P5K-E Speedstep + Overclocking
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
P5K-E Speedstep + Overclocking
Is there anyone out there that is using Speedstep on a P5K-E while overclocking a CPU? More importantly, does it work properly?
On idle, not only should the multiplier drop, the vcore should also drop. I have verified this to work on my Gigabyte P35-DS4 but the board is some 40 bucks more expensive than the P5K-E which has everything the DS4 had. There lies my dilemma.
EDIT: P35-DS4 + Q6600 @ 3.00 GHz stock volts - SPEEDSTEP works properly
Almost everywhere I go, everyone says the same thing. "Turn off Speedstep!" With electricity bills becoming more expensive, I wish to keep power consumption to a minimum. I also want to keep temperatures to a minimum since that room can reach 35+ degrees celsius in the summer.
I plan to use it with a E8400 cpu. Any ideas whether Speedstep will work properly on this board?
On idle, not only should the multiplier drop, the vcore should also drop. I have verified this to work on my Gigabyte P35-DS4 but the board is some 40 bucks more expensive than the P5K-E which has everything the DS4 had. There lies my dilemma.
EDIT: P35-DS4 + Q6600 @ 3.00 GHz stock volts - SPEEDSTEP works properly
Almost everywhere I go, everyone says the same thing. "Turn off Speedstep!" With electricity bills becoming more expensive, I wish to keep power consumption to a minimum. I also want to keep temperatures to a minimum since that room can reach 35+ degrees celsius in the summer.
I plan to use it with a E8400 cpu. Any ideas whether Speedstep will work properly on this board?
Re: P5K-E Speedstep + Overclocking
Let me chime in here. Turn off speedstep if you are overclocking. Speedstep kills performance too especially memory performance since it relies on CPU for that. Run some tests with it enabled and disabled and compare those 2...jimmyzaas wrote:Is there anyone out there that is using Speedstep on a P5K-E while overclocking a CPU? More importantly, does it work properly?
On idle, not only should the multiplier drop, the vcore should also drop. I have verified this to work on my Gigabyte P35-DS4 but the board is some 40 bucks more expensive than the P5K-E which has everything the DS4 had. There lies my dilemma.
EDIT: P35-DS4 + Q6600 @ 3.00 GHz stock volts - SPEEDSTEP works properly
Almost everywhere I go, everyone says the same thing. "Turn off Speedstep!" With electricity bills becoming more expensive, I wish to keep power consumption to a minimum. I also want to keep temperatures to a minimum since that room can reach 35+ degrees celsius in the summer.
I plan to use it with a E8400 cpu. Any ideas whether Speedstep will work properly on this board?
There is a reason why everybody (mostly) is turning it off.
Wait a minute.
Unless I have something seriously wrong with my memory, the memory performance has nothing to do with the intel CPU. Only AMDs have memory controllers integrated into them at the moment. The only thing that could affect memory performance would be the bus speed.
How would Speedstep be any different on an overclocked machine versus a stock machine? By overclocked, I mean a minor one that stock volts can hold (Sorry.. I forgot to mention this). And not something that would require me to push the CPU to 1.6 vcore or some ridiculous value. Most of the users that require C1E/Speedstep disabled are indeed doing much higher overclocks at non stock voltage.
So far I have read one post of a user on SPCR doing exactly what I did on a DS4. Stock volts overclock + proper speedstep. There must be someone that is doing the same with a P5K-E... I hope.
Unless I have something seriously wrong with my memory, the memory performance has nothing to do with the intel CPU. Only AMDs have memory controllers integrated into them at the moment. The only thing that could affect memory performance would be the bus speed.
How would Speedstep be any different on an overclocked machine versus a stock machine? By overclocked, I mean a minor one that stock volts can hold (Sorry.. I forgot to mention this). And not something that would require me to push the CPU to 1.6 vcore or some ridiculous value. Most of the users that require C1E/Speedstep disabled are indeed doing much higher overclocks at non stock voltage.
