Slow cpu clock on my Latitude E4300 in windows 7 x64. 665Mhz instead of 2400. High Performance in power options, AC power..

If you use ThrottleStop and check off the Power Saver feature and the SLFM feature then your CPU will run slower at idle and run at full speed when loaded.

But do you know what? This probably won't make any significant difference to power consumption. Most 45nm Intel CPUs enter the C3 state when idle so the MHz and voltage settings are ignored. The CPU is using a lower voltage internally than what any software will show you.

If you want to test for this then run the Windows Performance Monitor and monitor your battery power usage. At idle, whether you use the highest VID voltage and multiplier or the lowest VID and multiplier makes no difference because the CPU ignores these settings anyhow. I'll post a couple of screen shots of this so you can understand this better.

It sounds like there is a problem in your bios and maybe it doesn't correctly support this CPU. Did this CPU originally come with this laptop? Can you post a CPU-Z screen shot when it is running at 665 MHz? Just upload an image to http://www.imageshack.us or anywhere convenient.

Edit: Here's what my Dell with a T8100 does at idle. It spends over 99% of its time in the C3 low power state.



If I set the multiplier as high as it can go and I set the VID voltage as high as it can go I get this for power consumption.



Now if I enable all the power saving features and slow my CPU down to a crawl and reduce the VID voltage as low as it can go I get this for power consumption.



Do you see any huge difference? The average data is in milliwatts so divide that by 1000 and you get a savings of less than 0.1 watts. (12.503W DC vs 12.426W DC average) Power consumption constantly varies a tiny amount depending on background processes so that is within the margin of error.

At idle there isn't any significant difference anymore. There used to be when Intel first released the 65nm Core 2 mobile CPUs but Intel changed how these things operate when they released the newer 45nm Core 2 mobile CPUs like your SP9400. It's not worth worrying about these settings anymore because these CPUs should enter a low power state automatically. Maybe there's a reason why Intel got rid of the Super Low Frequency Mode (SLFM) feature on the new Core i CPUs. It's not needed anymore because it makes virtually zero difference to power consumption.

If you can disable SpeedStep in the bios then you should look into using this trick to speed your computer up. It won't consume any more power at idle as I've proven here.

http://forum.notebookreview.com/win...c-acceleration-ida-both-cores-core-2-duo.html

 

1. BIOS: A17

2. Started High load test... temp increased to 93C (199F) in 4 minutes .. and i stopped it. (... i need to do smth. with this)

3. 90 watt block

 

I made this stress test on normal clock... 2400 mhz

 

unclewebb,
thanks for detailed explanation, here is the screen:


I'm really confused with this... so only thing i can do - is wait for new bios or windows patch....?

 

Don't worry, I'm confused too.

Is this SP9400 the original CPU? It's not an Engineering Sample (ES) so it should work correctly. It seems like a bios issue where the CPU is not set up correctly when the bios transfers control over to Windows.

In the Core 2, the minimum multiplier is 6.0. When Super Low Frequency Mode (SLFM) is used, this cuts the bus speed in half but some software like CPU-Z reports that the multiplier is half as much as normal.

At idle the CPU does this

6.0 x 266.0 MHz = 1596.0 MHz

when SLFM starts, that drops the bus speed in half so the CPU should be running at this speed:

6.0 x 133.0 MHz = 798.0 MHz

CPU-Z normally reports this as:

3.0 x 266.0 MHz = 798.0 MHz

It is physically impossible for a Core 2 CPU to have a multiplier lower than this.

When CPU-Z reports a multiplier lower than 3.0, that's usually a sign that clock modulation is being used inside the CPU to internally slow it down. There is no reason why the CPU should be set up like this when you first boot up so something is definitely wrong.

ThrottleStop is an anti-throttling tool so as soon as you start it, even in monitoring mode, it immediately disables all clock modulation and sets that to 100.0% and it also turns off SLFM mode. It makes sense that your CPU immediately jumps up to at least 1596 MHz as soon as you run ThrottleStop.

There are two registers in your CPU that control the multiplier, voltage and clock modulation setting. When you first boot up, without running ThrotleStop, could you read these registers and post a picture of what the contents of those two registers are? You can use my MSR Tool to read values of the model specific registers within your CPU.

http://www.fileden.com/files/2008/3/3/1794507/MSR.zip

MSR 0x199 contains the request multiplier and voltage. MSR 0x19A contains the clock modulation settings.

Download my tool, start it up and at the bottom is a MSR Number box. Enter 0x199 in that box and click on the Read MSR button and it will read what that register contains. Do the same thing for MSR 0x19A. You can run two separate instances of MSR Tool if you want so you only need to post one screen shot showing both registers.

I'm not sure if you are familiar with under volting but it might be a good idea to look into this. It involves lowering your core voltage which can really help reduce heat at full load. You need to do some Prime95 testing as you go lower to make sure you don't go too low and lose stability. If it was my laptop, I'd also pull it apart and redo the paste between the CPU and the heatsink and make sure it is making firm contact to improve the heat transfer.

