wPrime, A Replacement to Super Pi

Due to its single threaded nature, Super Pi is losing its relevance as a measure of performance in the current era of multi-core processors. As the calculation times become faster and faster, Super PI is becoming a rather poor CPU bench test as it cannot fully stress all the CPU cores due to its single threaded nature. Super Pi scores, as it cannot utilize all of the multi-cores/processors when running the test, are rather geared towards high clock single-core CPUs.

wPrime is an excellent replacement for multi core CPUs as it is multi-threaded. It utilizes all the core of any multi-core CPU during testing. It is very similar to Super PI, but has a bit more features to retrieve most of the hardware information (CPU, Motherboard, etc.).

You can download the software from the following link;
wPrime DOWNLOAD

You can see the difference which one is taxing your multi core CPU more from the following 2 pics



HOW IT WORKS:

wPrime uses a recursive call of Newton's method for estimating functions, with f(x)=x2-k, where k is the number it is squaring, until Sgn(f(x)/f'(x)) does not equal that of the previous iteration, starting with an estimation of k/2. It then uses an iterative calling of the estimation method a set amount of times to increase the accuracy of the results. It then confirms that n(k)2=k to ensure the calculation was correct. It repeats this for all numbers from 1 to the requested maximum. The aim is to make a perfectly threaded benchmark, such that it would consistently use 100% of the CPU while in use. This is achieved by using the WMI to detect the CPU count and use exactly that many processing threads to avoid any performance losses due to multiple threads running on any single physical thread. Each thread is designed to do 1/n of the work, where n is the number of threads. For example, if you're calculating 16 roots on 4 CPU's, each CPU will calculate 4 roots. Some might argue that this style of threading is unrealistic in real-time performance, but in fact it is quite indicative of performance in several real world tasks such as F@H which allows you to run several instances of the work at any one time.

CPUz is used to detect the CPU count and use exactly that many processing threads to avoid any performance losses due to multiple threads running on any single physical thread. Each thread is designed to do 1/n of the work, where n is the number of threads. For example, if you're calculating 16 roots on 4 CPU's, each CPU will calculate 4 roots. Some might argue that this style of threading is unrealistic in real-time performance, but in fact it is quite indicative of performance in several real world tasks such as F@H which allows you to run several instances of the work at any one time.


Check the following link for multi threaded comparison of SuperPi and wPrime at MADSHRIMPS

OVERCLOCKING Q6600 WITH ASUS P5K-V

There might be a handful of people out there who will be using Asus P5K-V or VM motherboards for their rig as these are the only available mid-end boards from Asus for quite some time after P5B. Most of them already know there great overclocking potential just like there old sibling P5B. For the last couple of days, I’ve been working on this P5K-V board with a Q6600 (65nm Quad Core) with a G0 stepping to check its ocing capabilities especially for a quad core, and I must confess that I am really impressed with it.

People can think of this as a small guide for those who are using P5K-V and may be VM I assume, as both are same except the later one being a micro ATX board. I will try to explain some key BIOS settings here required for a good stable overclock. Kindly note that these are the settings that worked best for my rig but that doesn’t mean that it will work best for you. You may need to change a bit here and there depending on your other hardware but hopefully it will provide you a solid base and understanding to work around with your bios. I would also suggest to read some basic overclocking guides before coming to this.

Throughout this guide I’ll be using Q6600 2.4GHz with G0 Stepping and a 9x multiplier. Keep in mind that I used THERMALRIGHT ULTRA 120 with 95CFM VANTEC 120mm fan to cool the CPU inside CM 690 casing with 6x120mm fans installed for intake and exhaust. If you are planning to do a decent overclock use a good CPU COOLER. These boards only come with 3 Phase power design rather than 6 phase or 8 phase (usually available in high end boards), which makes the Mosfets and CPU socket surroundings quite hot during overclocking.

Following softwares are used for testing and monitoring:

OCCT for stress testing
EVEREST ULTIMATE for monitoring temps and voltages
CPUZ 1.44.1 for monitoring voltages
SPEEDFAN 4.34 for monitoring temps and voltages
CORETEMP 0.98 for monitoring core temperatures (most accurate)
CRYSTAL CPUID for cross checking core temperatures.

OK, let’s start from BIOS, which in my case is 0603. The good thing about P5K-V bios is that all of the overclocking settings, including memory timings and voltages are all on the same page, so that make things quite easier.

Open the BIOS and go to “Jumper free configuration”, and open that page.

