TechREACTION » Guides

[Blog] FFF&FM Rev2.0 – water cooling built for benching!

miahallen — Sun, 29 May 2011 01:34:49 +0000

The original recipe

A couple years ago, I put together a simple water cooling system that I dubbed FFF&FM.

Form
Follows
Function
and
Fully
Modular

And the name really tells you everything you need to know. It was not a pretty system, nor did it need to be….but it did work very well for cooling my HW while pretesting for more serious benching session with liquid nitrogen. Here is a link to the original forum post where I detailed the build.

FFF&FM Rev1.0 parts

The original system used two Swiftech triple 120 radiators sandwiching three 120×38mm San Ace fans. It used a single DDC style pump, and performance was excellent, especially in the winter when I could hang it out the window for some extra cool benching sessions.

FFF&FM Rev1.0

While the original system has served me well, I recently decided that I needed to correct a couple flaws with the design, and also push the performance to the limit….so with that in mind, I bring you the next generation….

Revision 2

The idea with revision 2 is to keep the original design goals, but enhance performance, and correct two flaws in the original design. What were those flaws?

First off, there was no reliable indication of flow, this is pretty important especially when disabling the safety mechanisms built into modern HW like we do with benching. When you tell the CPU not to throttle if it’s overheating, then if something goes wrong with the cooling system you’re liable to kill some HW. This almost occurred this past winter when I had a pump fail on the original FFF&FM and my CPU temps skyrocketed. Luckily I was running a Bloomfield CPU which has proved to be extremely durable.
The second flaw in the original design was the lack of pump redundancy, as I just described pump failure was a major concern in the new revision. So with the new design I knew I wanted to support two or more pumps in case one fails.
The last issue with the first desing was the use of thin walled tubing. I used 3/8″ ID and 1/2″ OD tubing which means the wall thickness was only 1/16″ thick. With the new system, I’ve upgraded the tubing to 3/8″ ID and 5/8″ OD which means the wall thickness is now twice the original, or 1/8″ thick. This new Primochill tubing is MUCH more kink resistant, and this is super important for a system thats continuously being reconfigured and played with.

All of the other design goals remain, and the most important ones are performance, and the “modular” design made possible with the Koolance quick disconnects. These have made this water cooling system as useful as it is.

The build

So, I wanted to take you through a quick step-by-step build log, showing how I put the system together. The biggest part of the new system is the new Watercool radiator, which is a monstrosity designed to use 18 140mm cooling fans, nine on each side. This is the “light” version, which only allows for mounting fans on one side.

Watercool MO-RA3 front

In the picture above, the radiator is resting on the box for a Swiftech MCR-320-QP quad 120 radiator….that should give you some scale, this thing is huge! There are actually 6 fill ports, two on the front, two on the rear, and two on the top which provides tons of flexibility for your desired configuration.

Watercool MO-RA3 rear

Based on this testing I wanted to put some shrouds on the fans, and I also wanted to use some good 120mm fans, due to the lack of decent 140mm fan options. So, I decided to use a 30mm thick piece of wood as a shroud for the entire bank of fans.

pum & res position

I started by getting the board cut to the correct size at the hardware store. Then I drilled holes to mount the radiator, and positioned the pumps and reservoir to ensure the board would accommodate the desired layout. The reservoir selected for this project is the Tecnofront Challenger X1 from Italy. It has an integrated flow meter, which will be a nice visual indicator for me. The two DDC based pumps (10W OEM versions) are linked with the EK Dual DDC Turbo Top.

fan hole placement

Next, I drew out the spacing for all the fans. My goal was to have each fan placed directly over the intended location for the standard 140mm fan, to get as even a distribution of air as possible.

cutting fan holes 1

Then I slowly began cutting the fan holes, going 1/2 way through on the first side, and then finishing the cut from the opposite side.

cutting fan holes 2

After cutting all the holes, I then routed the rear side (the side facing the radiator) to help airflow spread out from the fan to cover the gaps in-between the fans.

test fit

Then, I bolted the board up to the radiator for the first time to ensure a proper fit and alignment.

res & pump placement

With the radiator mounted, I was able to trace the edges of the reservoir and pump wires to plan for proper placement of each.

res & pump cuts

Next, I made some rough cuts for the reservoir and the pump wires, and I also routed the corners of the fans intake side.

weatherstrip

I took my Dremel sanding wheel to the corners and then some good ol’fashioned elbow grease and got them nice and smooth. Now the holes are proper square with rounded corners to match the fan’s frame. Then I placed weather strip around the edges to ensure a good seal on the radiator.

final fitting

Next, I mounted the radiator for the final time; the first tube was also connected with a compression fitting to the front side of the rad.

tube routing 1

You can see here the reservoir and pumps placed in their final locations. Both are secured to the top of the radiator with double sided sticky tape.

tube routing 2

You can see the basic tube routing here, the water comes in through the lower white tube on the left (into the front radiator port, not seen), exits the radiator into the black tube on the right and goes into the reservoir, the out of the reservoir into the pumps, and out of the pumps in the upper white tube and out to the system.

tube routing 3

Another angle.

tube routing 4

And another.

fans mounted

The next step was to mount my fans, nine San Ace 120×38mm 110+CFM gems. Been really happy with these fans over the last few years and they should be perfect for this project as well. You can also see the reservoir mounted at the top.

base 1

With everything mounted up, I also wanted to create a stable base for the system. I used the leftover wood from the original board.

base 2

I used some additional scrap to make some strengthening plates to put on each side.

base 3

Another shot from the front.

fan seals

No matter how hard you try, if you don’t have the proper tools to complete your cuts perfectly, you’re never going to have perfect alignment. That’s where silicon sealant comes in.

wired and running

The last step was to wire the fans and pumps together and fill and blead the system. I have to say this Tecnofront reservoir made that job easier than with any other reservoir I’ve ever used…99% of the air bubbles were bled within a couple minutes….very impressive!

system complete

After getting everything hooked up and running, I played with a few benchmarks. But first I wanted to give you a quick sound check, and made the video below. Also notice the flow meter showing you a good indication of flow.

Results

Well, this is not a review, but I still wanted to leave you with some results. I have not spent much time with the system yet. But I did have a chance to test out a new 990X and improve my previous best 5970 Vantage score.

3D Vantage with 5970

The CPU cores reached a max temp of 72C and the GPU cores hit 47C and 43C….not bad considering the 20C ambient temp in my room at the time.

I’ll have tons more testing with this system, and I still have a few more minor improvemtnes to make (fan grills and weather protection)…but the meat and potatoes are all here, I hope you liked it and it inspires some of you.

3 Step Overclocking Guide – Sandy Bridge

miahallen — Wed, 05 Jan 2011 03:26:08 +0000

Introduction

Many users are searching around the net these days looking for advice on how to overclock their new systems, but aren’t sure where to start. To help everyone out, I decided a how-to guide was in order. Searching around forums can be confusing and intimidating. There are so many people willing to give advice, but who can you trust? It’s hard to know, and I have seen many users sent on wild goose chases because they follow advice that doesn’t solve or even address their specific problem. I have also seen too much trial and error overclocking, and unless you get lucky, it tends to be far too time consuming and frustrating. What I have attempted to do is create a very simple three step guide for overclocking Sandy Bridge based CPUs. If you want to continue searching out other opinions, please consider each suggestion with caution. Some will undoubtedly be great, while some will not.

***Note*** We are still planning to create a BIOS terminology cross-reference chart (like in the other guides), but I need your help! Thanks, and happy overclocking! ***

Disclaimer

I am not responsible for any bad things that happen to you or your computer as a result of you following this guide, nor is TechREACTION.net. My goal is for this guide to be a safe overclocking guideline, however the burden for damaged hardware always lies on the user performing the overclock! Overclocking can damage hardware and in most cases will void your warranties.

Prerequisites

In the prior version of this guide, I requested that you have some basic knowledge of your motherboards BIOS. While I have not addressed every motherboard on the market, I have included details for the top enthusiast brands. But as before, please do not be afraid to get into your BIOS and have a look around, if you are ever concerned that you may have changed a setting erroneously, you can always load defaults, and start over. Most boards have a CMOS reset button on them now-a-days, if not check your user manual for the location of the CMOS reset jumper…please ensure you know the location before getting started.

This guide is independent of your cooling system. Whether you are using the stock Intel cooler or if you’re pushing to the extreme with phase change cooling, the basic steps remain the same. One thing that is far too common are mistakes mounting your cooling system, specifically the application of the thermal interface material (TIM). If you don’t have much experience mounting cooling apparatus, please refer to this excellent guide from Arctic Silver.

Methodology

Determining methods for finding a stable overclock are highly controversial, everyone has their own definition of a stable system, but when I refer to “stable” in this guide, I am referring to the stability of your selected “stability test.” So for a power user or gamer who wants a reliable system that won’t ever crash due to an overclock pushed too far, you’d need to test with a program that will load all of the cores and threads applicable to your CPU, OCCT and IntelBurnTest are two popular choices. OCCT uses the same algorithm as Prime95 but has a friendlier interface. IntelBurnTest uses the Intel linpack binaries to stress the system and also has an easy to use interface. In this guide I may use testing that is insufficient in your opinion. It is only a guideline and if you feel more testing is necessary for your system, by all means feel free.

So with that in mind, we will attempt to isolate each portion of the system and overclock one step at a time. This may seem time consuming at first glance, but rest assured this will potentially save you hours of troubleshooting and frustration. So go slow, and follow each step very carefully.

BIOS familiarization

If you’ve found my guide online, my guess is you’re looking for more than a basic overclock. If you’re not, and all you’d like is something simple, please redirect your attention to your motherboard manufacturer’s website and download the latest overclocking utility. For basic 10-20% overclocking, they work pretty well. There is “Gigabyte EasyTune6“, “Asus TurboV EVO“, “MSI Control Center“, and “eVGA eleet“. This guide is written to take it to the next level, for THAT we need to do the overclocking from the BIOS.

Speaking of which, before we begin, please check your motherboard manufacturer’s website for the latest version of your BIOS. Usually enthusiast level boards will have BIOS engineers tweaking them for months or years to improve overclocking support. Unless you have a reason to stay with your current BIOS, I’d suggest updating to the newest version.

