rsh2k1

Gah... So I have to reinstall everything I assume? Grrr. Advice time: Would you recommend just getting a smaller IDE or SATA drive for booting into Windows, maybe some key apps (Prime, etc.), and then setting up the RAID array for data files and such? I've read somewhere that nvidia RAID doesn't really like mixing RAID with a standalone HD, but that could be just wrong. -r

Hmm... I think I changed the PSU (to the PowerStream) and the heatsink on the CPU (to XP-90) but beyond that... oh yeah... I installed a wireless card (a belkin pre-n) to move the computer to the bedroom. Could something that small really corrupt a RAID install? -r

I think I'm having a problem with my RAID array under Windows.... Set some background: I have the drives in my sig running in RAID-1 (mirror) with Windows XP installed on it as the boot drive. I followed the directions from Angry's Guide pretty much to the letter and had successfully installed Windows on my RAID array. While I was overclocking the rig, I was always able to see only just the C: drive, which is good. Recently, after installing some games, continuing to test, etc., I noticed that I now have a C: and a F: drive.... I went into each one to see what was going on, and it appears to me that Windows somehow "lost" the RAID array. That is, all the files I downloaded and installed while overclocking -- such as OCCT or Prime95 or whatever -- are appearing on both drives, while the newer files such as games are not. They're appearing only on the C: drive. When I boot, it goes through the normal sequence, recognizes the RAID array, and says it's healthy, and so forth. It's only in Windows that I am now able to see both drives, and worse, they're not mirroring. This is unacceptable, since the whole point of setting up RAID-1 was so I would have a fallback in case I lost one drive. I've done a search and found this thread but it didn't really go into any detail. There was another thread mentioning something about nvidia's RAID controller software being bad, but I didn't see much resolution there either. Any help or suggestions would be appreciated. Thanks, -r

Is http://www.silentpcreview.com/Sections+ind...e-artid-84.html this what we're talking about? -r

Okay, so my rig is all set up now; 12 hours Prime95 stable, loaded FEAR and HL2 and played through a few rounds, checked MBM5 for temps and everything seems good. Here's the problem, however. The rig is a bit on the loud side, enough so that I don't think I can keep this in the bedroom, which is where I wanted to keep it. Granted, I have a lot of fans going: 1x92mm Vantec Stealth on the XP-90 cooling the CPU 2x92mm Vantec Stealth on the front of the case 1x120mm Antec TriCool fan on the rear of the case 1x?? fan on the chipset 1x?? fan on the 7800 GT 2x?? (80mm?) fans on the OCZ PowerStream 520W. One of the main problems is that 120mm TriCool -- it only has three settings: low/med/high. I've tried it on low, and the case heated up a bit too fast for my tastes, so I have it at medium. Of course, to change the setting involves opening up the case, etc. which is impractical. I have become aware of such things as fan controllers to manage precisely this sort of thing. Anyone have any recommendations for a fan controller -- or some other solution -- to deal with fan noise? What I'd like to be able to do is to leave the computer running 24x7 if need be, as there are anti-virus updates, scheduled programs, and such I'd like to run at 3am, but have the fan speeds and therefore the noise turned down while the comp is idle (i.e., overnight). The ability to see case and CPU temperatures while playing a full-screen game is a bonus, but not all that required. Thanks, -r

The only thing I'd like to comment/add as a fellow DFI and system building newb is to make sure that you get components you could overclock if the urge strikes you. I felt the same way as you do, that I just wanted to build a stock system that was stable. But the fact is that the DFI mobo is an overclockers mobo, and the AMD is an overclocker's CPU... the urge to bump your performance up by 10% or 15% without a whole lot of pain (e.g., watercooling, etc.) will strike you. I tried to save money by going with value RAM -- I did okay, but I wish I had the foresight to just buy good overclocking RAM to begin with, as the bug did bite me, and I'm forced to make do with what I have. I also tried to save money on the PSU initially, and use stock AMD cooling -- guess what? I've upgraded both of them. From that standpoint, it might not be a bad idea just to get 2x1GB DRAM that is friendly to overclocking, as well as a rock-solid PSU, and some aftermarket cooling solutions just in case. -r

Ah, yes, forgot to mention that. Temps are pretty darn low actually. I'm running around 1.6v Vcore (1.45 +110% -- shows as around 1.59 on MBM), 1.3v LDT, and 2.65v DRAM; temps are around 36-38C idle on the CPU, and around 48-50C while running StressPrime. [Edit: I had forgotten to enclose the case completely when I ran the previous measurements.] PWMIC fluctuates between 48 and 52C, while the NF4 Chipset is also around 48 to 53C or so. My only reason for concern is that the CPU fan (a Vantec Stealth 92mm) hits 10K+ RPM's when the CPU is at full load, but... that doesn't seem so bad. So I'm pretty sure I can push the CPU more; I suspect my RAM just won't take it. The system is stable at 233Mhz (DDR 466) but on any divider, it won't boot and won't boot at 235Mhz. I could theoretically try to finetune memory settings, push it more volts, etc. but that just doesn't seem worth it with value RAM, when I'm getting 2.56Ghz. I'll save up for a bit and invest in OC'ing RAM and go at it again. I have the sort of "pre-OC'ed" eVGA 7800GT, and this being my first time with overclocking at all, I'll probably wait before tackling things like that, hehe. -r

