Blogging Hub show why you must never protect your wireless network with WEP

Think you're secure with basic WEP? You're not. We're going to demonstrate exactly why you need to tighten up your Wi-Fi security by showing you how easy it is to hack a relatively unprotected network. We'll give a working demonstration of the weakness in your network's security and how they can easily be exploited. This is the best way of creating the incentive to shore up that security by using together encryption.

In this case, we're going to crack wide open the WEP security that far too many home users still rely on. Whether this is through a lack of knowledge or the belief that they have no assets worth stealing is neither here or there. WEP is insecure, as we'll show you.

Before we begin, however it's imperative that we make it absolutely clear that this not a technique you should ever try in the wild. Doing so is clearly illegal and we do not condone hacking of any kind. The tools and techniques we will be demonstrating here are not toys and they could wreak havoc in the wrong hands. So please be responsible with what we're about to show you.

Backtracking

For this demonstration, we're going to use some of the tools that come bundled with the Swiss Army Knife of online security: BackTrack Linux. This contains all the security tools that the network security professional and the hardened hacker alike would need to carry out a completely exhaustive examination of a network. Everything we need is installed, configured and ready to go. Being Linux, it's also open source and therefore free to download. The developers have just released a new version called BackTrack 5.2. You can download the ISO file here: www.backtrack-linux.org/downloads. This is designed to run from the DVD drive as a live CD, so copy it to a fresh DVD using you favorite ISO burner. If you don't have one, you can download and install ImgBurn free from www.imgburn.com.

After creating the BackTrack DVD, boot it from a computer with a wireless networking card. When the Boot: prompt appears, press [Enter]. Select the default text-only version from the bootup menu, and when it finishes booting you should be met with a # prompt and a cursor. Type the word startx and press [Enter] to load the graphical desktop.

At the top of the screen you'll see a menu bar with a picture of a terminal. Click this and a terminal window appears. Enter the command iwconfig and a list of network interfaces appears. One of these should be call wlan0. This indicates that the wireless drivers supplied with BackTrack can see your wireless network card.

The utilities we'll use are all compatible with several industry standard Wi-Fi chipsets - namely rtl8187, rt2570 and rt73. This means that just about all modern wireless cards should be okay to use. However, if wlan0 doesn't show up, you could try booting BackTrack on a different computer instead. Otherwise, borrow a USB Wi-Fi dongle to see if BackTrack recognises it.

Collecting data

The standard advice is never to use WEP to secure a Wi-Fi network, but why is this? Put simply, it's because when given enough packets to analyse, cracking software can recover the password (called a passphrase in Wi-Fi speak) to join such a network, as we'll demonstrate.

Any hacker worth their salt begins by passively collecting data about potential targets, and this is exactly how we'll begin. In a terminal window enter the command airmon-ng start wlan0.

This command begins collecting data about nearby wireless networks and sending it to a special dump where other programs can pick it up and use it.

When you type in the command, you may get a warning about DHCP. Ignore this and look instead for the message 'Monitor mode enabled on mon0', which indicates that everything is running as it should be in the background. We'll now find out what is being recorded to this mysterious mon0. Start by entering the command airodump-ng mon0.

Note the extra letter 'o' in the airodump-ng command that wasn't present in the airmon-ng. This catches many people out. The screen begins to fill with a list of networks. The BSSID (basic service set identification) column contains the globally unique hardware MAC address of each router or base station within range. The ESSID (extended service set identification) column is the given, friendly name of each of the networks.

Below this is a shorter list of connections being made to those networks to which they're connecting are displayed in the BASSIC column, and the MAC addresses are also displayed in the Station column. Usefully, the Probe column gines you the name of the network that's currently being accessed.

If you live in an area with lots of Wi-Fi networks and plenty of activity, both lists will expand and contract of ten, making them difficult to read. To prevent this, with the terminal selected, press the [R] key twice. This switches off the auto sort facility. Pressing [R] again will switch it back on again. You can also cycle through the columns and have the program sort by these using the [S] key.

To stand any chance of gaining access to a WEP network, there  must be a current connection to it. On the basis that you're cracking your own network, connect a computer to it and it will appear as a station in the lower portion of aidodump-ng's output. Next, make a note of the station MAC address and then leave the utility running for a few minutes until all networks within range are recorded, then press [Ctrl]+[C] to stop the program.