If you don’t know how to access the BIOS, please refer to your motherboard’s owner’s manual for instruction. While you’re there, find out how to “clear CMOS”. As I mentioned in the introduction to this guide, it’s important you know how to properly “clear CMOS” before we begin.

Secondly, the first thing to do after powering up the new system is to enter the BIOS and find the “hardware monitor” area and verify the CPU temperature is reasonable based on your cooling. If not, please power down the system and verify the mounting of your cooling apparatus (refer to the guide linked in the “prerequisites” section.

Goals

The variety of users reading this guide is vast, and each user’s goal will be unique and specific to his/her needs. It would be impossible for me to address every user’s specific needs. But I’ve attempted to be as broad, yet specific as possible. My goal is to assist the maximum number of users as possible, despite your specific needs.

Based on user feedback from the previous version of this guide, I decided to better address overclocking with power features enabled. The easy answer was to follow the old guide and then attempt to enable your power features afterward, but that rarely worked when approaching the limits of a given system.

Just as before, if you want to maximize your overclock, you should disable all the power saving features in the BIOS as detailed in step 1. However, if you’re after a more moderate overclock, and you’d like to save power (especially while your system sits idle) you can leave those setting enabled. Just follow the guide as written (I’ve added tips for you along the way) to find your best settings. While your potential overclock will be more limited, the benefit will likely be worth it to many of you.

Terminology

I’d like to start off by writing briefly about the BIOS and more specifically, differences in terminology between the different manufacturers. Obviously there are too many motherboards on the market to show you every single one in this guide. But looking at boards from the four top manufactures, we should be able to better identify specific terminologies used by each.

Understanding “total system performance”

Before we go into how we overclock these CPU’s let us look at what determines how fast your whole system will run. CPU frequency is very important. However, there are many other factors that play into your total system performance. All of your primary BIOS overclocking revolves around the Base Clock or “bclock” and clock ratios. The base clock’s default speed for all H67/P67/Z68 based systems is 100MHz. overclocking methodology changes a lot with this new platform. We are no longer focused on bclock changes because Intel has buried the clock generator into the CPU itself, and has locked it down so only minor changes are possible.

Sandy Bridge die map

CPU frequency = bclock x CPU clock ratio

This is a biggest change from the last generation (H55/H57/P55/X58) is that most overclocking will come from changes in the CPU multiplier, not the bclock. But the formula doesn’t change; the bclock multiplied by the CPU clock ratio will still determine your overall CPU frequency.

Core i3 & i5 2100 series = fully locked
- These CPUs are “fully locked”, the multiplier cannot be increase beyond the factory setting.
Core i5 & i7 non-K series = partially unlocked
- These CPUs are “partially unlocked”, the multiplier can be adjusted up to +4 the factory setting.
Core i5 2500K & i7 2600K = fully unlocked
- These CPUs multipliers can be freely adjusted; in effect, they are “fully unlocked”.

H67/P67/Z68….what’s the difference?

There will be three “chipsets” supporting LGA1155 Sandy Bridge, from an overclockers perspective, here is the difference.

H67 will not allow CPU multiplier overclocking, only base clock, which as I’ve already mentioned is very limited. If you intend to overclock your CPU, please do not choose H67. Memory overclocking is still possible with, and the GPU integrated in Sandy Bridge is enabled with H67. Please don’t bother buying a “K” series CPU with H67, it will be a waste of your money unless you really need a stronger iGPU.
P67 will be the enthusiast choice for Sandy Bridge, it will allow CPU and memory multiplier adjustments for overclocking. P67 does not support the integrated GPU, if you plan to use the integrated GPU, please do not buy a P67 based motherboard.
Z68 Is now available and supports fully unlocked overclocking for the K series and is iGPU enabled. You will be able to adjust the CPU, memory, and GPU multipliers with Z68, a great choice if you plan to overclock and use the GPU. Also keep in mind, this would be a good choice for someone who’d like to run multiple monitors, but dedicate their discrete graphics card to the primary monitor.

Memory frequency = bclock x System Memory Multiplier (SPD)

Depending on the motherboard, memory overclocking is fairly straightforward with Sandy Bridge. Bclock multiplied by memory ratio will determine the memory frequency.

DDR – The other part that can be quite confusing for users who are not familiar with DDR technology is the difference between the memory clock speed and the memory’s DDR speed. For instance, DDR3-1600 actually runs at 800MHz, it’s just that DDR (or dual data rate) technology allows the memory to process twice per clock cycle. Back when we switched technologies from SDRAM to DDR for the first time, the manufacturers started saying DDR-400 when it ran at 200MHz because it was better marketing to sell their memory over the older SDRAM technology. This is why CPU-Z shows 800MHz for your DDR3-1600, or 1000MHz for your DDR3-2000.

Memory speed and bandwidth can have a huge effect in some applications, and negligible impact on others. But overall, top shelf memory is one of the worse items you can spend your money on from a value perspective. Faster CPUs and GPUs will give you much more performance for your hard earned cash.

Important Voltages when Overclocking

There a few important voltages which you will need to manipulate while overclocking; below are the primary ones. Not every motherboard BIOS is identical, but all enthusiast level motherboards should provide control of the voltages as shown below.

CPU Vcore – Directly related to the CPU frequency. As you increase the CPU frequency you would need incrementally increase the v-core as well. Sandy Bridge is very new, and a “safe” voltage range for long term reliability is not yet known. As we spend more time and learn more about this platform, I will update this guide with a more educated estimate. For now, I’d suggest staying below 1.45V or 80C load temperatures. I feel those are both fairly conservative settings. Another big change from previous architectures is that the L3 is now tied directly to the CPU for both power and clock speed. So, from an overclocking perspective, L3 is now part of the core, where it was previously part of the uncore.

For now, my recommendation is that while you are stress testing, monitor your CPU core temperatures with Real Temp and if the temperature is under 80C, you can increase the voltage up to 1.45V max. If you don’t mind the risk, feel free to push further and make sure to publish your results for the community to learn from, whether positive or negative.

Nothing I’ve ever used my computer for come close to generating heat like LinX. Because it generates so much heat, it has become my favorite stress testing application. As long as I can keep my CPU cores below 80C while running LinX, then for me that’s safe. If you are more conservative/cautious than me that’s perfectly OK.
*** NOTE *** Please ensure you’re running Windows 7 with SP1 for LinX testing, it enables AVX which increases temps a fair amount.

Load Line Calibration - This actually goes by a few different names, but they are all meant as a means to reduce or prevent v-droop. It does typically ease the overclocking process at the cost of violating Intel’s design specs. However, overclocking in its essence violates Intel’s design specs, so you’re not breaking any new ground with this feature. This feature was very useful for increasing the overclocking potential of the last few generations of Intel CPUs, and it’s already being effectively used with Sandy Bridge as well. For more insight on the theory of LLC, refer to this excellent explanation at anandtech.com. There was also some real world testing recently; feel free to check out Bobnova’s LLC investigation here.

VccSA – This is the voltage with controls the “System Agent” (new “uncore”). Since the L3 cache has been moved to the core, the only thing left on the System Agent that concerns us as overclockers is the integrated memory controller (IMC). It’s already been discovered that the IMC on Sandy Bridge is quite robust, and usually won’t need any additional voltage for speeds up to DDR3-2000 and possibly even higher. I found I needed about 1.15V for maximum potential when running very fast memory speeds. This may also be important with very high density DIMMs or when fully populating the DIMM slots on your motherboard. At this time, I would caution using any more than 1.2V on the VccSA.

VccIO – This is the voltage which controls the SA’s IO. Many users and manufacturers are taking issue with my claims of the SA voltage being most important for IMC overclocking. While I cannot explain my personal results, they definitely go against the majority. With that being said, the motherboard manufacturers and many uses will tell you to only adjust the VccIO, and leave the VccSA alone…I’m recommending you try both, and see which works better for your CPU. My testing was difinitively VccSA reliant. At this time, I would caution using any more than 1.2V on the VccIO.

DRAM voltage – This is directly related to your RAM modules and increases will allow increase in MEM speeds. There has been a lot of debate as to the limitation 1.65V limitation Intel has published. With the older platforms, the rules no longer apply. With a few months past now, it seems safe to say that this platform is robust enought to handle running memory at higher voltages, at least for the short term. Many overclockers still want to push the limits, but since all the current memory seems to scale less with voltage than older stuff, this is becoming a moot point. I’d still suggest staying at 1.65V or below for a regular daily system, but I’ve pushed up to 1.85V for short benching sessions without any adverse effects.

Sample overclocking goals to use as a reference

I decided not to use “sample systems” as in the last few guides. Because of the vastly different methodology used in this overclocking process, I don’t believe they’ll be helpful this time.

Next: 3 Step Overclocking Guide Continued…

[BLOG] TechREACTION CPU cooler testing methodology

miahallen — Mon, 15 Nov 2010 02:00:30 +0000

In an effort to better serve the enthusiast community. TechREACTION is working on defining a long term methodology for testing various products. Our desire is to be consistent, accurate, and transparent with out methodologies.

We have spent a couple of months deriving out first complete testing methodology, and I’m sharing it here for “peer review” as-it-were. From this point forward, all CPU cooling devices formally tested by TechREACTION will adhere to this methodology. This will allow us to compare various heat sinks against one another with reasonable accuracy, regardless which staff member is performing the testing. We know it is not perfect, but we hope that it will improve out reviews from this point forward, and allow us to begin compiling the test results of various cooling devices for quick reference over the long term. Please have a look, and see my commentary after the outline. Then feel free to join a discussion on the topic in our forums.

