Not being an avid gamer and primarily focused on building production boxes, never spent much time seeing what I can get outta any individual box.....if it took more than 15-20 minuted configuring BIOS, I wasn't interested. MY usual approach was to use built in BIOS OC settings to get that 20% OC and then try and tone down things a bit to make it run a little cooler w/o sacrificing that 20%.
With my son building his 1st box paid for w/ his own money, this gave us a project to work on together which, as any Dad will tell ya, 18 year old kids asking to spend time w/ Dad is a rarity to be treasured . So after a month of chasing down Event Viewer errors w/ Win7-64, the 4 day weekend was our time to try and see what we could do on this.
Have downloaded all the OC guides and the R2E MoBo gives us plenty of tools. He blinged it up w/ an OC Station which I only went along with cause it was a great fan controller and seeing as he spent his own money, he was entitled to a little bling.
System includes R2E MoBo, i7-920 (DO), Mushkin CAS 6 (998692).
Here's the approach we took .....
Used built in OC profile to get to 3.2 GHz using "CPU Level Up" feature. That left us w/ the following:
CPU Multiplier - Auto
BCLK - 160
CPU Voltage - 1.3125 (reads 1.304 in CPU-z)
QPI/DRAM Core Voltage - 1.3875
DRAM Bus Voltage - 1.56406
memtest successful for 18 hours
Only other BIOS changes (that come to mind atm) were to turn off HT and set CAS timings to 6-7-6-18. Thought the CPU Voltage and QPI voltage was a little high, and RAM voltage a little low, but figured I'd chill those out later. was looking to get to 3.66 for everyday use but also curious as to how far it could go.
Jumped to BCLK of 165, no issues ... continued w/ no changes to BIOS other than jumping BCLK by 3 and testing w/ prime 95 for an hour till I got to 174 which, w/ 21 multiplier, gives me the 3.66 we were going for.
Got to 177 w/ no apparent issues (see later note) then tried 180 w/ no luck. Upped voltages incrementally couple of time, then just went all out and did:
CPU Voltage - 1.375
QPI/DRAM Core Voltage - 1.3875
DRAM Bus Voltage - 1.64xxx
Still didn't run at 180 till I set multiplier to 20 which is slower than 177 x 21. Rebooted to 177 BCLK and after some forum researching was typing here on a temperature thread and BSOD'd. As I type this am at BCLK 174 w/ following voltages / temps as per Asus Probe
CPU 1.35 (CPU-z says 130.5, Speedfan says 1.30)
CPU PLL 1.81
QPI/DRAM 1.34
DRAM Bus 1.56
CPU 60C (60 per HW Monitor & SpeedFan)
MB 25C
NB 45C
SB 40C
SpeedFan indicates core temps of 58-61 which sounds off but so does CoreTemp and RealTemp's 74/72/71/71 ...as i recall shouldn't be more than 5C between CPU Temp and core Temps.
Failures were determined by error messages in prime95 and / or any unusual messages upon landing in Windows such as some service not starting.
Oh and another odd note....if we boot into Windows and use the OC Station or Asus TurboV software to boost voltages and BCLK we can easily get to 4.0 GHz. Prime 95 runs stable
Tho satisfied to hit our target did hope to learn some more about more aggressive OC'ing. Mostly was surprised to get that high with those low voltages but even more surprised to that jumping them up those large increments at the end did nothing to improve things. Any ideas where we might be choking ourselves off ?
With my son building his 1st box paid for w/ his own money, this gave us a project to work on together which, as any Dad will tell ya, 18 year old kids asking to spend time w/ Dad is a rarity to be treasured . So after a month of chasing down Event Viewer errors w/ Win7-64, the 4 day weekend was our time to try and see what we could do on this.
Have downloaded all the OC guides and the R2E MoBo gives us plenty of tools. He blinged it up w/ an OC Station which I only went along with cause it was a great fan controller and seeing as he spent his own money, he was entitled to a little bling.
System includes R2E MoBo, i7-920 (DO), Mushkin CAS 6 (998692).
Here's the approach we took .....
Used built in OC profile to get to 3.2 GHz using "CPU Level Up" feature. That left us w/ the following:
CPU Multiplier - Auto
BCLK - 160
CPU Voltage - 1.3125 (reads 1.304 in CPU-z)
QPI/DRAM Core Voltage - 1.3875
DRAM Bus Voltage - 1.56406
memtest successful for 18 hours
Only other BIOS changes (that come to mind atm) were to turn off HT and set CAS timings to 6-7-6-18. Thought the CPU Voltage and QPI voltage was a little high, and RAM voltage a little low, but figured I'd chill those out later. was looking to get to 3.66 for everyday use but also curious as to how far it could go.
Jumped to BCLK of 165, no issues ... continued w/ no changes to BIOS other than jumping BCLK by 3 and testing w/ prime 95 for an hour till I got to 174 which, w/ 21 multiplier, gives me the 3.66 we were going for.
Got to 177 w/ no apparent issues (see later note) then tried 180 w/ no luck. Upped voltages incrementally couple of time, then just went all out and did:
CPU Voltage - 1.375
QPI/DRAM Core Voltage - 1.3875
DRAM Bus Voltage - 1.64xxx
Still didn't run at 180 till I set multiplier to 20 which is slower than 177 x 21. Rebooted to 177 BCLK and after some forum researching was typing here on a temperature thread and BSOD'd. As I type this am at BCLK 174 w/ following voltages / temps as per Asus Probe
CPU 1.35 (CPU-z says 130.5, Speedfan says 1.30)
CPU PLL 1.81
QPI/DRAM 1.34
DRAM Bus 1.56
CPU 60C (60 per HW Monitor & SpeedFan)
MB 25C
NB 45C
SB 40C
SpeedFan indicates core temps of 58-61 which sounds off but so does CoreTemp and RealTemp's 74/72/71/71 ...as i recall shouldn't be more than 5C between CPU Temp and core Temps.
Failures were determined by error messages in prime95 and / or any unusual messages upon landing in Windows such as some service not starting.
Oh and another odd note....if we boot into Windows and use the OC Station or Asus TurboV software to boost voltages and BCLK we can easily get to 4.0 GHz. Prime 95 runs stable
Tho satisfied to hit our target did hope to learn some more about more aggressive OC'ing. Mostly was surprised to get that high with those low voltages but even more surprised to that jumping them up those large increments at the end did nothing to improve things. Any ideas where we might be choking ourselves off ?