try to see if it will hold at 2x? As far as the subtimings,
ya, those are on auto. I don't know enough about tuning the ram.

[zoson]

Do you mean multiple, like 2x or 1.5x? Changing the bclk to get the diff. ram speed changes the uncore.
What am I missing here? Can you just key in the speed you want and I just never tried, thinking you had
to chose from what is available in the drop down menu? I can't go see at the moment cause my daughter
is using it for an online class.

Not quite 2x. I've been keeping it between 1.5x and 2x. Should I try to see if it will hold at 2x? As far as the subtimings,
ya, those are on auto. I don't know enough about tuning the ram.

Bloomfield used to require that uclk be 2x memory speed. Gulftown only requires uclk to be 1.5x memory speed. However, uclk is the speed of the memory controller, so if you increase uclk you will see better memory performance. Some gulftowns can hit 2x uclk, others like mine can not. Finding your uclk should be a separate endeavor from finding max bclk and max memory speed, then you dial it in as close as possible with your max bclk.

CPU-Tweaker can reveal the multiplier in use for your uclk. For example, with 200 bclk and 3600mhz uclk, you're using the 18x uclk multiplier(or 1.8x memory speed). This is right in line with 200 bclk with 2Ghz ram because you would use the x5 memory multiplier(ddr speed = multiplier x2), making 2x memory speed the 20x multiplier or 4000mhz.

As for your sub timings, if you click the spd button on cput, it will show you the extended subtimings for your memory. All I did was individually change them one at a time from auto to their xmp setting, then run memtest to see if it was stable. If it wasn't I would increment that setting by 1 in bios and repeat until I was fully stable with all my timings tightened as much as possible. Lots of trial and error, but it was a single sitting on a weekend.

I didn't bother with any timings not listed on the spd xmp profile.

[Andreezy]

since the topic is memory, this are my settings for a bloomfield and diff mobo. how can i increase the read/write speeds.

[l0ud_sil3nc3]

Just thought I'd comment re: my voltage measuring before of vdimm which I measured as high vs BIOS. Well after a new DMM battery, its actually reading spot on what I set in BIOS to the nearest 0.005v. Much better than I thought before! Also makes my new ram kits appear much better than I thought when I first tested too as I was only feeding them 1.61v

well that's good news

[donmarkoni]

AIDA , Everest readouts also depend a lot on how you configure the CPU bios settings ( C states and co ) Same when comparing different versions..; sometimes bandwith readouts are over 2K different

And strangely sometimes the gained boost by tweaking, hardly reflect in eg SuperPi 32...

I concur.

Here are my results with daily settings as example:

Target CPU Frequency		4600MHz
Target DRAM Frequency		2005MHz

LN2 Mode:			Disabled***
QPI Loadline Calibration:	Enabled***

Sync Mode			[Enabled] ** Gone since BIOS 0003/1301
AI Overclock Tuner		[Manual]
OC From CPU level Up		[Auto]
2D Performance Booster		[Disabled]
CPU Ratio Setting		[23.0]
CPU Turbo Power Limit		[Disabled]

>CPU Configuration----------------------------
CPU Ratio Setting		[23.0]
C1E Support			[Enabled]
Hardware Prefetcher		[Enabled]
Adjacent Cache Line Prefetcher	[Enabled]
MPS and ACPI MADT ordering	[Modern]
Intel(R) Virtualization Tech	[Enabled]
CPU TM Function			[Disabled]
Execute Disable Bit		[Enabled]
Intel(R) HT Technology		[Enabled]
Active Processor Cores		[All]
A20M				[Disabled]
Intel(R) SpeedStep(TM) Tech	[Enabled]
Performance/Watt select		[Power]
Intel(R) C-STATE Tech		[Enabled]
C State package limit setting	[Auto]
C1 Auto Demotion		[Enabled]
C3 Auto Demotion		[Enabled]
----------------------------------------------

BCLK Frequency			[200]
PCIE Frequency			[105]
DRAM Frequency			[2005]
UCLK Frequency			[4009]
QPI Link Data Rate		[7218]