You will see the main page with all your overclocking settings. Some of them will be hidden at first, so you need to change a few things, so that you can see all the options.

Had to split the page into 2 images due to its length.





AI Overclocking - Change to MANUAL

CPU RATIO CONTROL – I set this to MANUAL and then CPU RATIO SETTING to 9x for my Q6600. You can leave it on AUTO if you want or change it if you intend using a lower Multi.

FSB STRAP TO North Bridge – I kept it to AUTO (will try to play with it later)

FSB Frequency – Mine stock was 266, and I changed it to 378 to get a CPU Clock of 3.4GHz.

PCI-E Frequency – I always set it to 100. Its bit safer to keep it like that, though, most of the time, people don’t find any issues keeping it to AUTO.

DRAM Frequency – Initially I would suggest keeping it at the lowest divider. 1:1 would be the best option if you are at 333 MHz FSB or higher. If your FSB Frequency is set at 266 at the moment, you would only be able to see 667 setting, use this. You can change it later after finding your best CPU overclock.

DRAM Timing Control – Set this to MANUAL, and you will see all the RAM timings and sub timings. For best stability, enter the first four timings manually exactly matching your ram timings. Usually you can find these written on your ram label. I’ve always found stability issues by keeping those to AUTO, so manually entering it is always better. You can always tighter you timings later if you really want to, off-course after finding your best CPU overclock.

NOTE: Incorrect RAM timings and voltages are one of the most common reason for poor or unstable overclcock, therefore, it is best to enter the correct values manually.

DRAM Static Read Control = AUTO – (Somehow its related to FSB STRAP TO NorthBridge)

Transaction Booster = ENABLED/DISABLED –It helps increasing the overall system performance by a very small margin, if ENABLED. But I set it to DISABLED for ocing. You can keep it ENABLED if you don’t have any stability issues.

Clock Over-Charging Mode = AUTO -- (This option sets the additional voltage supplied to the board clock regulation circuitry. Useful for achieving extreme clocks)

CPU Spread Spectrum = DISABLED
PCIE Spread Spectrum = DISABLED

These Spread Spectrum features are used to reduce the EMI of a given BUS but may compromise your system stability during overclocking. It is best to set the above two DISABLED for best stability.

CPU VOLTAGE = This is where you can add/subtract extra Vcore if needed. (I set it to 1.375 after performing the VDROP and VDROOP MOD to OC my quad at 3.4GHz. You will need to set it according to the type of your processor and the amount of VDROP and VDROOP you have.

CPU Voltage Damper = ENABLED or DISABLED. This is actually meant for eliminating VDROOP and lot of people has found it useful. Didn’t work for me so I set it to DISABLED and did a VDROP and VDROOP MOD for my board (more on this later). You can set it to ENABLED and see if it helps reducing or eliminating the VDROOP, off course if you are experiencing some VDROOP. I found terrible VDROOP on this board. Before doing the MOD mine VDROP was 0.05 and VDROOP was greater than 0.12 (Initially needed to set my Vcore at 1.52 in BIOS for 3.4GHz to get an IDLE Vcore of 1.47 which was dropping to 1.35 under LOAD….resulting in higher core temps. After MOD, BIOS Vcore is set to 1.375, actual is 1.36 and under load is 1.35, temps go only 22C above room temps).

NOTE: Following shots from BIOS HARDWARE MONITOR and CPU-Z showing actual voltages at IDLE and LOAD after VDROP/VDROOP MOD


For those who don’t know about VDROP and VDROOP, here is a short explanation.

If you set a Vcore value in BIOS, there is some difference between the voltage shown in the BIOS, and what you actually get (VDROP). And there is also some difference between your idle Vcore and Vcore under load (VDROOP). Amount of VDROOP plays an important role in system stability under load and CPU temperatures.

Click the following link to learn about applying a VDROP/VDROOP PENCIL MOD

VDROP/VDROOP PENCIL MOD FOR ASUS P5K-V

CPU Voltage Reference = AUTO --(Read CPU GTL Reference below)

CPU PLL Voltage = AUTO–(Use to overcome the FSB Wall issue. I won’t suggest playing with it unless you are an extreme tweaker. May damage your hardware)

DRAM Voltage = I would suggest to set this manually. You can find this info on your ram label. I have found AUTO settings quite unstable on most of the systems.

NOTE: Ram is potentially the easiest component when it comes to killing. Overvolting may damage your ram quickly. Kindly consult your ram's label or manufacturer website for correct volatge.