For a separate test, with ThrottleStop running and set to the highest multiplier, 9.5, run a single thread of Prime95. Go into the task manager and use set affinity... to lock Prime05 to one core or the other but not both. ThrottleStop should clearly show one core mostly using the 9.5 multiplier while the other core will mostly be using 9.0. That's a sign of Intel Dynamic Acceleration mode. Your default multiplier is 9.0 but when only one core is working it can use the 9.5 multiplier for a little boost in performance.

With the way your bios is, you are going to have to run ThrottleStop or RMClock to get your CPU set up correct and running at full speed. Either should work OK but RMClock does not support the half multipliers so if you want to use RMC, you will have to do a registry hack if you want to use the highest IDA (9.5) multiplier.

The first post in the under volting thread is full of information about RMC.

http://forum.notebookreview.com/har...arket-upgrades/235824-undervolting-guide.html

 

John, is RMClock still actively supported? It's a shame if it's not...always one of my favorite programs.

 

Unfortunately, there's been no development for the past two years. However, it works OK with the Core 2 Duo CPUs.

John

 

unclewebb, Thanks again, here is the screenshot

 

Your screen shot of those registers explains exactly what is going on. I have no idea why but at least we know what.

Register 0x19A is the clock modulation register. Software that makes changes to this register, including the bios, results in your computer running slower. Here's an example of what a value of 0x1C gets you.



The percentage numbers are an approximation according to Intel. When set like this, your CPU will be performing at about 75% of what Intel intended. Clock modulation should only be used in extreme situations like when your heatsink falls off. It should not be used when your CPU is sitting idle at the desktop. Dell has a few issues with randomly using clock modulation for a variety of issues and then forgetting to turn it off.

As soon as you start ThrottleStop, it immediately turns off clock modulation. As long as the first digit in 1C is a zero, there will be no clock modulation going on. The second digit tells you the approximate percentage.

E - 87.5%
C - 75.0%
A - 67.5%
8 - 50.0%
6 - 37.5%
4 - 25.0%
2 - 12.5%

That final clock modulation setting will make your computer run like a real slug.

If that register contained 18 then the 1 means that clock modulation is being used and the 8 means that your computer is running about half as fast (50.0%) as it should be.

The next register 0x199 is the multiplier and voltage (FID / VID) request register and it also tells a story. The 8 in 0x8611 shows that the CPU is using Super Low Frequency Mode (SLFM) which internally drops your bus speed in half from 266.0 MHz to 133.0 MHz. The next digit is the multiplier and it is set to 6. A 6 multiplier combined with the bus speed running at half of its maximum speed gets you the equivalent of a 3.0 multiplier and then when you add in some clock modulation to that, CPU-Z shows that as the equivalent of about a 2.5 multiplier. Now that you know what's really going on, that's not technically accurate but it does give you a clue that there is something wrong. The 0x11 in 0x8611 is the voltage VID request. 0x11 equals 17 decimal and then you plug that number into this formula.

VID Voltage = ( 17 x 0.0125 ) + 0.7125 = 0.9250 volts

Clock modulation should rarely if ever be used and it's fine for a CPU to use SLFM mode and a 6 multiplier when the CPU is idle but there is a problem when these settings are stuck like this at full load. ThrottleStop just adjusts these registers so your CPU can run at full speed. You can play around with ThrottleStop and then you can read these registers to see how things change.

It is very likely a bios issue so I guess you'll have to contact Dell to ask them what's going on.

Here's the version of RealTemp I was using that correctly reports clock modulation.

http://www.fileden.com/files/2008/3/3/1794507/RealTempBeta.zip

RM Clock or ThrottleStop can be used to turn off clock modulation (ODCM) and to get your multiplier up at its proper value. RMC does not directly support the extra 0.5 multiplier that your CPU is capable of during IDA mode but there is a way to fix that with a registry mod.

Edit: The programmer of Core Temp has been working so his program correctly reports the speed and multiplier of Core 2 mobile CPUs when SLFM is enabled.
Here are his latest beta versions in either 32 or 64 bit.

http://www.alcpu.com/CoreTemp/CoreTemp32Beta2.rar
http://www.alcpu.com/CoreTemp/CoreTemp64Beta2.rar

 

93C is hot. The fan should be kicking in around 80 and hold temps to under 90.

Might want to make sure that the fan is blowing AND make sure that hot air is getting out. I had an E6400 that was completely clogged with dust and cat hair from the previous owner. It was showing the same symptoms. It would heat up, and throttle down to protect itself. You may also want to make sure that your work surface does not block any vents. I have seen cushioned laptop pads cause issues.

90 watt should be fine. My E4310 came with a 65 watt supply.

How old is the unit?