Memory Recheck			[Disabled] ** BIOS 1301
Memory Configuration Protect	[Disabled]

>DRAM Timing Control--------------------------
CAS# Latency		      7	[7]
RAS# to CAS# Delay	      7	[7]
RAS# PRE Time		      7	[7]
RAS# ACT Time		     20	[20]
RAS# to RAS# Delay	      4	[4]
REF Cycle Time		     72	[72]
WRITE Recovery Time	     10	[10]
READ to PRE Time	      6	[6]
FOUR ACT WIN Time	     20	[20]
Back-To-BackCAS# Delay	      0	[Auto]

Timing Mode		     1N	[1N]
Round Trip Latency on CHA    53	[53]
Round Trip Latency on CHB    55	[55]
Round Trip Latency on CHC    57	[57]
tRL Channel A			[-2] ** BIOS 0003
tRL Channel B			[-2] ** BIOS 0003
tRL Channel C			[-2] ** BIOS 0003

WRITE To READ Delay(DD)	      8	[Auto]
WRITE To READ Delay(DR)	      8	[Auto]
WRITE To READ Delay(SR)	     19	[Auto]
READ To WRITE Delay(DD)	      8	[Auto]
READ To WRITE Delay(DR)	      8	[Auto]
READ To WRITE Delay(SR)	      8	[Auto]
READ To READ Delay(DD)	      7	[Auto]
READ To READ Delay(DR)	      6	[Auto]
READ To READ Delay(SR)	      4	[Auto]
WRITE To WRITE Delay(DD)      7	[Auto]
WRITE To WRITE Delay(DR)      7	[Auto]
WRITE To WRITE Delay(SR)      4	[Auto]
----------------------------------------------

CPU Differential Amplitude	[Auto]
CPU Clock Skew			[Auto]
IOH Clock Skew			[Auto]
------------ Extreme Engine Digi+ ------------
Digi+ PWR Mode			[T-Balanced]
PWR Volt. Control		[Auto]
Load-Line Calibration		[Full]
CPU Voltege OCP			[Disabled]
CPU PWM Frequency		[Auto]

Extreme OV			[Enabled]*
Extreme OC			[Auto]

CPU Voltage		1.415	[1.42500]
CPU PLL Voltage 	1.349	[1.35150]
QPI/DRAM Core Voltage	1.376	[1.38750]
DRAM Bus Voltage	1.773	[1.77550]

>DRAM REF Voltages----------------------------
DRAM DATA REF Voltage on CHA	[Auto]
DRAM CTRL REF Voltage on CHA	[Auto]
DRAM DATA REF Voltage on CHB	[Auto]
DRAM CTRL REF Voltage on CHB	[Auto]
DRAM DATA REF Voltage on CHC	[Auto]
DRAM CTRL REF Voltage on CHC	[Auto]
----------------------------------------------

IOH Voltage		1.204	[1.20575]
IOH PCIE Voltage	1.508	[Auto]

ICH Voltage		1.111	[Auto]
ICH PCIE Voltage		[Auto]

*QPI Voltage OCP		[Disabled]
*IOH Voltage OCP		[Enabled]
*DRAM Voltage OCP		[Disabled]
*QPI PWM Frequency		[Auto]
*IOH PWM Frequency		[Auto]
*DRAM PWM Frequency		[Auto]

---------- Spread Spectrum Control -----------
CPU Spread Spectrum		[Disabled]
DRAM Spread Spectrum		[Disabled]



***on-board jumper dependant

[Hondacity]

glad to see aida being used ....

[Andreezy]

i played around taking Zozon's advice of using memory's XMP settings and punching numbers manually and enabling C states.
i have to say i under estimated my memory.






P.S.......
sorry for posting here results of a diff motherboard.

[zinsco]

I concur.

Here are my results with daily settings as example:

DRAM Bus Voltage 1.773 [1.77550]

Isn't that voltage too high?



Edit: Oh ya, I remember now. Just have to keep the DRAM Bus and QPI/DRAM Core voltages within 0.5V