Homing in

Next, we need to collect data running between just the router in question and the computer connected to it. To do so, we re-run airodump-ng, but with some filtering switches in place: airodump-ng -c<channel>-w<output>--bassid<MAC> mon0.

<Channel> is the channel number of the router in question and <output> is a filename in which to store the captured packets (use 'dump' or something similar for name). <MAC> is the MAC address of the router.

By running this command, you only see your router and the computer connecting to it. The collected data is also captured in the dump file (actually more than one dump file, but we don't need to worry about that).

As the command runs, pay special attention to the #Data column in the top of airodump-ng's output. These are the actual packets going between the computer and the router, and they contain the WEP passphrase. The problem is that we need between 5,000 and 25,000 packets to crack the passphrase.

We need to generate a large number of packets, and luckily BackTrack contains another utility that will do this for us. Enter the following command, where <BSSID> is the MAC address of the router and <STATION> is the MAC address of the computer connected to it: #aireplay-ng --arpreplay -b <BSSID> -h <STATION> mon0.

Note the double dash before arpreplay. Airplay-ng generates packets that are faked to seem to have come from a particular computer and spewed to the router. You define the MAC address of the source PC with the -h switch. In this case, it's the computer connected to the router.

Every packet sent from the router contains an encrypted version of the WEP passphrase. The reason for generating lots of extra packtes is that with a large enough sample of encrypted passphrase, we can use another utility to analyse the captured data and start to make a statistical guess at what the underlying plain text might be.

Sudden traffic

Airebplay-ng might be slow to get going, but after a few minutes it suddenly begins generating plenty of extra traffic. Let it keep on going until airodump-ng shows something over about 5,000 in the #Data column, after which time you can stop it using [Ctrl]+[C]. You can also stop the airodump-ng command at this point. It doesn't matter at all if you massively exceed the number of packets. In fact, more data makes it easier to crack the passphrase.

Note that while you're generating these packets, the WLAN light on the router should usually be flashing to say that data is being sent and received. As the associated internet light isn't also flashing to signal through traffic, the target of the packets must be the router itself. This serves as a quick method of telling if someone is attempting this kind of attack against your own network, even (or rather, especially) when none of your own computers are connected at the time.

With the results of bombarding the router with packets containing the encrypted WEP passphrase stored in a temporary file, we can finally attempt to crack it. To do so, we use the aircrack-ng utility. There's no guarantee that this will work first time, but if it fails, you just need to collect more than the previous 5,000 or so packets. To run aircrack-ng, enter the following command: #aircrack-ng -z -b <BSSID> dump*.cap.

Again, <BSSID> is the MAC address of router. The argument dump*.cap enables aircrack-ng to read all the files that airodump-ng created in the current directory, which you can see by entering the command ls.

If Aircrack fails to recover the passphrase, it will tell you and give a recommendation for the number of packets it thinks it'll need to make an accurate guess. A WEP passphrase of 'hello' took nearly 15,000 packets to crack, but once aircrack-ng had these at its disposal, the whole process took 10 seconds. With 30,000 packets, this time dropped to just four seconds.

So, now we have the passphrse to the WEP router we have compromised it to the point where we can join the network from Windows just like any legitimate user would. Once joined, open a command prompt and then type the command ipconfig to check your network connection.

Everybody needs...

You can surf through the compromised WEP network, use a network mapping utility like our friend Zenmap (www.nmap.org/zenmap) to find other computers on the network and, in some cases, even mount network shares to read the information they contain or deploy a network traffic sniffer to possibly capture some useful username and password pairs. This is why WEP security is no security at all. If, during the process of working through this tutorial, you discover one or more WEP protected networks in you vicinity, it would be illegal to hack them, but the act of a good neighbour would be to knock on the door and help bring th eowner into the 21st century.

The new go-to gamer's chip, but is it a Sandy Bridge upgrade?

This is it. The biggie. The replacement for our favorite CPU of the past year or so. The new Quad-Core Intel Core i5 3570K, one of two shine processors from Intel's new 22nm Ivy Bridge family of CPUs, has plenty to prove.

We'll be measuring this new chip against our old favorite, the Intel Core i5-2500K from the Sandy Bridge generation. This is the best all around gaming CPU ever, and all the chip most will ever need. Or maybe that should be the Core i5-2550K, which was very minor clockspeed bump over the 2500K. In truth, they're much of a muchness.