We’ll be using the following hardware configuration
1. X58 motherboard
2. Bloomfield based CPU
3. 3×2GB RAM
4. low power graphics card
5. Open test bench
6. high quality PSU
Mount the heatsink three separate times (if you find poor contact during dismount, throw out the results and re-run) with Arctic Cooling MX-3 as shown in the images below (with slightly less TIM).
1. For each mount take a before and after photo of TIM application and spread (as shown below).
2. If the heatsink you are testing comes with fan(s), then mount it a fouth time with the stock TIM and stock fan, and run through testing (as described below) one time.
For each mount, use the following configuration
1. Windows 7 x64
2. Prime 95 for 30 minutes for each test
3. stock speed testing, run once with Scythe Gental Typhoon 1850RPM fan and once with a 109R1212H1011 San Ace fan (or similar)

i. bclock – 133MHz

ii. CPU clock ratio – x21

iii. Vcore = 1.15V – LLC enabled

iv. VTT = 1.2V

v. uncore ratio – x21

vi. DDR3-1066 7-7-7-20 1T

vii. Intel SpeedStep (EIST) disabled

viii. Intel C-States (C1E etc…) – disabled

overclocked testing, run once with Scythe Gental Typhoon 1850RPM fan and once with a 109R1212H1011 San Ace fan (or similar)

i. bclock – 200MHz

ii. CPU clock ratio – x20

iii. Vcore = 1.4V – LLC enabled

iv. VTT = 1.35V

v. uncore ratio – x16

vi. DDR3-1600 8-8-8-24 1T

vii. Intel SpeedStep (EIST) disabled

viii. Intel C-States (C1E etc…) – disabled

RealTemp will be used for core temperature monitoring/logging
A high precision digital thermometer will be used to capture ambient temperatures at approx 3″ from the center of the intake fan. These temps should be logged accurately at 1 minute intervals (minimum) throughout testing as described below.
After each mount, follow these steps for testing four times….once for each configuration listed above.
1. Start Windows and allow the system to idle at the desktop for 5 minutes.
2. Start the Prime95 application, and three instances of CPU-Z for CPU/mobo/MEM, then manually start RealTemp (with logging enabled) application, and allow Realtemp to log idle temps for five minutes.
3. When the Realtemp clock reaches 5 minutes, start the Prime95 “blend” torture test and allow to run for 30minutes.
4. When the RealTemp clock reaches 35 minutes, stop the Prime95 torture test.
5. When the RealTemp clock reaches 45 minutes, close the RealTemp application.
While each 45 minute set is running, the you should be monitoring ambient temps as described in step 4 (above)…so at this point you should have a minimum of 46 data points for the ambient temp graph.
Save the RealTemp logand the ambient temp log into your data folder named appropriately for the test….I use the following naming scheme – mount_clock_fan….so 1st mount, stock speed, Gentle Typhoon fan would be “M1_stock_GT.csv”
Copy and save this spreadsheet to a new google doc and name it for your HSF. Now follow these two easy steps and end you’ll up with a nice result…
1. Paste the values from each RealTemp log into columns A-H
2. Paste 45 temp readings from each ambient temp log into column K.

First of all, we decided the X58 was the platform of choice for enthusiasts these days. While we do support AMD, the lack of reliable temperature monitoring hardware and software for AMD platforms meant it was not as logical a choice as Intel. Combine that with the performance advantage Intel holds over AMD, and we decided the X58 would be the best platform for cooler testing. For CPU, Bloomfield was chosen for many reasons, primarily due to the fact that many of you should be able to replicate our results with reasonable accuracy in your own testing.

We talked about many possibilities for # of mounts, OC settings, and test applications. We settled on two settings, stock speed at OC’ed to 4GHz with 1.4V. We decided the 4GHz at 1.4V represents the typical worst case scenario for a Bloomfield overclocker. So we’ll test with those two profiles on Bloomfield (D0 stepping only) CPUs.

We settled on testing by averaging the results of three mounts, and throwing out the results from any mounts where proper contact is not achieved (based on TIM spread). These results will give us a good idea of the average results obtainable with decent mounting. The TIM selected for all testing is Arctic Cooling MX-3 for its availability, good performance, and ease of use. There are many other good choices available, but the MX-3 provides a good basis for our testing.

And lastly, we wanted to standardize airflow testing with specific popular fan models. Specifically, the 1850RPM Scythe Gental Typhoon series which have recently become one of the most sought after fans for their performance/noise ratio, and the long standing San Ace 109R1212H1011 fan….or another 120×38mm San Ace fan with the same specs (102CFM), of which there are a few. If the fan has a fan integrated in its design, or will not accept the 120mm fans we have selected, the cooler will only be tested with the integrated version, or the fan that ships with the unit (at the testers discretion). In addition, all coolers will also be tested with their factory supplied TIM and fan (single mount only) to compare how well the boxed solution performs.

So, at a minimum, four mounts will be performed, one with the stock TIM, and three with MX-3. For the stock TIM mount, the factory provided fan will be tested (if supplied) at both speed settings. For the other three mounts, we’ll use Arctic Cooling MX-3 and will be run through testing four times, twice for each fan at each of the speed settings.

Prime 95 was selected to generate a load for the CPU, as we feel it more accurately represents real world use. Temperature monitoring will always be performed with the most recent version of RealTemp available at the start of testing. Concerning temperature monitoring, the other very important measurement is ambient temperature, which we will poll a minimum of once per minute with a probe mounted at the air intake of the cooler being tested.

Our results really focus on the “rise over ambient” measurement….that is, the amount of core temperature increase above the ambient temperature in the testing location. This is a very important measurement and is much more useful for a wide range of readers in various environments around the globe.

Lastly I’ll mention the google spreadsheet we’ll be using to compile the data we generate during testing….it’s still pretty rough, and will likely be improved and streamlined as time goes on. Please feel free to use it in your own testing (according to the instructions listed above) and if you blog at TechREACTION, use it with pride.

We are very open to your comments, questions, and criticisms….please post in our forums.

AMD Phenom II IMC and RAM Overclocking Guide

Slappa — Fri, 15 Oct 2010 00:07:12 +0000

Introduction

Over the recent months I have watched and observed as the Phenom II CPU found its niche within various enthusiast communities and online forums. Users have been provided with an excellent, all around, 45nm AMD quadcore.

However, there is one factor that people still find confusing and unclear. Overclocking the Phenom II. Why are people so confused? One reason may be because they add a lot of new variables and overclock differently than the competitors chip. Some may choose to see Phenom II overclocking as a headache, or a fresh challenge in their hands. I choose to see it as a fun challenge.

Now, one area that people have mass amounts of problems is Ram/Memory stability and overclocking, as well as what role the IMC (NB) plays in affecting ram and performance.

The IMC

The IMC links the CPU to the memory in a system. In the case of the Phenom II AM3 series, DDR2 or DDR3 memory is supported. One important performance factor that people neglect when applying daily overclocks to their Phenom II systems is that the faster the IMC speed is, the better your memory will perform. Additionally, you have to take into account that overclocking the IMC may require voltage increases on the CPU/NB. This may add unwanted heat which happens to be the Phenom II’s arch-enemy. While overclocking though, always keep in mind that the IMC is a very touchy factor in an overclock. If it’s not working out for you, you perhaps need to try different combinations/ratios. When stress-testing the IMC, the most preferable method is Prime 95 blend test for at least 2-3 hours.

The Ram

The memory within an AM2+ or AM3 configuration acts and requires different tuning than your average Intel setup. The RAM ties strongly in with the IMC. If the ram is unstable it can cause instabilities in the IMC and vice versa. Ram on this platform tends to like lower memory frequencies and timings. When tuning ram, take into account every setting you can find in your BIOS. I mean it. Experiment. It may seem like a lot to take in, but hey, grab a coffee, a pad of paper, and sit down for 2 hours and do some testing. The second important portion of RAM clocking is that you must know what type of IC’s your RAM uses. This already will put you in a great position to fine tune the ram. Plenty of online resources can help you identify your IC’s if you do not already know. Once again, after clocking your ram, 100% stability can be found using Prime 95 linked above. Stability testing for benchmarking overclocks can be done in SuperPi 32M.

The CPU Cores

Since the focus of this article is geared towards Ram and IMC clocking, I will be quick on my CPU Cores description.

Phenom II CPUs are usually very intolerant to high voltages when using air setups. Be aware that more voltage may decrease stability going past 1.5V on quad core versions of Phenom II. Also remember that these cores love cold temperatures and scale brilliantly with it. This can be seen in my Canadian Winter blog series, as I will soon start to compare the scaling from normal air temperatures to very cold air temperatures. Even without winter air, every degree that you can lower your temperatures in an air setup is worth it towards overclocking.

Pulling It Together

This creates a 3 component chain wherein lies the secret to stability and performance. You have to realize that if any component in this chain (CPU, NB, Ram) happens to be unstable, the chain will not be complete and you will never see stability. In order to maintain stability, you have to find a sweet spot for each variable according to your setup. There are a plethora of ways you can approach this. However, I discovered that some specific methods are more efficient than others.

The Overclocking – Guide

Starting With The RAM

It is unarguable that even out of the box, with bare bone stock settings in a BIOS, memory will cause the most headaches for the AM3 setup. For example, prior to Phenom II X6 chips which include a brand new IMC revision, you could not take a standard 1800MHz 9-9-9-24 set of DDR3 and get it to run at it’s rated speeds without doing some tweaking to settings in the BIOS . Even sticks that can in fact run at rated speeds out of the box on an AM3 platform can be a pita for stability. This is in contrast to the CPU and IMC where on stock settings, they are supposed to operate flawlessly without any rifts.

1) To begin your overclock start with a default ram clock and timings. For example: 1333MHz 6-8-6-24. This can also depend on the stock specs of your ram to begin with. That is always a good starting point (excluding stock 1800MHz and 2000MHz sticks, as it is unlikely you will get them to run properly at first starting at such a high frequency).

2) You can start speeding up your ram in a variety of ways. By increasing clock speed (Through dividers and HT Ref Clock), tightening timings (adjusting latency timings, and sub-timings in the BIOS), or combining both for the ultimate overclock. Start in small increments on the frequency and timings. Remember that the 4 major ram timings (CAS, tRCD,tRP, tRAS) are very dependent on what IC’s your ram has, and changing these more then one value at a time can cause instant instabilities requiring a CMOS reset. Also remember that it is contradictory to loosen all timings significantly to increase frequency and vice versa. What increase in speed will there be if you are just trading off one thing for the other? However fiddling around with some sub-timings and loosening can help stabilize higher frequency overclocks, while minimally affecting performance. Generally, the most important part of ram clocking is getting familiar with your ram, and what settings greatly affect performance and stability. This is a lengthy process and there is a lot of rebooting involved. General ranges to shoot for would be: 1333MHz (CL 5 or 6), 1600MHz (CL 6 or 7), 1800MHz (CL 6-8), and 2000MHz (CL 7+). What range you want to fall into also depends on your ram, the quality of the IC’s and your ability to tune the timings. Remember to always experiment as much as you can and play around with everything available!

