Reiter rig parts ahoy and a bit of a set back

So today I picked up the motherboard and memory for the Reiter rig. Gotta say, wasn't expecting all the original paperwork (and CD) and most of the original cables to come with it (even a spare 775 pin cover). The eBay seller even saved the case stickers for the Gigabyte motherboard. Gotta say, if it powers up and functions great, the seller is definitely going to get a 5 star rating for this. The package even came from Estonia, and there's what looks to be no damage at all during shipping (which is excellent). Also the seller  included a double fan connector (which I guess he was using for his CPU heat-sink fans). The north bridge heat-sink looks to have faded a bit (it used to be more of a golden yellow metallic color). Not really something to be bothered about, just something I'd thought to mention. The board I received, by the way, is a Gigabyte GA-965G-DS3 (in case you didn't read up on it in a previous post). It is a bit older than the P5Q in terms of generation of chipset (uses the 965G chipset which is a few generations older than the P45), but it is still a quality board with some overclocking potential (not as much at the P5Q though). The RAM the seller included was 2x2GB of Crucial Rendition RAM (which is their budget RAM without any heat-sinks on the outside). As I won't be overclocking the ram too much, it shouldn't really be a problem that there's no heat-sinks on the outside, just something that I aesthetically prefer on systems or if I need a bit of headroom on the overclocking side. Quite impressed thus far, lets see how it does in the rig.

Contents in the Gigabyte GA-965G-DS3 box

All of the included contents of the Gigabyte GA-965G-DS3 box

The Gigabyte GA-965G-DS3 motherboard

I unwrapped the Core 2 Duo E6550 CPU that was to be installed into the system, took the pin cover off of the Gigabyte motherboard and slotted the CPU in to place. Then I noticed that the north bridge heat-sink seemed rather loose. I looked at how long the hold down pins were and was astounded at how little spring tension they were getting. I pinched the two pins out, and took off the heatsink and compared the pins with pins from an old ASUS north bridge heat-sink to find that the ASUS ones were very short by comparison. Seeing this, I cleaned off the old thermal paste from the north bridge heat-sink and the chip itself, put some Phanteks PH-NDC thermal paste on (it's what I had handy at the time), and reinstalled the heat-sink with the smaller hold-down pins. It fit way more securely and believe that it's a significant improvement over (what should be) the stock pins. Finally, I cleaned both the heat-sink and CPU surfaces, put the thermal paste on and mounted the heat-sink fan. I fitted the motherboard into the case, plugged everything up, powered the system on, and.... nothing. OK, I thought to myself, must be a BIOS update required or something along those lines. Go to the motherboard's page on Gigabyte's website, look over everything and then realize I made a rather rookie mistake. The motherboard doesn't support the E6550 (as it doesn't support 1333+ FSB based CPUs). Well, I'm a bit disappointed in myself for not looking further into the FSB support prior to purchasing the E6550. I assumed that because the E6500 and E6600 functioned OK in the Gigabyte board, a E6550 would be OK as well, but unfortunately the FSB being different was an immediate problem. According to some sources though, the E6550 is compatible with Revision 3.3 of the board, so that's also a bit odd. OK, well I can rectify me totally spacing out on which CPU I was getting. I decided to swap the CPU's between the Reiter rig and the Broadway rig and that should fix all my headaches as the P5Q does support the E6550 and the Gigabyte board supports the E6300 (as the 6300 is 1066 FSB). Kinda upsetting considering I spent a few hours putting the wires in nicely etc. only to have to take apart some of it to redo this big mess I've made. Luckily I just need to take off the heat-sinks, so don't need to remove the motherboards again to put the backplates on (as they will not be adjusted) and don't need to wait for more computer parts in the mail (aside from the Mushkin RAM).

The motherboard with the E6550 installed

Pin size difference (ASUS on the left/Gigabyte on the right)

Bottom of the north bridge heat-sink (prior to cleaning old thermal paste)

North bridge chip (prior to cleaning old thermal paste)