At first glance, you might wonder whether Intel Core i5 3570K is actually a new chip, so similar are the headline specs to its predecessor. With four cores and no Hyperthreading support, there's not a lot of extra CPU hardware. The clockspeeds and cache haven't budged an inch, either. As before, we're talking 3.4GHz nominal, 3.8GHz Turbo and 6MB of cache.

Dig a little deeper and the differences emerge. For performance enthusiasts, the most important upgrade is the shrink from 32nm to 22nm process technology and the introduction of Intel's 3D Tri-gate transistors. The upshot is what Intel is calling a 'tick-plus'.

A tick in Intel speak means a die shrink of an existing architecture, where as a tock is a new design using the old manufacturing tech. So the existing Core i5-2550K is part of the Sandy Bridge tock family and the new Intel Core i5 3570K is an Ivy Bridge tick. Got that?

Anyway whether or not the tick-plus label makes any sense, it'll be Intel's own CPUs that Intel Core i5 3570K will have to beat. AMD's FX Bulldozer chips, like the AMD FX 8150, simply can't compete when it comes to per core performance, and that's what you need for a great gaming CPU. That's what the old 2500K and 2550K were all about, and is what Intel Core i5 3570K will have to deliver to take over where those two left off.

Intel Core i5 3570K old and new

So, the new Intel Core i5 3570K is intriguing mix of old and new. What does it look like when you life the bonnet?

The CPU cores themselves are pretty much the same as before, barring a few minor changes. It's the process tech that promises the most. Intel reckons the new 22nm 3D Tri-gate transistors are more than just your typical process shrink. They offer much better current control, less leak and improved flow.

In theory, that give Intel lots of options. Firstly, with each core consuming less power and taking less space, it could add more cores. But it didn't. Alternatively, Intel could have ramped up the clocks. But it didn't. Finally, the new Intel Core i5 3570K could be more power efficient than the old 2550K. Here, finally, we have some tangible evidence that the new process deliver as billed. Where the 2550K is rated at 95W, Intel Core i5 3570K has a 77W TDP. That said, with the core count, clocks and 6MB cache carried over, one thing is clear: if Intel Core i5 3570K is to deliver improved performance, those minor core revisions will have to do the work.

Shift your attention to the graphics, however, and things look a lot more interesting. If you care about integrated graphics, that is. The Intel Core i5 3570K gets the full fat HD Graphics 4000 core, which means 16 execution units where the old 2550K had just 12. Each unit gets an upgrade too, with more power and an upgrade from DX10 to DX11 support.

The final change of note is the addition of PCI Express 3.0. As before, the main PCI Express controller is integrated into Intel Core i5 3570K's die and as before you get 16 lanes to play with regards discrete graphics cards. But with double the bandwidth of PCI Express 2.0, Intel Core i5 3570K and other Ivy Bridge chips make a more robust platform for multi-GPU graphics.

We've established that the CPU side of the Intel Core i5 3570K is little changed, but big things have happened with the graphics. How does Intel's new mainstream crowd-pleaser perform when you fire her up?

"As For The Graphics, Intel Has Indeed Upped Its Game With This Chip"

With the clockspeeds and cache staying mostly steady and the cores revised rather than redesigned, we didn't have high hopes for a dramatic boost in performance. That's pretty much how it plays out. In the Cinebench 11.5 professional graphics rendering test, the 3570K edges the old 2500K 1.6tps to 1.49tps. Factor in the 100MHz clockspeed advantage of Intel Core i5 3570K and it's clear Ivy Bridge isn't doing anything dramatic.

It's similar story of very modest performance gains in x264 HD video encoding and the CPU -intensive World in Conflict game benchmark. Intel Core i5 3570K is quicker, but the difference isn't nearly enough that you're ever going to feel it.

That said, we have been expecting some showbiz moves when it comes to overclocking. After all, Intel has been bullishly bigging up is fancy new 22nm transistors in all their 3D Tri-gate glory. Oh, and as before, that 'K' suffix indicates a full unlocked CPU multiplier for easy overclocking. Time for them to deliver. What you make of the 4.9GHz result depends on the angle you're coming form.

On the one hand, that's 1.5GHz higher than the Intel Core i5 3570K's guaranteed clockspeed for four loaded cores, which is one hell of an overclock with a simple air cooler. On the other, it's just 100MHz higher than the old 2500K manages. So much for that fancy 22nm tech, then?