Here are a few sources that list different DDR3 IC’s:

http://www.jmax-hardware.com/bdds/ddr3.html

http://ramlist.i4memory.com/

http://www.hardwareluxx.de/community/f13/hardwareluxx-ram-datenbank-ddr3-404293.html#post7147707

3) With each small change you make it is important to know if you are stable or not. In windows you can use Super Pi 32m as a preliminary stability test. This will not gauge 100% stability, but it will help you figure out what range you will be able to overclock within. Passing 32m also means that you are stable for most benchmarks. If you cannot pass 32m, then your ram is not remotely stable, and you can then adjust your settings accordingly. For full stability testing for 24/7 overclocks, utilize Prime 95 for at least 3 hours. During any stability test, it is important to eliminate the risk of cpu cores and IMC (NB Frequency) causing any stability problems. This is to ensure that if there are instabilities, they are occurring within your ram, not within your processor cores or memory controller. So make sure that the NB and CPU are as close as possible to stock settings prior to testing.

4) If you do find instabilities in your overclock there are a few things you can do to help stabilize them. First things first. Do not go crazy with the ram voltage. In moderation though, it can help. Recent DDR3 modules cannot take too much voltage on standard cooling. It is common for a lot of ram to come stock at 1.65v. Usually this ram should not exceed 1.75v for risk of degradation or damage to the ram. However some IC’s are known to like voltage more then others. Micron D9’s come stock at 1.8v usually and love voltage. Now that that’s out of the way, here are a few tips for stabilization. Increasing the CPU-NB voltage slightly is always helpful, regardless of whether the NB Frequency is at stock. NB Voltage is a setting in the BIOS that refers to the physical northbridge on the motherboard. Do not confuse this with the IMC. In some cases, increasing this voltage in moderation can help stabilize high ram overclocks. Another tip is to ensure that both of your ram sticks are in slots 3&4. On some Gigabyte 890FX boards these slots offer higher headroom for ram speeds. Keep the ram cool, pointing an extra fan toward it can always help. The final tip is to adjust drive strengths, which I will talk about in depth in the next step..

5) In the BIOS, there is an advanced set of ram settings. These are called drive strengths. Essentially you can use these to stabilize the ram further. It is very hard to understand which changes in drive strength values are useful unless you utilize memtest. You need to boot from this program off a CD or USB before going into windows. When changing the value of a drive strength always remember to do it for both sticks of ram (sometimes labeled DCT0&DCT1, or A&B). Only changing one value at a time, run a round of tests in memtest. You want to change drive strengths until memtest gives you as little errors as possible on the test suite. Once done so, you have essentially helped stabilize the ram or have created more headroom for overclocking. Remember that you cannot do this if you are getting preliminary tests of hundreds or thousands of errors in memtest. You need to begin with a handful of errors, and this is what you must use as your baseline value that you will start tuning from.

Adding In The Northbridge Overclock (IMC)

Once you have a proven ram overclock that is stable, it is time to work on the IMC. The IMC can vastly increase performance of the memory subsystem in all aspects. It allows a memory overclock to work to its full potential. Overclocking the memory controller is very similar to overclocking the CPU cores. It is much easier to overclock and understand then ram overclocking, as there are less variables in play.

1) When overclocking the IMC you need to know what will affect it’s headroom. On the Phenom II X4, X3 and X2 chips, the NB Frequency has a max stable range of 2600-3000MHz depending on your chip. For 24/7 overclocks this frequency usually falls around 2700MHz. This will require voltage increases in the CPU-NB option in the BIOS. However, Phenom II X6 chips are slightly different. Their max 24/7 range is around 2800MHz-3000MHz. They also require less voltage then previous chips.

2) Start by increasing the NB multiplier till you hit around 2300MHz-2500MHz. Try stability testing this in Prime 95 combined with your ram overclock. Do not adjust the voltage to the IMC just yet. If this passes, then move up a notch. Go until the IMC causes Prime 95 to crash, and then start increasing voltage in moderate increments on the CPU-NB option. When overclocking take into account that the HT Ref Clock (which also controls your ram and cpu frequency) will become the base that is multiplied by the NB multiplier to get your NB frequency. This means that each multi increase will have a bigger increment as opposed to if the HT Ref Clock was left at 200.

3) For CPU-NB voltages, do not exceed 1.5v. After that the IMC won’t really scale well. Also note that each time you add voltage to the CPU-NB, you are adding to the overall heat output of the chip. So this is something to watch under air cooling, as it could hinder your CPU core overclock that you will later add in. The rule of keeping the chip cool also applies to the NB. A cooler chip could equal a higher overclock headroom.

4) When stability testing your final NB overclock, run Prime 95 for at least 3 hours on the blend setting. This is to ensure that the ram and NB are both stressed to the maximum during the 512K FFT iterations. You may wish to run this longer then 3 hours if you like, but that is more stress then any real-world application can offer. After this is stable, then you have yourself a final overclock of the IMC and Ram.

The Final Step

For the final step, you would add in your CPU core overclock. A great guide to check out for overclocking the CPU is my previous guide here. All this information should help in completing your final stable, 24/7 overclock. Here’s an example (it was only 2 hours because of time constraints):

I hope you guys enjoyed my second guide, stay tuned for more in the future.

———–

To see more of my articles, check out my list of techreaction blog posts, click here.

To see all of my AMD overclocking and experiences, check out my personal blog, click here.

3 Step Overclocking Guide – Lynnfield

miahallen — Tue, 07 Sep 2010 04:00:42 +0000

Introduction

So many users are searching around the net these days looking for advice on how to overclock their new systems but don’t know where to start. To help everyone out, I decided a how-to guide was in order. Searching around forums can be confusing and intimidating. There are so many people willing to give advice, but who can you trust? It’s hard to know, and I’ve seen many users sent on wild goose chases because they are following advice that doesn’t solve or even address their specific problem. I’ve also seen too much trial and error overclocking, unless you get lucky it tends to be far too time consuming a frustrating. What I’ve attempted to do is create a very simple three step guide for overclocking Lynnfield based CPUs. If you want to continue searching out other opinions, please consider each suggestion with caution. Some will undoubtedly be great, some will not.

Disclaimer

I am not responsible for any bad things that happen to you or your computer as a result of you following this guide, nor is techreaction.net. My goal is for this guide to be a safe overclocking guideline, but the burden for damaged hardware lies on the user performing the overclock! Overclocking can damage hardware and in most cases will void your warranties.

Prerequisites

Methodology

Determining methods for finding a stable overclock are highly controversial, everyone has their own definition of a stable system, but when I refer to “stable” in this guide, I am referring to the stability of your selected “stability test.” So for a power user or gamer who wants a reliable system that won’t ever crash due to an overclock pushed too far, you’d need to test with a program that will load all of the cores and threads applicable to your CPU, OCCT and IntelBurnTest are two popular choices. OCCT uses the same algorithm as Prime95 but has a more friendly interface. IntelBurnTest uses the Intel linpak binaries to stress the system and also has an easy to use interface. In this guide I may use testing that is insufficient in your opinion. It is only a guidline and if you feel more testing is necessary for your system, by all means feel free.

BIOS familiarization

[Sticky] Intel Core i3, i5, and i7 – 3 Step Overclocking Guides

miahallen — Tue, 07 Sep 2010 04:00:40 +0000

Introduction

So many users are searching around the net these days looking for advice about which components to purchase and how to overclock. This guide is not meant as a comparison between AMD, and Intel. This guide assumes you’ve already decided to purchase an Intel based system, and takes you to the next step…..deciding which one. The emphasis here is on value, overclocking, and overall system performance. If you want to continue searching out other opinions, please consider each suggestion with caution. Some will undoubtedly be great, some will not. This isn’t a comparison test, but a simple overview of the different choices from Intel and their inherent strengths and weaknesses.

Overclocking guides

Once you’ve made your decision and purchased your system, the overclocking guides below will help you gather a better understanding of how your system works, and will give you a very methodical approach to overclocking it. If you have not yet made your purchase decision, continue to read below to assist you with your decision. After you have selected your hardware, please come back to visit the overclocking guide pertinent to you.

3 Step Overclocking Guide – Sandy Bridge

3 Step Overclocking Guide – Bloomfield & Gulftown

3 Step Overclocking Guide – Lynnfield

3 Step Overclocking Guide – Clarkdale

Micro-Architecture

When the Nehalem Micro-architecture was introduced to the world in the fall of 2008, everyone took notice. The Intel Core series of CPUs that started with Conroe in 2006 was very strong, but I’m not sure anybody anticipated that Intel would be able to be successful with two micro-architecture generations in a row. So, when Nehalem first hit the scene, it was a shock and awe campaign. Not only did Intel have the performance to stay well ahead of the competition, they actually improved on the previous generation so greatly, that a large quantity of users already on the stellar Core 2 Duo/Quad platform were compelled to upgrade without skipping a generation. Since AMD has not had anything competitive from an outright performance standpoint, Intel has cornered the high end market for quite a while. What follows is a brief look at the various platforms that Intel has introduced since Nehalem’s release.

First we have the Nehalem family which consists of 45nm quad core CPU’s

4-core Bloomfield (Core i7 920 to i7 975) – socket 1366
4-core Lynnfield (Core i5 750 to i7 875K) – socket 1156

Next is the Westmere family which consists of 32nm, dual core, quad core (Xeon only), and hex core CPU’s

6-core Gulftown (Core i7 970 and i7 980X) – socket 1366
2-core Clarkdale (Pentium G6000, Core i3 530 to i5 655K) with integrated GPU’s – socket 1156 only

There are many other versions of CPUs designed for the enterprise (Xeon) and the mobile markets. I will not include any details about them within this guide; some of the Xeons are virtually identical to the desktop counterparts and may be worth considering. For more details, please see this wiki about Xeon CPUs. The mobile equivalents won’t even fit into the desktop sockets, so I won’t spend any time on them.