FSB Termination Voltage = Leave at AUTO for now.

North Bridge Voltage = Leave this at AUTO if you are marginally ocing your system, say only couple of 100MHz. Setting this to 1.40 would be more than enough for 1000MHz OC provided that you don’t use lower multipliers and tighter RAM timings. You may need to increase this for very high FSBs and tighter RAM timings AND (may be) if you are using 4 sticks of RAM.

North Bridge Voltage Reference = AUTO--(Read NB GTL Reference below)

UPDATED
BIOS 1001

Ok, Bios updated to 1001. Found the following new settings.

LOAD-LINE Calibration = AUTO/ENABLED/DISABLED—(Previously CPU VOLTAGE Damper. Meant for eliminating VDROOP. I set it to DISABLED)

CPU GTL VOLATGE Reference = AUTO--(Previously CPU VOLTAGE Reference)

NB GTL VOLATGE Reference = AUTO--(Previously NB VOLTAGE Reference)

WHAT IS GTL VOLTAGE REFERENCE?

Gunning Transceiver Logic- Tweaking the GTL Reference voltages for the CPU and Northbridge can sometimes give them better stability. This is particularly noteworthy with (multi core) quad-core processors because, if you’re finding two cores drop off under load, it’s probably down to the fact that the two CPU dies are not identical and while core one and four can hit the FSB you’ve set, the other pair of cores are having trouble. Increasing this level will help bring stability to cpu's at higher FSB's

Memory and CPU suffers from clock miss alignment at higher frequencies. From what I understand, the GTv needs to be adjusted in order to cope with the additional electrical noise created by increased FSBs and electrical current through the motherboard. The GTLv has the ability to 'filter' out noise in the electrical signaling to the CPU and the memory that would cause instability.


Now lets come to the ADVANCED CPU SETTINGS page.


DISABLE everything here except EXECUTE DISABLE BIT (which is used to prevent BUFFER-OVERFLOW). You can DISABLE it if you like.

If you are not doing any bench testing i suggest you should keep the CPU TM Function ENABLED. It affects CPU protection management and help you when you don’t realize you’re pushing your chip too hard. This will keep you safe from damaging your processor.

This concludes our main settings for overclocking on P5K-V. I’ll update this blog as required. I hope it will be useful for some. Kindly leave a comment if you think there is some mistake or you want me to add further related to this.

Thanks!

DISCLAIMER

Overclocking may VOID your components WARRANTY. Overvolting any component carries a high risk of damage and/or failure. I take no responsibility for any loss or damage to your components as a result of using this guide. All overclocking should be done at your own risk.

ASUS SS-PRO vs ASUS SK II vs AC FREEZER 7 PRO SHOOT OUT

Until few weeks back I used to wonder about the level of performance of locally available Asus CPU coolers. I read couple of reviews on the net about those coolers but none was able to provide me with the performance details I was looking for. You must be wondering the reasons for that? Well here are few I would like to highlight.

No 1; Surprisingly, some reviews won’t even bother to mention the ambient temperatures at the time of testing, one of the most important factor to evaluate any cooler’s thermal performance as it uses air to dissipate heat.

No 2; Sometimes no mention what kind of test bed was used, like whether the testing was performed outside the case or inside the case? Usually temperatures in an open case test don’t go as high as in a close case, the later being the operating environment for majority of us.

No 3; Well, actually i always wanted to test these against my all time favorite middle weight champion AC FREEZER 7 PRO.

Fortunately, these days of wonder soon got over when I received the following two Asus coolers from Shingtech for review.

ASUS SILENT SQUARE PRO
ASUS SILENT KNIGHT II

As the scope of this MATCH UP is more towards the performance of these coolers rather than there packaging and construction details, I will still try to give readers a short brief about packaging and construction stuff.

Asus is a well known name and is one of the largest motherboard makers in the world. Besides producing many other computer peripherals, Asus also produce CPU and case coolers. Today we will be looking at two top of the line CPU coolers from this company and see actually where they fall in when it comes to cooling performance.


ASUS SILENT SQAURE PRO

Asus Silent Square Pro belongs to Asus top of the line Silent Square Series. Silent Square Pro package is quite eye catching due to its orange color scheme that is pretty much consistent with the cooler’s cladding color. The box has some nice shots of the cooler and some basic product details.