Perhaps, but what's really interesting for existing Sandy Bridge owners is that Intel's new Z77 platform seems to release more overclocking potential from the outgoing 32nm generation. We reckon the Z7 is worth a good 200MHz to 300MHz with Sandy Bridge chips, which is intriguing even if it undermines the attraction of the new Intel Core i5 3570K and its Ivy Bridge brethren

Intel Core i5 3570K game on

As for the graphics, Intel has indeed upped its game. World in Conflict at 1,280 x 800 jumps from a sluggish 19 frames per second on the 2500K to a borderline playable 27 frames per second on the new Intel Core i5 3570K, for instance. Skyrim performance is up from 21 frames per second to 25. In reality though, you still need a discrete graphics card for enjoyable gaming.

It's hard to know what to make of the QuickSync video transcode core though. Performance i sup by roughly 10 percent. It's certainly impressive to observe a whole hour-long HD episode of a TV show crunched down for smart consumption in just a few minutes, but with application support remaining somewhat limited, QuickSync still isn't the killer feature it could (and indeed should) be.

The new Intel Core i5 3570K has everything that made its Core i5 2550K such a winning CPU. Per core performance is the best you can buy, it overclock like there's not tomorrow and it's relatively affordable. We're also pleased that it remains compatible with Intel's LGA1155 socket. Intel has created a rod for its own back in recent years by delivering ever more performance. The new Ivy Bridge generation basically puts on the brakes. No more cores, no more clocks, and no more overclocking headroom. The improved graphics don't make up for that.

Intel Core i5 3570K technical analysis

With AMD not really turning up to this party, the comparison is with Intel's Sandy Bridge chips, most notably our old favorite the 2500K. We've ticked our usual boxes with Cinebench rendering, x264 HD video and gaming on the CPU side. Overclocking is a key competence for the Intel Core i5 3570K, and we're pleased to note it's another 5GHz chip. We've also had a look at the integrated HD Graphics core, both in games and its QuickSync transcode engine. It's a decent step forward compared to the previous core, but it's still not a great gaming solution.

Single-threaded CPU performance

---

Multi-threaded CPU performance

---

CPU encoding performance

---

QuickSync encoding performance

---

CPU graphics performance

---

Overclocking performance

---

A diminutive orange offering

This really is more like it. We've been impressed by the overclocked AMP! cards from Zotac before - most recently and fantastic GTX 670 AMP! - and it's trying to continue that trend with the GTX 660 Ti AMP! edition. Where the other manufacturers have approached the new Nvidia card as just another big GPU with high clockspeeds, Zotac has been a little smarter.

Using the same reference PCB as the GTX 670, it's managed to create a card with a far smaller footprint than the competition. You've got to give it props for taking a different approach to the others - especially considering it's managed to hit the same GPU frequency as the slightly pricier Gigabyte card.

The AMP! is right up there with the big boys in terms of gaming performance. In fact, in some instances it manages to hit a higher boost clock for longer. That orange-tinged fan array helps the chip spit out the occasional higher frame rate than the GTX 660 Ti competition.

That still doesn't change the fact there are AMD cards for the same money, which are generally speedier in gaming frame rates. Yu can also pick up a vanilla GTX 670 for only another $32 more, and that will take both the GTX 660 Ti and HD 7950 outside for a good kicking. That's before you think about the incredible overclocking potential the GTX 670 still has in its pocket.

While the GTX 660 Ti AMP! edition is definitely the finest example of Nvidia's new card we've seen, that still doesn't mean we'd recommend it over the superior GTX 670 or HD 7950 it's up against. This is not the GPU to pick your battles with, Zotac.

Fans of the big green graphics company rejoice, Cyberpower is showing its love of all things Nvidia here with a rig that's ripe for the fanboys. The SI has put some decent parts together for your pleasure, all of which are bathed in the eerie, mushy-pea glow of the Cooler Master 690's LEDs.

We say 'decent', because we're not entirely enamoured with Cyberpower's choice of components. It's the choice of graphics card that really has us stumped. We know that it's a new graphics card, but the GTX 660 Ti really isn't the sort of component we'd want in a rig we'd just paid grand for.

That's especially true when we're seeing overclocked GTX 670s, GTX 680s and an overclocked HD 7970 in other machines.

Scan was the only other system integrator to opt for Nvidia's mid-range make-weight, and together they prop up our list of gaming benchmarks. up against the superior GPUs of the other rigs in the test, the GTX 660 Ti looks very, very weak.