Clarkdale

All Clarkdale CPUs feature two pieces of silicon; the first chip contains the dual CPU cores and L3 cache, it is manufactured on a 32nm node. The second chip is the graphics core, memory controller, and PCIe controller and is manufactured on a 45nm node. The pictures below should help you understand a little better.

Clarkdale package

Clarkdale die map

The new 32nm lithography is very exciting and powerful, and overclocks very well, unfortunately with Clarkdale, the 32nm CPU is tethered to the older technology which creates some unique challenges when overclocking. The memory controller is dual channel and officially supports up to DDR3-1066. Memory and graphics performance are both fairly limited with Clarkdale. If you plan to run multiple graphics cards, I’d recommend an X58/LGA1366 based system.

Pentium G6950

Currently, there is only one Pentium CPU based on Clarkdale, the G9650. This is a very good budget buy, and has very good overclocking potential, but lacks Intel® Hyper-Threading Technology and Intel® Turbo Boost Technology. It is also limited to only 3MB of L2 cache.

Core i3 530, 540, 550

These CPUs are similar to the Pentium G6950 but add Intel® Hyper-Threading Technology. They also come with 4MB of L2 cache.

Core i5 650, 660, 661, 670, and 655K

These CPU’s are very similar to the i3’s but add Intel® Turbo Boost Technology. The i5 655K has an unlocked multiplier for maximum freedom in overclocking. The i5 661 differs from all other Clarkdale based CPU’s because its integrated graphics core runs at 900MHz instead of the standard 733MHz in the rest.

Lynnfield

These CPUs are manufactured on the 45nm process. They have dual channel IMCs and PCIe controllers built into the CPU die. This is the closest Intel has come so far to integrating an entire system on a single piece of silicon for the desktop market (that will all change next year with Sandy Bridge, but that discussion is for another time).

Lynnfield die map

Each core has its own 256KB of L2 cache, and all four cores share 8MB of L3 cache. The memory controller is dual channel and officially supports up to DDR3-1333. Lynnfield CPU’s are known for their extremely high memory speeds and for very good value in the quad core marketplace. Similar to Clarkdale, I would recommend to anyone interested in multiple graphics cards, that you consider an X58/LGA1366 based system for additional PCIe bandwidth, just like Clarkdale, these CPU’s have only 16 PCIe lanes built in. All Lynnfield CPUs include Intel® Turbo Boost Technology.

Core i5 750 and 760

The only things that separate the i5 Lynnfields from their i7 siblings is their exclusion of Intel® Hyper-Threading Technology, Intel® Virtualization Technology for Directed I/O (Intel® VT-d) and Intel® Trusted Execution Technology (Intel® TXT).

Core i7 860, 870, 880, and 875K

So, these are just like their i5 siblings, except that they include all three technologies missing from the others. Intel® Hyper-Threading Technology, Intel® Virtualization Technology for Directed I/O (Intel® VT-d) and Intel® Trusted Execution Technology (Intel® TXT).

3 Step Overclocking Guide – Bloomfield and Gulftown

miahallen — Tue, 07 Sep 2010 04:00:32 +0000

Introduction

So many users are searching around the net these days looking for advice on how to overclock their new systems but don’t know where to start. To help everyone out, I decided a how-to guide was in order. Searching around forums can be confusing and intimidating. There are so many people willing to give advice, but who can you trust? It’s hard to know, and I’ve seen many users sent on wild goose chases because they are following advice that doesn’t solve or even address their specific problem. I’ve also seen too much trial and error overclocking, unless you get lucky it tends to be far too time consuming a frustrating. What I’ve attempted to do is create a very simple three step guide for overclocking Bloomfield and Gulftown based CPUs. If you want to continue searching out other opinions, please consider each suggestion with caution. Some will undoubtedly be great, some will not.

Disclaimer

Prerequisites

In an earlier version of this guide, I requested that you have some basic knowledge of your motherboards BIOS. While I have not addressed every motherboard on the market, I have included details for the top enthusiast brands. But as before, please do not be afraid to get into your BIOS and have a look around, if you are ever concerned that you may have changed a setting erroneously, you can always load defaults, and start over. Most boards have a CMOS reset button on them now-a-days, if not check your user manual for the location of the CMOS reset jumper…please ensure you know the location before getting started.

Methodology

Determining methods for finding a stable overclock are highly controversial, everyone has their own definition of a stable system, but when I refer to “stable” in this guide, I am referring to the stability of your selected “stability test.” So for a power user or gamer who wants a reliable system that won’t ever crash due to an overclock pushed too far, you’d need to test with a program that will load all of the cores and threads applicable to your CPU, OCCT and IntelBurnTest are two popular choices. OCCT uses the same algorithm as Prime95 but has a more friendly interface. IntelBurnTest uses the Intel linpak binaries to stress the system and also has an easy to use interface. In this guide I may use testing this is insufficient in your opinion. It is only a guidline and if you feel more testing is necessary for your system, by all means feel free.

BIOS familiarization

3 Step Overclocking Guide – Clarkdale

miahallen — Tue, 07 Sep 2010 04:00:09 +0000

Introduction

So many users are searching around the net these days looking for advice on how to overclock their new systems but don’t know where to start. To help everyone out, I decided a how-to guide was in order. Searching around forums can be confusing and intimidating. There are so many people willing to give advice, but who can you trust? It’s hard to know, and I’ve seen many users sent on wild goose chases because they are following advice that doesn’t solve or even address their specific problem. I’ve also seen too much trial and error overclocking, unless you get lucky it tends to be far too time consuming a frustrating. What I’ve attempted to do is create a very simple three step guide for overclocking Clarkdale based CPUs. If you want to continue searching out other opinions, please consider each suggestion with caution. Some will undoubtedly be great, some will not.

Disclaimer

Prerequisites

Methodology

Determining methods for finding a stable overclock are highly controversial, everyone has their own definition of a stable system, but when I refer to “stable” in this guide, I am referring to the stability of your selected “stability test.” So for a power user or gamer who wants a reliable system that won’t ever crash due to an overclock pushed too far, you’d need to test with a program that will load all of the cores and threads applicable to your CPU, OCCT and IntelBurnTest are two popular choices. OCCT uses the same algorithm as Prime95 but has a more friendly interface. IntelBurnTest uses the Intel linpak binaries to stress the system and also has an easy to use interface. In this guide I may use testing that is insufficient in your opinion. It is only a guidline and if you feel more testing is necessary for your system, by all means feel free.

BIOS familiarization

[Blog] Gaming in Tight Spaces…..mini-ITX (featuring Gigabyte, Silverstone, and Prolimatech)

miahallen — Fri, 13 Aug 2010 20:26:28 +0000

History of mini-ITX

I was sitting in my office, browsing tech news online almost a decade ago when I first saw the mini-ITX form factor. My mind immediately started fantasizing about the possibilities such a small system would afford. Thoughts of internet terminals built into kitchen cupboards, or car PCs entertained my imagination for years. As the early VIA based systems got into reviewers’ hands and performance results started hitting the press, like many others, I was left wanting. A PC that struggles to play flash based videos, or 720p video content doesn’t get me very excited.

The first system that really caught my attention hit the scene in 2006, the Albatron KI51PV-754. It was based on socket 754 for AMD Athlon 64 processors. But, since the socket didn’t support dual core CPUs like the 939 socket did, you’d be stuck with a single core CPU, and a single DIMM of memory, it also lacked a PCIe slot and the price was unreal! So, it was encouraging to see progress, but obviously not prime time.

Albatron kept at it and in the summer of 2007 released a system based around the new AM2 socket, the KI690-AM2, which would accept Athlon dual core CPUs. But like it’s predecessor, it also came with a few flaws, it was stuck using SO-DIMMS for memory duty, was still lacking a PCIe port for GPU duties, and the cost was still too high.

Then in 2008, nVidia debuted their new “ION” system, which it was said would answer everyone’s payers concerning performance per watt. Combining the low power consumption of the modest Intel Atom CPU, and a small integrated GPU by nVidia, the ION would have the combination of CPU and GPU power to be taken seriously, at least once everyone was on board. It’s taken some time to get everyone on board, but this ION system is proving to be very versatile these days. There is one area where it still does not make the grade…….gaming.

Then, another small company hit the scene in 2008, they had been selling nVidia based graphics cards for a couple years, and decided to try their hand in the motherboard market. In late 2008, Zotac released a rash of AMD and Intel based mini-ITX motherboards that were clearly aimed at enthusiasts. The boards were based on all of the latest chip-sets including ION, nVidia 6100 series for Intel, nVidia 8200 series for AMD based systems. And then in early 2009, the nVidia GeForce 9300 series for Intel based system.

Obviously there have been many other mini-ITX products that have been made. I don’t mean to play down any of the other products, but from my perspective, these were the milestone products.

mini-ITX and me

The ZOTAC GF9300-G-E was my first mini-ITX motherboard. It had all the features one could ask for at the time, socket 775 for a powerful Core 2 Duo CPU, dual channel DDR2 support for regular sized DIMMS, a PCIe x16 slot for dedicated graphics, and even onboard WI-FI! I brought the board home, and went straight to work rebuilding my HTPC around it with an E8400 Core 2 Duo at its heart. It didn’t take long for me to figure out that the system still had a fatal weakness……power consumption. I had built the system in an Antec ISK-300-65 case. But that “65″ at the end of the model number is significant, it represents the wattage supplied by the case’s PSU. 65W proved insufficient to power the system, even when underclocked, and undervolted. So, when it was working I was very happy with the performance, but it really wasn’t very well suited for a really small system.

Zotac H55ITX

Then at CES this year, Zotac announced plans to release a new mini-ITX board based around the LGA1156 socket for Intel Core i3/i5/i7 CPUs. I was drooling on my keyboard when images first hit the net. Was this going to be the answer to our questions, on paper it was just about perfect. Quad core support? Check. Dual channel DDR3 with regular sized DIMMs? Check. PCIe slot? Check. Onboard wi-fi N? Check! So, as soon as they showed up….I had two on their way to me.

This is not a review of the Zotac board though…the H55ITX from Zotac served me for about 4 months in my HTPC, and while it never impressed me, it worked in the Antec HTPC case I had, and 65W seemed plenty. The system was stable, and my HTPC was working well for the first time in years. I finally got around to putting the second board into a SFF gaming PC about 6 weeks ago. This machine I built with the intention over overclocking….whoops! Things went sour shortly after first power up. I’ll go into really brief detail.