Well, I managed to get to 2568Mhz (233x11) running the DRAM at 2.5-3-3-8 1T at 1:1 (233mhz, or 466DDR). Passed overnight Memtest (9 hours or so), passed 32m SuperPi, and passed 30 min OCCT, and 3dMark05. Currently chewing through StressPrime, and I plan on letting it run all day. For whatever reason, I can't seem to get anything to work on x10 multi. If it proves stable, I'm not going to touch the rig until I get some new RAM. I'm pretty happy with the performance I'm getting out of ValueSelect memory. The only thing that bugs me a bit is my 3dMark score: I'm seeing others on ORB with similar config getting over 10k 3dMark, while I'm looking at 7600 or so. I assume that difference has to do with the RAM, but just in case, I'll update the Nvidia Forceware drivers to the latest and see if that offers any improvements. Many thanks to everyone for their help. -r

Thanks everyone; just an update. I began trying to find the max Mhz for my DRAM... and more oddness.... I followed the Overclocking Guide very specifically (lowered CPU multiplier to x7, etc.), then put the RAM on 1:1 (200 Mhz). 220Mhz --> no problem, passed Memtest. 230Mhz --> got into Memtest no problem (!) This is significant because when I was trying to find my CPU max, I couldn't get past the POST where normally, I would get the "DMI Pool" message. I got past it this time, and now I'm happily running 25 cycles of Memtest #5 (as a quick test). There was a thread on DFI Street that I can't find anymore where someone mentioned possible issues with Corsair RAM and dividers -- that Corsair OC's higher running at 1:1 than using any dividers (which again, seems odd)... Could that have been an issue when I was trying to find my max mhz for the CPU? Keep in mind that I don't have any dividers on this, and if there was a problem component in my rig now, it's the RAM. Almost seems to me that the CPU is holding me back, and that... just seems weird. If I could get to around 2500-2600 stable using ValueSelect RAM, even if I have to relax timings to something like 3-4-4-8, I'd be more than happy for now. Oh, I'm running 3x LDT now, and I'll try upping the LDT voltage a step as well if I see instability or problems. -r

Both of these sticks have passed Memtest for over 8hours each and in combo, at least at stock timings, running 1:1. I guess if 2400 Mhz is most I can expect with my crappy RAM, that's fine. I can buy new RAM later and try again. What's really weird is that I can pass 10 hours of Prime running 220x11 with those odd Ram settings (to find CPU max) but I can't boot into Windows if I go 240x10, which is supposedly lower speed (2420 vs. 2400). All around strange. -r

I'm following the Overclocking Guide to the letter... so I kept the multi at x11. Tony, since I've put the Ram Ratio at 100Mhz (1/02) as per the guide, I doubt it's the RAM... I know my RAM is (a) Corsair, and (B) ValueSelect -- I'll be upgrading that as soon as possible. My LDT Multi is set at 2.5x (again, as per the Guide). I kept the voltage on those at stock: 1.20v on LDT and 2.6v on the DRAM (although MBM tells me they're at 1.15v LDT and 2.58 DRAM). Keep in mind that I'm just trying to find the max of my CPU. It might be that the max Mhz on the 3500+ is 2400 Mhz... that would be really disappointing to say the least... when a 3200+ can be pushed to 2700. -r

Hi everyone, So I'm dipping my toes into the OC'ing game. I've read, then re-read the Overclocking guide on DFI, and also read a number of other beginner's guides (e.g., the one on i4memory.com). I started the journey earlier tonight. After a scary episode when my CPU fan stopped working, and the MBM5 and my BIOS both said the CPU and Chipset temps were -127 C (that's minus 127), I followed the instructions on the Overclocking guide to try and find my CPU's Max mhz. Now... if I understand the guide correctly, the point of this is to minimize the role that my (admittedly crappy) RAM has on the CPU, right? By putting the memory on a 1/02 divider, lowering LDT, etc., I should be able to just OC the heck out of my CPU to see how far it would go, yes? Well, I got it to 220 HTT at x11 multiple, or 2420 Mhz, with all those lowered settings. That's only a 200 Mhz OC, or less than 10% from the stock 2200Mhz. I'm giving the vCore ~1.6volts (1.5 + 110%) and haven't touched the rest, but the computer won't POST at 225 HTT and won't boot into Windows at 222 HTT. And yet... at 220 HTT, it passes SuperPi, OCCT, and 3DMark03 no problem. Seeing as how I'm reading about people getting a 3000+ to 2500 Mhz and a 3200+ to 2700 Mhz, I have to wonder... does this sound right to you? I know each CPU is different and all, but I'm just trying to find the max Mhz of mine, and for a 2200 stock, 2420 max overclock seems well... low. Is there something I'm missing here? Or is this the best I can expect from this CPU? -r

Thanks to everyone on DFI-street, I now have a stable (24+ hours of StressPrime), home-built computer. Of course, since I bought a DFI mobo and a 3500+ Venice... I'm tempted to dabble in the arcane arts of overclocking. I've been looking for some good beginner's guides on AMD overclocking and on overclocking in general. The DFI Overclocking Guide is really great, but it's a bit of advanced juju for this noob. Anyone have a suggestion for a good beginner's guide to overclocking that wouldn't harm my wanting to move up the food chain with DFI-specific guides later on? When I say "Beginner", I mean literally, someone who doesn't know how to find things like the CPU multiplier. -r

Okay, I d/l'ed Tmod's CD -- fantastic tool that. I ran Memtest once again, just in case, but only for about an hour, since both memory sticks have been tested previously. Passed. Ran WD disk tool; passed. Flashed the 623-3 BIOS. Now, I'm getting a CMOS checksum error -- default loaded. Should I just clear the CMOS and try again? -r

Well, like I said, I've been waiting to flash the BIOS until something occurs that makes it seem necessary. Guess this is that thing which makes it necessary. Offhand, I didn't see anything in the descriptions of the BIOS that suggested it would fix a nvatabus.sys problem or any of the other BSOD/reboot issues I've been having. What major things got fixed in the 623-3 BIOS? -rs