Still, it means Cyberpower has been able to stick a Blu-ray drive in the rig, and we all know how useful they are for gamers. To be fair though, the cost of a Blu-ray ODD is probably not that much more than a DVD drive these days, so we can't be too harsh on it.

The cash freed up by going for a weaker GPU has arguably gone into the choice of solid state drive. At 128GB, the Crucial M4 boot drive certainly isn't vast, and will only let you install a few of your most frequently played games, but it's nevertheless one of the strongest SSDs in the test. It offers fantastic 4K random read/writes in the AS SSD benchmark, and delivers excellent boot times.

Along with the AdvanceTech rig, it's also one of only two in this test to use water-cooling on its overclocked processor, and like the AdvanceTech, it actually has relatively conservative overclock applied to it.

At 4.5GHz its i5-3570K is certainly no slouch, but 4.6GHz is what we've come to expect from even the lowliest Z77 motherboard. This build is hanging out around the bottom of the CPU performance charts, and with a high quality liquid chip chiller installed in that glowy chassis, we were expecting more.

We can't help but feel a little disappointed by this latest Cyberpower system. With its bold and brash Nvidia branding, we'd have expected graphics to have been one of the components it prioritised above the others. Sadly, by going for the GTX 660 Ti, Cyberpower has left the machine languishing in the benchmark doldrums.

With the slightly conservative overclock on the i5 CPU, combined with the weaker GPU, the Infinity Apollo comes up short almost across the board. Only the decent showing by its choice of SSD gives it any semblance of pride against the competition. This is by no means a poor gaming PC, but if you're thinking of spending this much money, the majority of the other machines in the test would get our vote.

Blogging Hub shows how to save time with Task Scheduler

Every day when I turn on my computer, I find myself firing up the same software and reading the same websites. But instead of launching apps manually or typing in the URLs, Windows opens them up automatically. While I'm away from my computer, it also sometimes performs routine maintenance by defragging the disk or scanning for viruses.

The tool that turns your docile computer into a caring virtual nanny is the Windows Task Scheduler. All you have to do is make a list of tasks you'd like to automate and configure the Task Scheduler to handle them for you.

Task Scheduler can launch any app that you have installed and gives you extensive options to specify the time intervals when you want these task to run.

Step-by-step: Run tasks automatically

Configure Windows Task Scheduler to carry out the maintenance jobs you might forget

 1   Launch Windows Task Scheduler

The easiest way to get to the tool is by typing Task Scheduler into the Start menu search. The main interface gives you an overview of the tool and lists all to active tasks. Before you create you own tasks, right-click on the 'Task Scheduler' library and select the 'New folder' option to create a folder that house all your custom tasks.

---

 2   Create a task

Select the folder you've created and go to 'Action > Create task'. This will open the Create Task dialog box, which helps you define the task. Start by entering the name of the task. It's safe to continue with the default Security options, which run the task as the current logged-in user, but you can also run the task as another user.

---

 3   Set triggers

Now switch to the Triggers tab and click the 'New' button to define the conditions that will trigger you chosen task. You can either start a task at a particular time of the day during particular events, such as the computer booting up. For now, just select 'At log on' from the drop-down list, which will start the task when a user logs in.

---

 4   Specify an action

Head to the Actions tab and click 'New' button to define the action that will take place when the task is triggered. The tool can perform three types of actions. For now, select 'Start a program' from the Action drop-down list and point to an app in the space below, with the name of a website in the optional 'Add arguments' box.

---

 5   Different action, same time

After adding the action in step 4, click the 'New' button and ask the tool to launch another program that you want to run at start-up. I set it to start Excel and my to-do list by pointing to excel.exe in the Setting  section, along with the name of the file in the 'Add arguments' box and its location in the 'Start in' box.

---

 6   Defrag disk when idle

Now create a new task that will defrag the hard drive with the Windows Disk Defragmenter whenever you PC is idle. When creating this under the Action tab, point to the app you want to launch (usually C:\Windows\System32\Defrag.exe) then go to the Trigger tab and make sure you select the 'On idle' option when creating the trigger.

---

 7   Set additional conditions

When you select 'On idle', you have to define additional settings that alert the Windows Task Scheduler when the computer is idle. To set these, go to Conditions tab and select one of the preset times from the drop-down list. For laptop users, there's also a checkbox that will ensure this task only runs when the PC is plugged in.