Actually, the very first time I booted the first board, I thought it quite odd that there was a message at the bottom of the post screen stating “FOR EVALUATION PURPOSES ONLY”. I thought it was just a fluke. But after powering on the second board for the first time (a few months later), I was greeted with the exact same message. Once into Windows, I opened CPU-Z and on the motherboard tab, all of the vendor specific fields show “to be filled in by O.E.M.”…humm. A quick check on their support page shows the very first BIOS update removes the post screen message, and that’s it. I think it’s a bit sloppy that Zotac would ship out their boards to distributors with a pre-release BIOS, but apparently, that’s exactly what happened. So, I updated the BIOS of both boards with the latest version at the time, version 330. And the message at the post screen did in fact go away. However, the fields were still not completed in CPU-Z.

So, I’m more concerned with performance, and I was willing to forgive all else, if high performance was present…but it wasn’t. My original goal was to build this gaming rig with a Clarkdale based dual core CPU overclocked to at least 4.5GHz. I had two Clarkdale CPUs which I had purchased for my extreme overclocking endeavors. My i3 540 ended up in the HTPC system, the other CPU I had on hand was my very strong i5 670; yes, the same one that I took to 6.8GHz a few months ago. Seemed like a good candidate, so I set right off to try to overclock it to my target…but I fell well short on my goal. There is no way to increase VTT/QPI voltage in the BIOS, so I was limited to around 154MHz bclock…with a maximum CPU multiplier of x27, that put my CPU clock at only 4150MHz. Not bad, but I knew I could do better…but without the option to increase the VTT voltage, I was stuck.

The second major problem came a short time later. I had ordered some low end GeIL DDR3-1333 DIMMs, and when they arrived, I popped them in. For the next two weeks I struggled to get the system stable, and I was not successful. Although the memory would pass memtest with flying colors…the system remained completely unstable with regular BSOD’s in Windows. It took me quite a while to determine it was a compatibility issue with the memory, because it would pass memtest as I mentioned. After switching to some extra OCZ Gold DIMMs I had, the instability and BSOD’s went away.

Around this time, Intel released their new Core i5 655K CPU with and unlocked CPU multiplier. I thought this would be the perfect solution for my Bclock issues. If I didn’t have to raise the bclock, I should be able to reach my goal by simply raising the multiplier and CPU voltage as necessary….as long as the CPU was up for those speeds. But after purchasing the new CPU and installing it into the board, I quickly realized my problems were not over. Right off the bat, the board was feeding my brand new CPU 1.4V with all stock settings loaded. And undervolting back to normal levels resulted in instability. In addition, the “unlocked” multiplier didn’t work! I’d seen many people with this issue, so I went through methodically, manipulating every setting in the BIOS trying to figure out how to get the extra multipliers to work.

After a couple days of searching support forums and asking for help, I decided to e-mail tech support. After about a week, I heard back from the USA branch with contact info for the BIOS techs in Hong Kong. So I sent my request for help to them…and waited…and waited…and about a week later I got an e-mail for one of the techs claiming they would try to provide a beta BIOS to fix the issues. And about a week later he sent me the beta BIOS…IT DIDN’T CHANGE ANYTHING! I wrote back and to report my findings, and over the next few days the BIOS engineer and I exchanged several e-mails in an attempt to correct the problems. The culmination of these efforts ended with zero changes to the problems, and the engineer claiming that it worked fine for him. The only theory I have is that he was using an engineering sample, and my CPU was a retail sample.

It has now been over six months since the board has been available, and Zotac has had 4 BIOS releases, non of which provide vendor information in the motherboard tab of CPU-Z. The last release was in April, almost four months ago. It seems crazy to me that they wouldn’t have fixed these simple issues yet. But, I digress……

Gigabyte H55N-USB3

About the same time I started building the gaming box, Gigabyte announced plans to produce their first ever mini-ITX motherboard. At first I wasn’t interested, because the onboard wi-fi on the Zotac was a slightly important feature for me. However, after working with the Zotac tech support and failing to resolve my issues…I decided it was time to move on to a motherboard manufacturer I could trust, enter the Gigabyte H55N-USB3. As the very first product in the mini-ITX market from Gigabyte, you’d expect a few shortcomings, here and there. You might expect an immature BIOS, limited overclocking, or buggy drivers and/or features. However, I’m very happy to report that you’d be dead wrong!

In case you’ve been living under a rock, I happen to be very familiar with Gigabyte motherboards, I’ve played with quite a few over the last year or two. It would take me a few minutes to distinguish between the BIOS on this board and that of the P55A-UD7. This board is built just like its bigger siblings…it’s amazing! We’ve come to expect lower end motherboards to have a lower end BIOS. But that’s just not the case anymore with Gigabyte! And sure, you lose the onboard wi-fi to the Zotac board, but you gain SATA3 and USB3.0, which is great for future expansion and compatibility.

Gigabyte H55N-USB3 vs Zotac H55ITX

Here you can see the two boards side by side. Very similar in layout and design. But you’ll notice the Zotac’s onboard wi-fi module, and six (instead of only four) SATA ports. Also, on the back panel, Zotac provides ten (wow) USB2.0 ports to Gigabyte’s four USB2.0 & two USB3.0 ports. So, feature wise, I’m still a fan of the Zotac board. But the real game changer is overclocking performance.

H55N-USB3 with box and contents	H55N-USB3 I/O panel
H55N-USB3	H55N-USB3
H55N-USB3	H55N-USB3

Building a mini-ITX based SFF gaming computer

So, what components did I use? How exactly did I put this system together? How high was I able to overclock it? How well does it perform? How much power does it consume?

These are the questions I suppose you would like me to answer right? Well, read on…

What components did I use?

Case – Silverstone SG05 The case may be the single more vital aspect of a SFF build. The who point of a SFF build is to be small….and the case you choose will determine the precise amount of “small” that you get. To my knowledge, the Silverstone SG05 is the smallest case on the market which will support a 9.5 inch long, dual slot graphics card. Actually, I’m still a bit shocked by just how small this thing is….my main PC is in a Coolermaster Cosmos S case which is almost 6100 cubic inches…..the SG05 is only 658, barely more that 1/10th the size!!! It has one other very important feature for a gaming system, a powerful PSU….300W and 80+ certified made by FSP.

Motherboard - Gigabyte GA-H55N-USB3 As seen above, a very well endowed mini-ITX motherboard.

CPU – Intel Core i5 670 I tested both this and the i5 655K extensively, and this CPU proved to be a vastly more capable overclocker. A better choice for most users would probably be an i3 540 which have been shown to be very capable overclockers for a much more sensible price. Another worthy candidate for a build like this would be a Lynnfield based quad core i5 or i7 CPU, although it will put a greater strain on the PWM on this motherboard, and a greater draw on the PSU.

CPU cooling – Intel Core stock HSF, and the Prolimatech Samuel 17. There are a couple other good choices on the market. But I chose the Prolimatech due to my highly positive experience with their tower cooler, the Megahalems.

Memory – 2×2GB OCZ Gold DDR3-1600 8-8-8 I removed the factory heatspreaders because I think they are ugly. The DIMMS will have plenty of cooling from the CPU fan. These DIMMS represent a good value, although there are many other great choices on the market. I would advise against anything high end if you choose a Clarkdale based dual core CPU, and the integrated memory controller (IMC) on these CPUs is not strong enough to properly drive most high speed memory. If you choose a Lynnfield based quad core CPU, the high speed memory may be worth your while

Video card – VisionTek Radeon HD 5850 1GB This is a reference design 5850, so brand is not important here…..all boards based on the reference design are only separated by their BIOS. This GPU represents a very good value, it has near top end performance for a reasonable price. It has a very high performance-per-watt factor, and it fits! The other very sensible choice from the green team would be the new nVidia based GTX 460 GPU. However, I wanted maximum performance, and the 460 doesn’t quite match up to the 5850.

Storage – Intel X25-M 80GB SSD and 1TB Western Digital Caviar Green I had the SSD sitting around after upgrading to a G2 drive in my main office PC, it’s being used for Windows 7 x64 only. The 1TB drive is used to store all of my games and media files. I’m not using an optical drive in this build, as all of my games are digitally downloaded from Steam. If you are on a tighter budget, you could forgo the SSD and still have stellar performance.

Monitor – Dell UltraSharp 2007FP All testing (except 3DMark tests) will be performed at the native 1600×1200 resolution.

I’d like to talk about price for just a moment. Systems “designed for gaming” do not always live up to their hype, especially for one on a budget. The reason for this is the graphics card. There are a ton of online PC vendors around, and many of them will build you a “gaming system”. They’ll use a high end Core i7 CPU, triple channel memory, and a bunch of other cool stuff that sounds great on paper, but then they put a low end graphics card into the mix (to save on costs), and it just ruins the whole thing from the perspective of gaming.

The system I build here would cost about $1300 (not including the monitor), although following my advice an purchasing a Core i3 540, and forgoing the SSD could cut the cost down to well under $1000. Yet you’d be hard pressed to find a machine as capable for gaming from one of the previously mentioned online vendors. Please heed this advice, if you want to build a gaming system, start with the graphics card, and go from there.

How exactly did I put this system together?

First, I stripped the case down to the basics to ensure all the components had clearance, to ensure all the PSU connectors were long enough, and to form a basic game plan of how to organize everything to go in it properly. This is not an exact science, and it took many attempts to get everything situated and happy. You’ll also notice I modified the PCIe power connector by soldering on a splitter at the end, giving me the dual connectors required by the 5850.

222 mm (W) x 176 mm (H) x 276 mm (D)	Will fit a 9.5" graphics card without modification
Only room for an SFX PSU	300W FSP built PSU
The PSU has 1 80mm cooling fan	cables are a bit long for such a small case

Next, I stripped the stock heaspreaders off the OCZ DIMMs….they’re so much sexier when naked

2x2GB OCZ Gold DDR3-1600 8-8-8-24

Then, I unpacked and prepped the Prolimatech Samuel 17 heatsink. This thing is a very impressive looking cooler to me….totally means business, considering its size.