The design concept is quite similar to Tuniq Tower 120 in a sense that the 90mm fan is sandwiched between the fin arrays on both sides. I found it quite weird that Asus has used very uncommon fan size of 90mm instead of 92mm. This will leave user with nothing but to stick with the provided fan. SS PRO consists of 5 copper heat pipes with aluminum fins and copper base. Though I am still in doubt about the claim of copper base that rather look likes aluminum to me…hmm well guess I am not a metallurgist. The fan spins at a rate of 2500RPM and is capable of producing quite decent 49.4CFM with respect to its size. The accessory kit includes mount adapters for Intel 775 with two different backplates for different motherboard configurations. Mounts for AMD 754/939/940/AM2, and Intel 478 are also included. ASUS Silent Square Pro includes a very useful 3.5" bay fan controller with a blue display for real time RPM monitoring of the CPU fan. This makes it very easy to control fan speed and noise of the Silent Square Pro. Small syringe of premium ASUS thermal paste and a clear set of color installation manual are also included.

The size of SS PRO is not a mediocre by any means. Actually it’s quite big and will occupy quite a real estate once inside a case. It weighs 745gm which is by no means can be considered light or normal. SS PRO doesn’t come with the best and easiest fastening mechanism, the standard push-pin style, like the one we find on AC FREEZER 7 PRO which I used for reference. Back plate installation means one need to remove the motherboard in order to install the cooler, a procedure not liked by many but necessary for heavier mounts like these. The mounting method appears to be very secure and manage to distribute the weight quite evenly. On motherboards with heat pipe setup surrounding the CPU socket like Striker Extreme or CPU socket near the top edge of the board, pushing the holding clamp is quite difficult as there just isn't enough space between the socket and the heat pipes or socket and power supply with the CPU near the edge. In that case you may need to unscrew the board just to push and lock the retention clip. Have a look at the SS PRO details from Asus website:

ASUS SILENT KNIGHT II

Silent knight II comes in a nice box with a cutout hole in the middle displaying a side view of the heatsink itself. Images along with basic design features and specifications are printed on its sides and the back.

Silent Knight II is constructed of all copper from its base, to the heatpipes, as well as the fins, and arrives at a reasonable 610g weight. The cooler is based on six copper heatpipes originating from a copper base plate which then pass through 128 copper fin arrays. It comes with a 92mm PWM BLUE LED sleeve bearing fan that can rotate up to 2300 RPM at max. The fan draws air into the heatsink through the middle where it’s located, and then exhausts it out the back. The blue fan combined with its heatsink design creates airflow onto components around the CPU socket to more effectively cool them down. The overall construction quality of the cooler is pretty excellent with clean welded metal joints and good finished base, though there is still room for improvement in base finishing. The heatsink fins are pretty much elevated above the mainboard PCB, thus reducing the chance of any component conflict.

The accessory kit includes mount adapters for Intel 775 with two different backplates for different motherboard configurations. Mounts for AMD 754/939/940/AM2, and Intel 478 are also included. Fastening and installation mechanism of Silent Knight II is same as its sibling, the Silent Square Pro.

Here is what Asus has to say about their SK II;

AC FREEZER 7 PRO

Founded in 2001, Arctic Cooling is a well known manufacturer of cooling solutions for personal computer systems focusing on CPU, GPU and PC-Case cooling. AC FREEZER 7 PRO is designed to be used with intel LGA775 socket CPUs. There is an AMD version available under the name AC FREEZER 64 PRO.

The package is small and mostly monochromatic and the only other things included in the box are an instruction sheet and a small, branded sticker.

The Freezer 7 Pro is built around three U-shaped heatpipes (for 6 heatpipe paths) and a stack of 42 aluminum fins. The heatpipes receive heat from the CPU at the bottom of the U and transfer it up to the ends where it is dissipated into the aluminum fins and then eventually into the surrounding air. The fins are large and tightly spaced, so the total surface area is high. The 92mm fan is rated for providing 45 CFM at 2500RPM. AC FREEZER PRO weighs very modest 520gm, exceeding Intel and AMD recommendations by only 70 grams. AC FREEZER 7 PRO last three fins have been bended downwards to provide a bit of airflow to the VRMs (Voltage Regulators Module).

This image has been resized. Click this bar to view the full image. The original image is sized 840x697.

The mounting system is the same as used by Intel's stock heatsink, so mounting the Freezer 7 Pro is just a breeze. You just need to remove the fan before installing the FREEZER and then putting it back again after installation which is not at all difficult. AC FREEZER 7 PRO comes with pre-applied premium quality MX-1 or MX-2 thermal paste so you don’t need to worry about how much you should apply. Installation will take less than 5 minutes.