---

 8   Display a message

For a task that runs in the background, it's helpful to create an action that displays a message to alert the user. For our defrag task, create a new action, select the 'Display a message' option from the drop-down list and enter a message describing the event. I also use this feature to display messages reminding me of the day's chores.

---

 9   Control task behaviour

You can further tune the behaviour of a task from the Settings tab. In most cases, the default settings should suffice, but some might need tweaking. For example, if you've set up a virus scan at a particular time and day, and your PC isn't switched on at that time, there's an option here that will run the task the next time the computer is on.

---

 10   Export the task

Once you've set up Windows Task Scheduler , it will improve your Windows experience considerably. I've just scratched the surface and show you how I use it every day. If you have multiple computers with the same programs installed, you can right-click on a task to export it, then import it into a custom folder on another computer.

---

  What'sthe real value of future-proofing?

While every man and his dog is throwing SATA 6Gbps interfaced SSDs at us as if there's no tomorrow, where does that leave everyone that is still packing SATA 3Gbps systems? Surely they deserve some modern SSD love too? This is the premise behind Crucial's latest v4 family drives: build a 3Gbps SSD at a good price so people with 3Gbps systems don't have to buy expensive 6Gbps models with features and performance they can't tap into. Seems vaguely sensible, right?

While that would have been a sound idea when 6Gbps drives were relatively expensive, real life has dealt the Crucial v4 a kick where it hurts. The SSD market has become a little more cut throat sooner then most people expected and more than many ever thought possible. Right now you can buy SATA 6Gbps drives for almost the same price as the list price of the v4. Even Crucial has had to reduce the price of the 6Gbps m4 128GB to the point where it's only ten dollar more expensive than this drive. Go figure.

Performance

Crucial quotes sequential read/writes figures for the v4 at 230MB/s and 175MB/s respectively, which appears to be a bit on the conservative side as we managed to get 274MB/s reads and 233MB/s writes via the ATTO benchmark with review drive. That's actually pretty decent performance for a mainstream 3Gbps SSD drive, so things looked relatively promising at first.

Interestingly, the v4 contains something that's rare in these days of Marvell, Samsung LSI SandForce controllers. It's actually built around a Phison PS3105-S5 chip. The controller uses custom Micron firmware (version S5FAMM22) and supports wear leveling, garbage collection and, importantly, TRIM.

Unlike the LSI Sandforce controllers, the data isn't compressed - instead a cache chip is used to buffer data, much the same way as the Marvell controller in the m4. This shows up in 4K AS SSD tests where the two Crucial SSDs are close in performance terms, unlike in the AS SSD sequential read performance, which shows the performance advantage of the 6Gbps interface rather well.

Internally the drive is neatly laid out with one side of the PCB occupied by a Micron 166MHz DDR chip, which looks after the cache duties sitting alongside the Phison controller. These two are joined by four of the eight 16GB 25nm Micron MLC synchronous NAND chips that make up the storage capacity of the drive, with the remaining four NAND modules sitting on the other side of the board.

While the prices for the faster 6Gbps drives were much more expensive than 3Gbps, Crucial's v4 would have made a lot of sense. Unfortunately the bear pit that is the SSD market at the moment has, in many cases, reduced that premium to almost nothing. It's pretty much a no-brainer then - pay the extra and get a drive that's ready for you to upgrade, should you wish to at a point in the future.

Technical analysis

It may seem unfair to pitch a 3Gbps drive against a 6Gbps drive, but in the v4's case it's relevant as Crucial's own m4 is so close to it in price. You won't get the performance in a 3Gbps system, but you'll have a ready-to-go drive should you wish to upgrade.

Sequential read performance

Sequential write performance

4K random write performance

And in the Asus corner...

The battle between Asus and Gigabyte has traditionally been a heavyweight contest, though in recent years it's become more or a ritualised beating, as the Asus boards have consistently out-punched Gigabyte at every level. With the release of Intel's Ivy Bridge setup and Z77 platform, however, Gigabyte has made a dramatic return form.

Now the head-to-head is once more a battle of the giants, and nowhere more so than between the two premium Z77 motherboards we're checking out this issue.

They are both around the $400 mark - fortune compared to most of the fairy resonable Z77 boards tested - and come with more ringing and whistly things than you could shake a thin piece of dead tree at. They also both come with connections for the new Intel/Apple love child, Thunderbolt.

Lightning Thunder

Where the Gigabyte board has a pair of connections for the new interface, this Premium effort from the Asus has just a single port on the rear. Still, it shouldn't have trouble rocking triple-screen configurations, with HDMI and fully Displayport connectors alongside that Thunderbolt socket.