Prolimatech Samuel 17 box	mounting brackets included for all modern systems
it will accept a 120mm cooling fan (not included)	the mounting surface is slightly convex

Then, I assembled the motherboard, CPU, CPU cooler, and memory outside the case. Here you can see one of the delta fans I’m using for cooling, very powerful and fairly loud.

mobo, CPU, memory, and HSF installed	such a pretty little package
RAM is hard to change with the cooler installed	the fan is a 120x25mm 3400RPM Delta which will provide 113CFM of air flow

Then, I did a quick test fit of the graphics card to give you an idea of the clearance.

very slim clearance for the graphics card

the meat and potatoes

Next, I mounted the motherboard assembly into the case.

motherboard assembly installed

This next part was the trickiest. I now had to wedge the graphics card into place. It looks as if it’d fit with room to spare. However, with the reference 5850 design, the PCIe 6-pin power connectors are on the end of the graphics card, and it was impossible for me to plug them in after mounting the card. So I had to get the card into the case with the power hooked up. But if the PSU was installed first, there wasn’t enough room to get the card in. So I had to have the power connected from the PSU to the card, and both outside the case. The other thing in the way was the fan mounted on top of the Samuel 17 cooler, it had to be removed prior to getting the graphics card in. Then I reinstalled the fan, and finally the PSU. I also reversed the PSU; normally, the fan faces down and draws air out of the case. But as one of my major concerns is the PSU, I turned it around to allow it to draw cool air in through the vents in the top of the case.

added the graphics card and PSU

slim clearance for PCIe connections on the end of a long graphics card

After getting the graphics card and PSU installed, I installed the second Delta as a front intake fan, and the front panel.

the front fan mounted crowds the area even more

front panel installed

The last step was to install the storage bays. Obviously, I’m not overly concerned with appearances (ie wire management) with this build. For me, it’s all about the performance!

SSD and HDD installed	system complete - rear
system complete - front	system complete - left

So, there you have it…small enough to be easily portable.

How high was I able to overclock it?

Overclocking can be intimidating to first timers, but as long as you take it step by step, and approach it methodically, it’s not nearly as difficult as it may seem. I wrote a 3 Step Guide for overclocking your i3/i5/i7 CPU, be sure to give it a read if your attempting an OC for the first time.

I started by testing the overclocking potential of my CPU with my custom water cooling loop. I found that this i5 670 CPU is a very strong chip. 4.5GHz was much easier than expected, and 4.6GHz also fell with ease. After a few hours of playing and tweaking, I was able to get a maximum frequency of 4848MHz stable for 20 passes of LinX.

i5 670 LinX stable at 4.85GHz on water pretesting

Obviously, I wasn’t going to be able to run that high with the Samuel 17 cooler, but I still find it valuable to find the capabilities with a little more cooling. I had too many Intel box coolers sitting around recently, and threw out the non-copper-core versions. So this one has a copper core like the one that would come with a top end i7 870. I don’t overclock with stock Intel coolers very often, and this exercise reminded me why….they’re terrible! I was unable to give the CPU any more than about 1.2V or it would go into the temperature danger zone. For me, that means above 95°C when running LinX. Now keep in mine, my ambient temps were very high, about 31°C during this testing. On the left, my maximum OC with the Intel cooler, 4140MHz with 1.2V and getting pretty hot! On the right, the Prolimatech performs the exact same test with the exact same test conditions….but keeps the chip 26 degrees cooler! Nice… silentpcreview also recently reviewed this cooler and came away impressed, but they also compared it to the Scythe Big Shuriken, which when equipped with the same fan was able to outperform the Prolimatech cooler. So be sure to consider that one as well.

4.1GHz on Intel cooler hitting - 98C max

4.1GHz on Samuel 17 - max temp 72C

Next, I went to work pushing the CPU back up to the limit of the Samuel 17….I ended up with a very respectable 4.6GHz with 1.35V. I also spent a little time overclocking the 5850 graphics card. With stock voltage I was able to increase the GPU clock from 725MHz stock to 850MHz, and the 1GB of GDDR5 memory from 1GHz stock to 1200MHz (4800MHz effective). I was pretty impressed with that overclock on stock voltage. I spent a bit over time pushing the GPU core further with the help of some additional voltage. But considering power consumption, my performance-per-watt diminished very quickly as I raised the voltage. So I decided to stay with the 850/1200 I had already established. These will be my 24/7 clock speeds.

final stable OC of 4.6GHz

5850 FurMark stable OC 850/1200

How well does it perform?

Well, this is where I run a bunch of tests to see what kind of performance a system like this can provide. I’ve run all the tests twice to compare the stock speeds with the benefits of overclocking. So, on the left are the results with the system at stock speeds, and on the right….overclocked.

Stock settingsCPU – 3.6GHzMEM – DDR3-1066 7-7-7-16

GPU – 725/1000

Overclocked settings CPU – 4.6GHzMEM – DDR3-1840 9-9-9-27

GPU – 850/1200

For all testing, I was using ATI’s Catalyst 10.7 driver package in my everyday Windows 7 x64 OS. There were no special benching type tweaks applied, and I didn’t change any settings from test to test. I configured each bench as best as possible to maximize image quality, including AF x16, and AA x4 unless otherwise noted in the screen shots. I configured the ATI Catalyst Control Center 3D settings as seen here:

CCC 3D settings used for testing

I started with the four Futuremark synthetic gaming benches, and then moved on to 8 real world gaming tests. Here are the results all summed up. I’ve also included the screen shots for each run at the bottom of the article.

3DMark results

Futuremark testing is my speciality, but not the point of this article, and have pretty limited relation to real world testing. However, it is nice to see some significant gains can be had from overclocking. If you’re interested in pushing your overclocks to the max for this sort of thing, please check out the Benchmark Buffet! which will help you understand the basics of competative benchmarking.

Game test results

As you can see, most of the tests scaled very well with my OC’ed settings. All of the games were into comfortable frame rates with the system OC’ed, although there were a couple that were just barely.

Crysis, as we all know is a beast of a game. I’ve played through the entire original game again, and Warhead (for the first time) with this system, with maxed settings and AA x4….and for the most part my frame rates stayed in the 30s and 40s. Which is very comfortable in my opinion. There are certain action situations where frame rates can drop, but they are rare. Overall, I was very impressed with the performance. To be able to play this title on a box too small for my shoes, that cost less than $1300 is fantastic!

Another title that was struggling is STALKER, this game is a beast as well, and shows nice improvements with multiple cores, it’d probably have a nice boost from a quad core CPU in this situation, although there probably are areas where it is GPU limited as well. I’ve been playing the game with this system, and it’s not perfect with the settings maxed. For better gameplay, I’ve disabled the DX10.1 features and it plays nicely.

The last game I’d mention is Grand Theft Auto IV which really needs a fast, multi core CPU for best performance. This title is definitely CPU limited on this system, and could really benefit from a quad core. You can see in the screen shot that it makes best use of the CPU power available, in comparison to all the other titles here sitting at an average of about 75% CPU utilization. Despite this, my settings are maxed, and gameplay only slows down to bothersome rates when it’s dark and foggy, and I’m driving around with my headlights on….in these situations, police chases can be difficult as FPS drop into the low 20s.

How much power does it consume?

This is a very serious concern for a SFF build because of the space limitations for a power supply. Most SFF cases designed for mini-ITX motherboards either have pico power supplies (with an external brick), or SFX sized units, like the one in the SG05 shown here. As I mentioned before, Silverstone has selected an 80+ rated 300W PSU made by FSP. FSP has been pretty generally well regarded over the years, and we’d expect this to be a decent PSU based on the specs. However, I’m throwing a lot at it, more that what is recommended by Silverstone, they specifically suggest upgrading to a larger unit in the case of running a 5850 GPU. And I’m pretty sure they’re also assuming a system running at stock speeds. So, I’ll really be putting this thing through it’s paces.

I started by measuring power consumption at 100% stock settings. First idle, then with CPU only, then stressing the CPU and GPU simultaniously. I did this with LinX set for 3 threads, and FurMark at the same time (if you set LinX to 4 threads, there isn’t any CPU power left to run Furmark, and the GPU is not pushed to the max). Then I restarted and applied my final OC’ed settings for the CPU and GPU, and repeated the testing with idle, CPU only, and CPU + GPU stress testing. Stock runs on the left, OC’ed runs on the right.

Idle power consumption. As you can see, idle power consumption was 75W with the system at default clock speeds, and 114W with my OC’ed settings.

stock idle - 75W

OC idle - 114W

CPU load only. Here, the stock system was pulling 133W max, and OC’ed it was pulling 194W max.

stock load - 133W

OC load - 194W

CPU + GPU load. Here, I ran three threads in LinX in order to load three virtual cores, but left the forth to process the GPU load with FurMark. If you perform this test with 4 threads in LinX, you’ll notice that the GPU load is not maxed out. Here I was drawing 272W max with the stock system, and 373W max with the OC’ed system. This was really getting close to the limits of the power supply, so I only ran it for a few minutes. But despite this, I was quite impressed!

stock CPU + GPU load - 272W

OC CPU + GPU load - 373W

*I also would like to point out that while the PSU is rated at 300W DC output, when I am monitoring the wattage with my Kill A Watt meter, I am monitoring the AC input power. In between the AC input and the DC output….the power is going through a conversion that is not free, the conversion consumes some of the input power (which generates heat)….exactly how much is lost during the conversion depends on many things….but basically, the 80+ rating on this PSU indicates that it is pretty efficient, but at max load it’ll probably consume more than 20%. But for safety, we should assume that it is 80% efficient at full load…that means if the system is drawing 300W, I would see 375W displayed on the Kill A Watt meter (300/0.8). So, by my calculation, any more than that 375W on the display for a sustained period of time should raise red flags for me.

In addition to the stress testing shown above, I kept an eye on the power meter during each benchmark run shown below, and annotated the maximum wattage that I observed during each.