The heat-sink reinstalled with shorter pins

Putting the Phanteks PH-NDC thermal compound on the E6550

The Zalman CNPS10X Flex installed with both Silenx fans

The motherboard installed into the Reiter rig

The backside of the case complete with all wiring for the system

After adjusting the CPUs and testing for functionality on both boards, got them both working. Because I had to redo the CPUs I also had to redo the thermal paste on the Broadway rig so I used Phanteks PH-NDC thermal paste as well as using it on the Reiter rig. Unfortunately, because of the limit of the length of the fan cables, the Reiter rig didn't turn out quite as tidy in the motherboard area as the Broadway rig. However, as there is no window on the side panel, it won't be noticeable unless you peek through the small opening for the fan vents. I also noticed that one of the Silenx fans on the CNPS10X Flex was making horrible vibration noise, so I removed it from the system (yet another one that failed on me). What was even more odd is that it was the newer of the two fans. The older one is still working just fine (not that I plan to rely on it forever). Having a second fan may be nice for a push/pull configuration, but I would rather have the system be quiet and me missing one fan than fairly loud and having the extra fan. Even without the second fan on the back of the heat-sink, the rear case fan is almost perfectly aligned to help suck out any hot air that is being blown out of the CNPS10X Flex. Out of idle curiosity, I removed the other fan as well to see how quiet the system was, and dang I would not really even notice that it was on if it weren't for the LEDs (with the side Silenx fan also removed). But, that being said, I would rather have a bit higher decibel than whisper quiet to allow for some more active cooling on the Zalman and inside the case to keep the temperatures down. I also changed the SATA cable to a black one to make the wires look a bit less messy due to them being all dark cables now instead of having the one red SATA cable sticking out.

The motherboard reinstalled after E6300 swap

The final look of the back of the case

The final look of the outside of the system

On to overclocking now. I don't expect that it will overclock as nicely as the P5Q (because of less power phases and a lower tolerance FSB) but lets see how good we can get it to perform. As a starter, just to see if it would work right away no problems, I cranked the FSB up to 400, set the RAM multiplier to 2.0, left the PCI-E MHz as stock, and adjusted the voltages up a bit on RAM, FSB, MCH, and CPU to appropriate settings. Unfortunately, it was too unstable, so I went down to 333 to get a baseline to make sure that it would atleast be capable of that. That booted just fine. I then started cranking it up till I hit 360 on the FSB (making it 2520MHz with the 7x CPU multiplier). It was still booting at this point, but it seemed like anything higher the system would be uncooperative and fail booting to the BIOS. Still, going from 1.86GHz to 2.52GHz is a hefty overclock (approx 660MHz or 35% overclock) is a significant improvement. I'm not sure whether the Option 1 or Option 2 is really doing anything on the BIOS though (it's apparently supposed to adjust some timings or make it more compatible with the system) but as far as I can see, it has no significant improvement in terms of getting a higher clock. I would have loved to get it to hit 2.8GHz like it was performing on the P5Q, but the FSB limitation on the motherboard seems to be preventing it, which is a shame because had I known that I was going to have a FSB issue I would have considered a different second CPU instead of the E6550 that I originally purchased. Maybe I can scour eBay and find a E6600 for about as cheap as the E6550 and put that into it and possibly just use the E6300 for another build in the future (to allow for a higher multiplier on the CPU to get higher clocks). In the meantime though, I did a Memtest86+ full test on the motherboard's current BIOS settings to see if it was 100% stable. As expected, it did just fine under those settings as the RAM was actually underclocked. So next I decided to push up the RAM multiplier to 2.5 (to make the RAM @ 900MHz). Booted up OK so time to do another Memtest86+ and see how it does. It's a shame that the BIOS doesn't allow me to manually adjust the timings of the RAM, but considering that it's an older generation board and a bit more on the budget side of their boards, I'm not too surprised by it. With the 900MHz clock on the RAM, the timings auto-adjusted to 6/8/8/20, and when I had it set to 720MHz the timings were being adjusted to 6/6/6/17. Not sure why Crucial decided to use those clocks rather than 5/5/5/15 like most standard DDR2-800 RAM, but oh well. At least the timings allow for a bit higher RAM multiplier because of the limited FSB so that the RAM can be speedy even though the CPU is only at 2.52GHz.

BIOS settings that seem to be stable

Passed Memtest86+ testing

Once I had verified that the RAM was stable via Memtest86+ stating that it passed the full cycle of tests, I then realized that the HDD was not showing up during overclocking at all, and then when at stock clocks it showed only 33MB? So I plugged it up to my computer, and sure enough, the drive seems to be DOA (shows the exact same amount of space available). Not sure what happened because a week ago, I formatted the drive to 1TB and everything looked just fine. Going to have to RMA it through NCIX and hope that everything gets sorted out. Seems like as I progress through getting these two rigs up and running, I seem to be running into hardware problems where I'd least expect it. Hopefully when I get the Mushkin RAM from Newegg the Broadway system will be at 100% and ready to go or else I may end up having a mental breakdown.