Posting this in Xu Gang room - Drinking beer, Hong Kong
First of all, we would like to express my appreciation to those who shown interest and care towards this competition. Secondly, we are really sorry for the result of AOCC2008. I'll try to make this short and clear. And this competition we do not have a 'cameraman' with use except for vearn27, so i guess i won't be be writing a full report with pictures like what i used to do. I think vearn has done a good coverage on the event on behalf of us. Thanks.
During the 1st day, we reached Hong Kong in the afternoon. At night we got all the hardware for the competition. We were the 1st team to pick the hardwares and a box of Dry Ice. And every hardwares S/N will be recorded by VR-Zone. So there's definately NO BIAS from the organiser. We get to pick our one hardware for the list, which were 20 sets. We only need to return the HDD to the organiser at 5:30PM, 27th July.
When we get back to our room we started to test the setup. For this competition, me and coolice actually decided to focus on Super Pi 8M and 3DMark01SE. Therefore we focus to find stable FSB. The 1st day, which was 26th July we only focus on testing Super Pi. During our testing for the 1st day, coolice did the main overclocking while i was incharged of tweaking. We managed to do 470 FSB. For the 2nd day, 27th July which was the official testing day, coolice tried very hard to find higher FSB and finally managed to clock it to 490 FSB stable for Super Pi 8M. Our prepa
During our testing from Day 1 to Day 2, we use our own HDD. And we didn't bother to install the VGA driver because we focus on Super Pi. So after going to the AOCC2008 Presentation Talk and dinner in the evening, we started Dry Ice cooling. And because coolice was quite familiar with the setup already, we managed to clock it to 5.2GHz stable for Super Pi 8M in a short time. Before this, Team China and Hong Kong told us they can only run Super Pi 8M stable on Dry Ice with 4.7GHz and 5.0GHz. If i remember when we got the 5.2GHz stable for Super Pi, it was already 1 or 2 AM. So i was really tired and fell asleep with knowing it. Coolice continue to insulate the VGA card.
The next day in the morning, which is the Competition Day, coolice told me that the VGA card got problem with driver once i woke up. So this is where to problem started. We suspect it was our HDD that we used which was a very old benching HDD. So once we got to the competition at 10:00AM we immediately test the VGA card with the given WD VelociRaptor with floppy from VR-Zone. So confirmed our VGA cannot install driver and it was actually DOA. We regret that we didn't fully test our hardwares.
So this is the start of our bad luck. coolice need to redo a new insulation on a replaced VGA. Because the 1st Benching is Super Pi, the organiser provided another extra VGA to all contestants so that we won't need to run LN2 on VGA for Super Pi. So i started to power up our rig and find coldbug. It turned out our coldbug on -135 degree, which was very good compared to other competitors. But our coldboot is on -80 degree. So i tried to clock the CPU with the oc saved profile which is 5.2GHz we achieved on Dry Ice. By right it suppose to boot on LN2 easily but it just couldn't. I tried going for higher/lower vcore, vdimm, multiplier, timing and fsb. But it just wouldn't boot. After coolice done insulation on the VGA, he took over to try. After spending around 15 minutes finally he managed to rectify the problem. The oc profile saved on the BIOS was corrupted.
By the time our setup was ready to go, it was around 1:45PM and the competition will be started on 2:00PM. So we do not have time to find our highest CPU clock with LN2. So we tried very hard to go higher 5.2GHz. Even on 5.2GHz it wasn't as stable as our initial testing on Dry Ice. Finally we stop at 5.35Ghz to run Super Pi 8M when the time was going to finish for the benchmark. It was stable to finish the Pi run. Therefore i took over to apply the tweak. Our tweak has a few steps and during the final stage of the tweak, the rig just turned off and the time was only 10 minutes left. So i asked coolice to just use Copywazza which has a shorter process to try to get a final score. When coolice took over for around 10-20 seconds, our rig turned off again and couldn't be switched on anymore. So there goes our Super Pi 8M which was our main focus in this competition.
So we immediately try our best to boot up the rig. From our experience we supposed it was due to condensation. So we removed the CPU pot and clean it up. Then we try to switch on the rig without LN2 but it still refused to. After wasting around 15 minutes then we thought that the mobo failed on us. So we changed a new mobo but still it couldn't boot up. We exchanged CPU to diagnose which hardware died. But still it wouldn't boot. After than we switch power supply and it booted up with single ram. Then we quickly plug in another ram to start the benchmark. But once we plug in the ram again it couldn't boot up. At the same time TEAM Taiwan has problem with their rig. They are kind enough to test our rams and finally the culprit is one piece of our rams.
By the time we are ready for 3DMark01SE, it was already 45 minutes over. So we tried to run the benchmark with 5.3Ghz without taking anymore risk and it passed. So coolice started to clock our VGA card to 760/1250/1600 just to test stability and it passed. After that, coolice applied VMOD and try to clock higher. Once VMOD voltage adjustment was finished, guess what. . . rivatuner refused to recognise our clock. After spending 10 minutes to rectify it, TEAM China confirmed with us it was the rivatuner compatibility with the driver issue. They faced the same issue during their previouse competition. So exchanging a new HDD is the only way. And we need to apply our tweaks all over again. By the time everything was up and ready, it was 15 minutes left. Once coolice tried to clock the GPU properly and run, our rig died again. This time it was due to condensation.
After another failure on 3DMark01, we was really really down. But we still tried to put up a fight. At least to summit score. So after fixing the condensation problem, we carried on with Vantage Performance. Guess what, Vantage went smoothly as ever. For Vantage when the time was around 15 minutes, our mobo died. We only managed to run it stable on 5.3GHz CPU clockspeed and 850/1275/1700 GPU clockspeed which our score was only good enough for 11th or 12th placing. From what we know, TEAM India without VMOD able to hit 820 core speed for GPU.
So this was what happened during the event. Will edit this post for mistakes/error when i get back tomorrow.
This post has been edited by Sanko: Jul 30 2008, 04:28 AM
Advanced Overclocking Championship 2008 Hong Kong, 1st Day Coverage Updated - Page 2
Jul 30 2008, 04:20 AM
Quote
0.0222sec
0.37
6 queries
GZIP Disabled