AC power usage, real world VS synthetic

I was really quite impressed as I ran through test. I was expecting Crysis to be closer to the limit than it is. As you can see from the data in the chart…..STRESS TESTING IS WAY MORE TAXING THAN REAL WORLD USAGE!!! I think I may even try upgrading to a quad core CPU, and I think this PSU will be perfectly adequate…..but I will not be stress testing the GPU & CPU simultaniously if I do

Building a mini-ITX based SFF HTPC

I also rebuilt my HTPC, replacing the Zotac motherboard with the Gigabyte offering. Still housed in the Antec ISK case, the power draw was my biggest concern. With the Zotac board, I was unable to undervolt and underclock the system and retain stability. However, with the Gigabyte board this was accomplished without any trouble on my very first try. The most rigorous activity this system is used for is watching streaming video via Hulu desktop, and for watching 1080p x.264 video’s through my wireless N home network. With the Gigabyte board installed, I had to revert to a USB wi-fi dongle….a small price to pay for a low power, low noise, STABLE system.

The system is configured as follows.

Intel Core i3 540 dual core CPU
Intel GMA HD iGPU
1×2GB OCZ Gold DDR3-1600 (single channel) @ DDR3-1333
30GB OCZ Vertex SSD for OS (all media stored elsewhere, on network shares)
DVD-R/RW optical drive

I tested both boards at stock settings, and they were virtually identical in power consumption. Again, measured at the wall, both system’s idled at about 45W. While streaming a HQ show via Hulu desktop, I saw draw as high as 73W. And streaming 1080p x.264 content via my wireless network, the power draw peaked at 62W.

After installing the Gigabyte motherboard, I underclocked the CPU to 2.66GHz by lowering the multiplier to x20. Then I undervolted the CPU to 1.0V, the iGPU to 1.0V, the VTT to 1.0V, the PCH to 1.0V, and the vDIMM to 1.5V….the result? Idle power consumption dropped to 40W, Hulu playback dropped to 65W, and the 1080p stream only needed 58W to run properly. I’m actually very impressed be the performance per watt here, and I personally don’t think the ION platform has anything on this.

My HTPC - 40-65W AC

streams media from internet and NAS via wi-fi N

Conclusion

So, I asked a question at the start of this journey:

“Is it now possible to build a competent mini-ITX gaming PC, or should this market better shrouded to the HTPC and misc PC sections?”

My conclusion? YES! It is possible, mini-ITX has come a long way since its humble beginnings, and it has now reached the point where it is a viable platform for a powerful gaming machine. It’s amazing to me how much power I am able to pack into such a small box. It easily fits in my backpack, and only weights a little more than a desktop replacement “laptop”, yet packs a MUCH bigger punch! I look forward to continued growth in this segment over the coming years.

Gigabyte GA-H55N-USB3

This board has performed so far beyond my expectations I would recommend it to anybody…not just somebody on a budget, or for a boring office PC…ANYBODY! This motherboard is fully featured, stable, polished, and the BIOS is far more mature than you’d expect. It feels very similar to the P55A-UD7 top of the line LGA1156 motherboard from the same company, and all this for a very agreeable price. If you want a small PC, don’t look any further! I don’t hand out awards in my review, but if I did, this would be a Platinum! Thank you Gigabyte for providing the two H55N-USB3 motherboards used in this review!

On a side note, I upgraded my HTPC with the new motherboard when my wife wasn’t around. A couple days later she asked me if I had done something to the system. I wanted to know why she was asking, and she said “because it hasn’t had a hitch or crashed in days!” I was laughing out loud and told her about Gigabyte supplying an upgrade for it….she wanted me to make sure you know she is very grateful to be able to watch an entire show on Hulu without a crash, “Thanks Gigabyte!”

Prolimatech Samuel 17

When I first started this build, the Samuel 17 was not yet announced. I started by using the Intel stock HSF, and was really frustrated by the lack of decent options available at the time. When Prolimatech sent me the press release for this cooler, I was extatic and couldn’t wait to try it. I am impressed with its performance, and considering the size, I feel like they did a very good job with it. I also have some experience with the Scythe Big Shuriken, which would be the other major player in this size category. I tested it a bit on my first mini-ITX board with the Core 2 Duo E8400, and honestly I was not impressed. Since I have not compared the two coolers side by side, I cannot tell you which one is better, but I do have a lot of faith in the testing methodology of SPCR, and they said it’s better. So, Prolimatech may want to look at some ways to improve the Samuel 17 to match or better the Shuriken. If they do, I’ll try to get the pair for comparison. Until then, what I can tell you is that the Samuel 17 didn’t dissapoint in my testing, and that I am very satisfied with my stable 4.6GHz overclock. Something I couldn’t even come close to with the stock Intel cooler. Thank you Prolimatech for providing the Samuel 17 cooler used in this review!

Silverstone SG05

Silverstone did not provide the case used in this review, I bought it second hand, off the shelf from a local store here in Japan. However, it’s combination of size, features, and performance deserve recognition. So, I wanted to make sure I mentioned it in my conclusion. It’s a very impressive piece, and does great things for this emerging SFF market. I’ve seen tons of new designs popping up, but the majority of them are too small, or too large. Take for instance the new Lian Li PC-Q08 looks like a great new addition to the segment….but don’t be fooled, by comparison IT’S HUGE! Almost twice as big as the SG05 I used. On top of that, it costs more and does not include a PSU. Well done Silverstone, I’m looking forward to more cool products from you!

Screenshots

Below, you’ll find the screen shots from each of the gaming tests performed, stock clocks on the left, and OC’ed results on the right.

Stock 3DMark03	OC 3DMark03
Stock 3DMark05	OC 3DMark05
Stock 3DMark06	OC 3DMark06
Stock 3DMark Vantage	OC 3DMark Vantage
stock StreetFighterIV	OC StreetFighterIV
stock GTAIV	OC GTAIV
stock Far Cry 2	OC Far Cry 2
stock Dirt2	OC Dirt2
stock Crysis GPU	OC Crysis GPU
stock Crysis CPU	OC Crysis CPU
stock Crysis Assault	OC Crysis Assault
stock ClearSky	OC ClearSky
stock Batman	OC Batman
Stock Aliens vs Predator	OC Aliens vs Predator
Stock Call of Juarez	OC Call of Juarez

[Blog] College-Kid’s-Budget Gaming PC Building – Things to Avoid and Things to Adore Like Your First Pony

TheDramaLlama — Thu, 01 Jul 2010 02:29:44 +0000

Hello, everyone! My name is Alvin (a.k.a. TheDramaLlama) and I’d like to poke fun at all the old… er, say hi to everyone here at TechREACTION.

For my first blog, I’m going to do a rough overview on budget PC building. No, not your run-of-the-mill budget PC building, I’m writing about college-kid’s-budget gaming PC building, since college kids like me belong to a fantastic little niche, filled with people in situations I like to describe as, “Mommy, I have no money but I love expensive things.”

However, I’m not going to go into too much detail, both because this is my first post, and also because there are plenty of sources for that kind of neck-deep technical stuff already. Instead, I’m going to go over the lessons I’ve learned in building my first budget gaming PC, a rather peckish $1,000 affair with which I am now 90% satisfied. The other 10%, of course, belongs to the group of people I like to describe as, “I’m filthy rich and want more more more more frames.” If only.

Anyway, let’s start off with the specifications of my rig.

For my monitor, I got a 25″ Hanns-G 2ms LCD at 1080p. Deals, deals, deals.

For my CPU, I got an AMD Phenom II X4 at 3.2Ghz. I have yet to overclock this, although I’ve been feeling the bottleneck lately and will probably aim for 3.8 once I get a nice new cooler. Although to be honest, the “stock” cooler included with the processor was surprisingly large and effective. (I promise I’ll have pictures for my next post.)

For my GPU, I got an XFX Radeon 5770. I’ve been having some driver issues with this card (see forum for details), so I’ve been unable to overclock it; however, I’ve heard good stories of >1Ghz overclocks, so I’m holding out for now. I’m also planning on another 5770 in the near future and some Crossfire goodness, if budget allows.

For my motherboard, I have the Gigabyte 790X AM3 USB3 SATA6. Quite a nice futureproof motherboard for about $150.

And for RAM, OCZ 2×2Gb DDR3-1600. Everything else is the usual.

My first piece of advice when looking for a budget rig is to never, ever, ever look at top-of-the-line parts. They will make you feel bad, and you will not be able to afford them. Instead, bump it down a notch. For example, instead of the Phenom II X6, show some love for an X4. Instead of an i7, grab yourself a nice, fat hunk of i5. AMD chips tend to be on the cheaper side as compared to Intel, although they’re naturally outperformed, as well. I would personally never pay more than $200 for a budget PC’s processor, but then again, it’s up to your budget.

My second piece of advice is to get the most expensive mid-range graphics card you can afford. Don’t even bother looking at the 5970, or even the 5870 for that matter; instead, go for a nice mid-range 5770. Keep in mind that the price margin between the 5770 and the 5850 is roughly 40% (i.e. the 5770 is 40% cheaper than the 5850), but, in a brilliant stroke of marketing genius by ATI, you sacrifice about the same amount of performance. I chose the 5770 because $150 fit my budget, but again, it’s up to you.

My third piece of advice is to get a good, futureproof motherboard. I’m not particularly familiar with Intel motherboards, so I won’t say much about them, but I’ve heard horror stories about socket compatibility with the newest i7 chips; make sure you do your research before you invest in a motherboard that will become incompatible with new chips in a year or two. A good idea is to invest in the top-of-the-line previous generation motherboard; they are usually relatively futureproof, as well as extensively tested and reliable.

My fourth piece of advice is to never skimp on essential components. You can skimp on a case if you’d like to, or get a cheap-ish power supply, or settle for value DDR3 RAM, but as far as your CPU and GPU are concerned, you should never go low-end. Remember, you’re building a budget gaming PC, not a budget bucket-o’-bolts. All the savings in the world won’t matter when you dent your tower with a sledgehammer fist out of frustration.

My last piece of advice is to wait for good deals. It’s not always financially prudent to order all of your parts at the same time, and certainly not from the same vendor; you’ll often find fantastic deals on parts from all over the place, and more often than not, it will be worth your time to order your parts piecemeal and hunt for deals. You’d be surprised at how much you could save just by building your rig over the span of a week or two instead of a single day.

Well, that’s all for me for now. If you’re a cash-strapped college student, I hope you found this first post somewhat useful; I doubt all you rich folks with i7s will find much to glean from this.

Till next time,

TheDramaLlama