The Asus board also seem happier running Thunderbolt, as it managed faster data transfer benchmarks is storage tests. Used a pair of Intel's SSD 330 drives in a RAID O array to make best use of the available bandwidth. In both the maximum read-write and sequential read/write tests, the Asus board posted faster results - but not by much. On the 4K random test though, which is more of an indicaiton of general Windows responsiveness, the Asus board's write speed is 14MB/s higher. That's a big difference in 4K random terms.

Sadly for the Asus board, its usual dominance of gaming benchmarks has been undone by the superior performance of the top-end Gigabyte Z77. It was similar situation with the Gigabyte Z77-UD5H, and show just how much the opposition has caught up.

Usual service is resumed with regards to overlcocking though, as this board gave a solid 4.85GHz clock speed on i7-3770K. That's only a shade quicker than the Gigabyte's 4.79GHz, but still makes a difference.

Another trick the Asus mobo has up its silicon sleeve is the volume of PCIe 3.0 slots. With four full x16 connections, this board will happily rock the four-way SLI world. Sure, it'll get cramped with four GPUs, but you're only going to be covering the power and reset buttons with the fourth card, and with your chassis closed that makes no difference at all.

It comes with a 32GB mSATA SSD too, allowing you to pair up a traditional hard drive with the speedier SSD to give it a boost. The Gigabyte board has the same slot, but doesn't ship with flash memory for the Smart Response lovin'.

It's a tough call between these Asus and Gigabyte boards. The Asus has a little more in the feature set, but the Gigabyte has excellent gaming performance chops. For the money the Asus is probably better value, but it's a close run thing.

Technical analysis

It's a real ding-dong battle between this and the Gigabyte motherboard, and for much of it you'd have to favor the Asus offering with its faster Thunderbolt speeds and higher overclock. The Gigabyte has the edge in gaming performance though, so there isn't clear overall winner

Video encoding performance

4K random Thunderbolt performance

Maximum overclock

As reference as you're likely to get

Pricing is such an important part of the graphics card war, the mere $32 separating the Asus and EVGA cards actually means a hell of a lot. When you're operating in such a saturated segment every penny counts. Like the Asus card, this Superclocked EVGA GTX 660 Ti is factory overclocked, but because of its policy of cherry-picking GPUs and leaving the reference cooling in place, the EVGA isn't clocked as high.

The card will still regularly top the 1GHz mark in-game, but the extra 65MHz at the base isn't really going to set the gaming world alight. And so it proves in benchmarks, with the EVGA resolutely coming in behind all the other GTX 660 Ti cards we've seen so far.

That's exactly what you would expect against the more expensive card from Asus, but there are also the Gigabyte and Zotac-shaped thorns in the EVGA card's side. Both cards are practically the same price as the EVGA, but come with higher base clocks among other things.

This also means that it sits behind the similarly priced vanilla AMD HD 7950 in most of tests - even more so if you bring the latest HD 7950 with Boost into the equation, with its own raised clock speeds. Things are pretty close in terms of regular gaming, but when you look at the more compute-focused titles the gap starts to look covernous.

Where EVGA stands out though is in its excellent support setup and the option to pick up a 10-year warranty for your new GPU. Still, it's the slowest of the current crop and can't make up for that deficit by being a few dollar cheaper.

The Chillblast machine is one big, fat bestie. The chunky Zalman chassis gives it a real sense of presence on your desk, and the internal goodies match that sense of scale, with a heftily overclocked i5 CPU and similarly overclocked GTX 670 going the graphical grunt work.

Sadly though, it's in the overclocking department that the Chillblast Fusion Thunderbird gets a serious black mark. The Core i5 3570K gets a headline-grabbing 4.8GHz clockspeed - easily the highest clocked chip in the test. If it ran solidly at that speed, we'd be all over this rig like a cheap LED strip. Unfortunately, while it boots happily with these settings and will allow you to navigate around Windows with impunity, as soon as you start stressing the chip it begins to throw a wobbly. We couldn't run through either Cinebench or X264 encoding tests without it crashing.

All it took was a quick trip to BIOS land, knocking the multiplier down a notch, to hit a rock-solid 4.7GHz. If the CPU couldn't run under load at that clockspeed, the rig shouldn't have arrived with those settings. It's a shame, because we've only had good experiences with Chillblast machines in the past, but if you've just spent $1k+ on a machine, the last thing you want it to do is fail under load.

