MSI European Overclocking Challenge 2008 Coverage
The UK Team Effort
The UK Team
Barron and Busta's bid unfortunately didn't start too well. With a shortage of the hardware in the UK, they had only received a board 2 days before they were due to leave for the event, and that one was faulty. Still, with Barron's experience at I34 the weekend before, they proceeded with vigour when they received a board and GPU at the start of the event.
To start with they set up the hardware as it was supplied. With the OCZ vendetta cooling just to get a feel for the hardware. Testing the limits of the chip, the team managed to crank around 600FSB from the little E8400 with a 6 multiplier and air cooling alone.
The next step was to install Busta's Dragon Evo 2.4 Dice/LN2 pot to the board. Starting out with insulating the board using liquid electrical tape, which I couldn't help think resembled tar somewhat. With the liquid tape dry, a layer of Armaflex was placed on the back of the board and a specially shaped one was placed around the CPU socket. All in an attempt to prevent the hardware meeting its maker via condensation caused by the extremely low temperatures the equipment was about to endure.
Once the socket was wrapped up warm and the pot installed it was time to start with the extreme cooling. This was my first experience with pot cooling, and I was quite surprised to find that even without either of the substances available to us inside, the chip still booted. But that was soon to change as the UK team got things started with a dry ice run before moving onto the LN2. This brought the temperature down to somewhere on the wrong side of far too low for the first few runs and achieved a bechable overclock of around 5Ghz. A few SuperPi 1M tests were run, pushing out scores of circa 9 seconds which was really quite admirable.
The next step was to attach the Mouse pot to the MSI 9600GT. This meant another lavish coating of liquid tape on both the top and bottom of the card. Then the pot simply bolted on like most other coolers. The initial concerns about the sheer weight of the pot snapping the card were relieved when the guys found that the pot in fact supported itself on the free PCI slots below the PCI-E the card was installed in. Somewhere along the line, however, no probe had been attached to the Mouse pot before it was attached to the card, and therefore measuring the temperature was impossible. Then, towards the end of day one there was a slight ice build up on the back of the card.
Day two began with high hopes, with the hardware starting first time at previously saved settings. A few teams had to kick-start their setup's to get them to boot after sitting overnight. Again though, misfortune struck and something bit the dust. At first thinking it was the graphics card, the team brought in their replacement and tried to get the PC to post, but it was a no-go. The problem turned out to be the motherboard being stubborn. And so that, and the weak E8400 were replaced with only a few hours left to bench.
Luckily the new chip and board combo proved to be an improvement on the old and the team wasted no time in pushing it to its upmost limit. The chip could not only bench at higher clock speeds but was also cold bugged at lower temperature. Meaning in short that there was less care needed when adding the LN2 to the pot.
Remounting the card with the pot Barron and Busta then went on to attack 3Dmark05 with full force. benching in at around 5.2ghz on the CPU and 850mhz on the GPU core. Obtaining the result in 3Dmark that they got proved quite tricky once again due to a cold bug, however this time it was quite a bit higher than on the CPU at around -30. The GPU would be fine while under load with the probe reading less than -30, but as soon as the chip came off load. It would simply lock up the system, causing the hard earned score to be lost as the machine reset.
Overall the UK team did very well however, just how well is revealed on the next page...
Most Recent Comments