So far I have read one post of a user on SPCR doing exactly what I did on a DS4. Stock volts overclock + proper speedstep. There must be someone that is doing the same with a P5K-E... I hope.
AFAIK if you push the FSB too high, the board will automatically overvolt the CPU Voltage... even if the CPU Voltage is set to "Auto"... damn you ASUS !
I had a E4400 CPU (FSB800) and Speedstep was fine up to 280MHz (1120MHz) or 40%, anything beyond that CPU voltage wont drop, the board will take over the CPU voltage and overvolt it.
Here's something interesting at X-bit labs:
http://www.xbitlabs.com/articles/mainbo ... html#sect0
I had a E4400 CPU (FSB800) and Speedstep was fine up to 280MHz (1120MHz) or 40%, anything beyond that CPU voltage wont drop, the board will take over the CPU voltage and overvolt it.
Here's something interesting at X-bit labs:
http://www.xbitlabs.com/articles/mainbo ... html#sect0
Yeah GB doesnt have such limitation.
It will however disable Speedstep if you manually choose a CPU Voltage, just like all other boards.
I noticed that the ASUS board tends to override the "Auto" option when the FSB is 40% beyond stock, at least that's what happened with the FSB800 CPU.
The Q6600 is a FSB1066 CPU so Speedstep should work up to 1489.6MHz FSB if my theory is right.
Well, but my theory could be wrong and the board might actually be programmed to overvolt the CPU when the FSB is over 1120MHz.
You'll have to test. I couldnt get ahold of any FSB1066 CPUs back then.
GB P35 boards in my opinion have better overall features, BIOS, overclocking capabilities, build quality, sound codecs.
They lack extensive fan control though.
It will however disable Speedstep if you manually choose a CPU Voltage, just like all other boards.
I noticed that the ASUS board tends to override the "Auto" option when the FSB is 40% beyond stock, at least that's what happened with the FSB800 CPU.
The Q6600 is a FSB1066 CPU so Speedstep should work up to 1489.6MHz FSB if my theory is right.
Well, but my theory could be wrong and the board might actually be programmed to overvolt the CPU when the FSB is over 1120MHz.
You'll have to test. I couldnt get ahold of any FSB1066 CPUs back then.
GB P35 boards in my opinion have better overall features, BIOS, overclocking capabilities, build quality, sound codecs.
They lack extensive fan control though.
-
- Posts: 42
- Joined: Mon Apr 10, 2006 6:04 pm
- Location: Montreal, Canada
The 40% threshold might not apply to all boards I guess.
The P5B-PLUS I had certainly had that threshold with E4500 CPU.
I also heard that ASUS P5E-VM HDMI will set a fixed CPU voltage as soon as you overclock, please check this thread over RightMark Forums.
As for the pin mod, it didnt work with the P5B Plus as the board wouldnt recognize the forced FSB1333MHz strap and would always boot using the defautl FSB800 strap.
Yes using SetFSB would be good but I also had S3 standby issues with high FSB, so that didnt help much.
The P5B-PLUS I had certainly had that threshold with E4500 CPU.
I also heard that ASUS P5E-VM HDMI will set a fixed CPU voltage as soon as you overclock, please check this thread over RightMark Forums.
As for the pin mod, it didnt work with the P5B Plus as the board wouldnt recognize the forced FSB1333MHz strap and would always boot using the defautl FSB800 strap.
Yes using SetFSB would be good but I also had S3 standby issues with high FSB, so that didnt help much.
-
- Posts: 42
- Joined: Mon Apr 10, 2006 6:04 pm
- Location: Montreal, Canada
There might be a 40% threshold, I wouldn't know since I always went straight to FSB 333.
When using SetFSB, the only issue I have is that when resuming from S3 standby, the FSB falls back to its default of 266. So I just execute SetFSB again when needed. Alternatively, you can run a script automatically when resuming from standby.
When using SetFSB, the only issue I have is that when resuming from S3 standby, the FSB falls back to its default of 266. So I just execute SetFSB again when needed. Alternatively, you can run a script automatically when resuming from standby.