That aside, things are decidedly rosy for the Chillblast Fusion Thunderbird. This rig has one of the most balanced spec sheet out there, weighing up straight-line gaming performance with general PC functionality. Part of the reason the machine functions at all at 4.8GHz is because Chillblast Fusion has used a strong Asus motherboard, where others have used more budget-oriented options.

It has also opted for an overclocked Palit GTX 670 Jetsream in the graphics slot, which delivers frame rates practically on par with those posted by the two machines with GTX 680 cards inside. It's still short of performance compared with the overclocked HD 7970 of the AdvanceTech rig, but then you also get the added loveliness of things like PhysX and TXAA with the Nvidia cards

One of the big bounses of this rig is that it's one of only two PCs in the test with a chunky 240GB SSD as a main OS drive. It's also got a 1TB HDD backing it up for all your data needs, but with the decent SSD installed you could get a large chunk of your current gaming library on those solid state memory chips without things getting too crowded. The Scan machine, the other PC with a 240GB SSD, has a weaker graphics card, and therefore falls well short in the gaming benchmarks. That Mushkin SSD also helps the Fusion Thunderbird get the fastest boot time of the lot.

If it wasn't for the major black mark around the overclock, this review would be sickeningly glowing. Chillblast Fusion has found an excellent balance between the speedy GTX 670 i5 CPU and 240GB SSD. It's not the fastest rig in the test, but it has been intelligently specced-out, despite the Blu-ray drive, and no mistake. You can't argue with that impressive two-year warranty either; good work, Chillblast.

Big, expensive and life-altering? Well, two out of three ain't bad

Like most PC peripherals, monitors mostly sell on price. That's because consumers are largely a feckless bunch of mouth-breathers who can't see past pricing and a small half handful of headline specs. It's $200? It's 27 inches? It's shiny? It's sold. In that context, Samsung slinging out a premium-priced 27-incher is either bold or bonkers.

Samsung's own Series 8 model is around the $800 mark, and recent months have seen a glut of $323 efforts being bought from Korean eBay merchants. Meanwhile, you can pick up a number of 30-inch alternatives for roughly the same money as this 27-inch screen. One thing is for sure - the Series 9 will have to be life-alteringly good.

So what do get? First up is the PLS panel. It's the same type used in the Samsung Series 8 27-inch screen, which is effectively Samsung's take on IPS technology with one or two twists. So far, so good.

Next up is what you might call premium industrial design. That includes a proper metal frame, a brushed metal rear enclosure panel, soft touch controls and a general lifestyle vibe. The stand is height and tilt adjustable, which puts it somewhere between tilt-only cheapo monitors and more widely adjustable pro displays.

Then there's connectivity, which not only includes HDMI, DVI and DisplayPort, but also MHL or Mobile High-Definition Link, which allows you to hook up Android phones. And that's about it. There's no integrated TV tuner or app support. This is a pure PC monitor.

The Series 9 will therefore live or die on a combination of image quality and physical desirability. Unfortunately, we're not entirely convinced by either. The chassis and stand are superficially sexy, but we'd rather have more adjustability and easier access to the inputs. The latter are buried in the base, which makes plugging things in tedious.

Clear as mud

The Series 9's biggest problem involves image quality. That's not a result of a flawed LCD panel - it's lovely. The problem is that it's behind a so-called Crystal Clear Glass panel, so you can't actually see that lovely pannel very well at all.

It's a shame, because Samsung's PLS technology is lovely. You get all the benefits of IPS technology, including superb viewing angles and color accuracy, along with what certainly seems likes improved contrast and definitely involves a smoother, less sparkly anti-glare coating.

Still, the good news is that you can buy the Series 8, which has the same PLS tech, in a more unassuming but actually more ergonomic enclosure, minus the stupid panel cover. Okay, you miss out on the MHL, but it's $485 cheaper, so it's no-brainer.

Technical analysis

IPS stands for In-Plane Switching. PLS stands for Plane to Line Switching. And for the most part, it is indeed a case of I say tomato you say, er, tomato. This analogy isn't really working, but the general gist is that Samsung's PLS technology is very similar to IPS. The biggest difference is that Samsung uses a much nicer, smooth anti glare coating than the horrid, sparkly things you get with most IPS screens. It's a shame it's hidden behind a silly glass cover in the new Series 9.