TESTBED AND TESTING MATHADOLOGY

PROCESSOR: DUAL CORE E2140 1.6GHz & OC @ 2.66GHz with 1.38 Vcore
MOTHERBOARD: ASUS P5K-VM
RAM: CORSAIR XMS2 2x1GB DDR2 PC5300-667MHz
GPU: Nvidia 7900GS XXX 256MB
PSU: CORSAIR EXTREME 460WATT
CPU COOLERS:

• ASUS SILENT KNIGHT II
• ASUS SILENT SQAURE
• ARCTIC FREEZER 7 PRO

OS: WIN XP PRO (SP2)

All the tests were performed inside standard HT casing with 1x80mm intake and 1x80mm exhaust fan besides PSU’s 120mm fan also pulling air out of the casing. Speed fan 4.33 was used to measure the CPU temperatures through on-board sensors along with Coretemp 0.98 to measure the core temperatures. In order to establish how well each of these heatsinks can perform (especially in our local environment with long and hot summer accompanied by hours of load shedding), i set up a worst-case scenario by switching the A/C unit off and letting the room temperature to get to max which was around 33°C +- 1.

Reason for doing this was plain simple. As all the air coolers use air to dissipate heat, the temperature of your room plays very important role in determining the max CPU load temps. The performance of any cooling system can be expressed as a "thermal resistance", in degrees Centigrade per watt (°C/W). A thermal resistance of 1°C/W means that a heatsink is getting one degree warmer for every watt of heat it's dealing with. Lower the thermal resistance, better the heat dissipation. Not all manufacturers specify this in their specs but some do like ARCTIC COOLING.

FOR EXAMPLE: AC FREEZER 7 PRO’s claimed thermal resistance is 0.17 °C/W which means for every watt of heat, its temperature will rise 0.17°C OR 17 °C/100W above air. In other words, if the CPU is producing 100 watts of heat which is thus passing through the heatsink, the heatsink with a thermal resistance of 0.17 °C/W will dissipate 100-(0.17x100) = 83 watts while retaining only 17 watts of heat, which will be the CPU temperature above air. So if the room temp is around 30°C and your case has a good airflow the 100 watt producing CPU temperature will be 30 +17= 47°C. There are some other finer points that need to be considered such as the conductivity of air and the voltage the CPU is set to run at. If you know a Cooler's thermal performance in °C/W, you can predict your processor’s average load temps in any given environment by doing a simple math.

Idle temperatures were recorded after letting the system in idle state for 15 mins. System was loaded using OCCT for 30 mins to record the load temperatures. ARCTIC SILVER 5 was used as a thermal interface material with all the CPU Coolers to ensure the accuracy of the results.

First I measured the temperatures using stock settings then overclocked the processor to 2.66GHz with Vcore at 1.36v to increase the processor maximum output (TDP) from 65 watt to around 120 watt.

OVERCLOCKED

The Delta T shows the difference between ambient and the processor CASE and CORE temperatures. Measuring Delta is a better way of judging the cooler’s thermal performance regardless of your room temperature as it gives you a better insight regarding your processor temperature under any given conditions. It helps you in measuring your heat sink’s thermal performance.

FINAL WORDS

It can be easily concluded from the above results that all coolers are more-o-less on par in performance with each other. The performance of both Asus coolers is quite impressive that leave us with some room for decent mid end overclocking on any Core 2 Duos. If you are in the market for a good performing CPU cooler with killer looks then look no further then SILENT KNIGHT II, besides it doesn’t occupy much space inside your case. This will definitely give it an edge over SILENT SQAURE PRO in smaller casings.

SILENT SQAURE PRO would have been my choice had ASUS used some standard 92mm fan instead of uncommon 90mm. Though the huge size of SILENT SQAURE PRO doesn’t seem to do justice with its cooling performance, I am sure inserting any high CFM fan between the fins will definitely improve the cooling performance, provided you can get hold of one.

Once again AC FREEZER 7 PRO has come on top on my list due its excellent cooling performance, reasonable size, ease of installation and premium quality pre-applied thermal paste. Unfortunately, it's not available for sale in our local market, but fortunately we can have it's cooling performance in the shape of ASUS Silent Knight II and Silent Square Pro that comes with some great looks as well.

I hope you will find this match-up helpful for your next CPU cooler purchase