I am having some problems with my new computer set-up. The hardware I have is as follows: Intel Core i5 750 with boxed cooler (at stock speeds) MSI GD65 motherboard (Currently on BIOS 1.20 - the one it came with) 8Gb (4 x 2Gb) Crucial DD3-10600 RAM, at stock 9-9-9-24 settings Corsair HX520W Power supply (this is connected to the MB with the 24 and 8-pin ATX connectors) Powercolor Radeon HD4850 PCS 512mb (6-pin PCIE connector is correctly inserted) Hitachi 7K1000.B 1Tb harddrive (SATA mode is set as 'AHCI') Samsung SATA DVDRW drive The machine POSTs fine. The first two things I did were: - Run MemTest86+ for 6 hours straight, which resulted in no errors - Run the Hitachi Drive Fitness test on 'exercise' mode which does a full scan and 30-minutes of intensive read-writes. Again, no errors or problems (although it refused to run unless I set the SATA mode back to 'IDE'). As I'm still impatiently waiting for Windows 7 on/after 22 October (and my SSD to arrive) I decided to install the Windows 7 Release Candidate - the 64-bit version to be precise. It's here that I've run into problems. Basically, once the installer gets to the 'expanding files' stage, the computer will either freeze (stuck mouse pointer, caps/numlock lights unresponsive) or reboot. The furthest I've managed to get is 54%. Things I have tried (all resulting in no difference): - Changing from AHCI to IDE mode in the BIOS (as this had an effect on the drive fitness test) - Disabling BIOS temperature-controlled fans so always at 100% (by the way, my BIOS CPU temps are about 30'C - even immediately after an involuntary reboot). - Disabling EIST/C1E and Turbo Mode - Disabling the 'phase switching' capabilities of the board so all phases are always active - Various combinations of the above Things I intend to try: - A 'known good' GeForce 7900GTO from my old PC - Flashing to latest BIOS (think that's 1.30, but Bit-Tech review mentions 1.41). - Using the 4-pin instead of 8-pin ATX connection (the board came with plastic cover over 4 of the holes, which I removed.) - A different PSU (although I only have 400W Corsair) - Removing 4Gb of the RAM so only 2 slots populated - An ever lower power graphics card (though all I have is a PCI(!) Radeon 9200SE) - Installing Windows 7 32-bit (although don't see why this should make a difference) - Installing Windows XP (I don't have a copy of Vista) The only other lock-up I've experienced was with the 'lite linux' CD that the board came with (the hilariously named 'MSI Winki'). This happened when it loaded the MSI website in Firefox 3.0 that contained some Flash. I put this down to driver support, but perhaps not. I'm concerned that this is a power-related issue when the CPU is given something more taxing than a memory or HD test to perform, but really feel my PSU should be enough as it's a high quality model with a decent 12V amperage. For the past 3 years it's happily powered my 7900GTO, Core2 E6600, and a collection of drives & USB peripherals. Would therefore like to rule out other possibilities before dropping £100 on a power-hungry 800W PSU. Any help or comments greatly appreciated. Know this i5 and P55 gear is all still quite new, so did expect a few teething problems, but would like to get it sorted in the next few weeks before my Windows 7 DVD arrives from Amazon.
have you tried using a different disc looks like a corrupt windows 7 disc. seen alot of people with this issue try buring the disc again at the slowest speed possible.
Yeah ive had this on Vista and XP over the years, it locks up, reboots or displays an error and you assume hardware but 9 out of 10 times it turned out to be bad media, worth a try like everyone has said bud
I hope you're right guys I did have to Torrent the ISO, but did an MD5 checksum on it and compared the file size to make sure it was legit. Did tick the 'verify' button on Nero and can't hear my DVD drive 'struggling'. Funnily enough the first one I burned refused to verify. If my ISO's bad I know my 32-bit copy works and have a spare RC key. Failing that, can anyone recommend a good 'stability test' program that runs from a bootable CD? Want to stress the CPU and see what happens.
OK - an update. Decided the best way around the corrupt media issue was to do a USB-stick install by extracting the files from the ISO. Tried this and it got to 94% before it crashed. The next two attempts didn't get anywhere near that far. Did some internet reading and found some people had issues with the installer when >2Gb RAM was installed, so dropped down to one 2Gb stick of DDR3 and tried again. This time it installed and got all the way to the desktop. Thinking I'd cracked it I decided to calculate the Windows Experience Index for my shiny new hardware. This locked-up halfway through and after I reset, seemed to have lost the file "explorer.exe" as the task bar was in 'Windows 95' mode, and I couldn't open 'computer' or 'control panel'. Something clearly isn't right. Ran out of time and had to go to work. When I get home I shall try to reinstall, then load all drivers, latest BIOS and Windows updates before trying out Prime95, CPU burn, Futuremark, etc...
isnt that >2GB issue a myth ? i ve installed windows 7 on lots of machines and had no issues with 4GB 6GB 3GB makes no diff.
Yeah - I suspect it reflects the incidence of people with undiagnosed hardware/RAM faults sadly; certainly if my post-install experience is anything to go by :-( Few things more irritating than an unstable PC! When I built my Core 2 Duo system 3 years ago it took me a while (and several bios revisions) to get the RAM voltages and god-awful jmicron controller issues solved. Didn't skip a beat in 3yrs once sorted though - will just teach me to buy on launch day...
Well, given that I did manage to install it over USB, extract all the files from the ISO file to my USB key, and the MD5 checksum worked I think it's more likely that I have a hardware fault. The only other OS I've booted (comedically named MSI 'Winki') also locked up on me. Going to fiddle with things like RAM timings and voltages when I get home from work later.
An update: - Still with 2Gb RAM only I tweaked the timing to 2T from 1T, fixed the RAM voltage at 1.506V (these DIMMS are rated at 1.5V) and disabled spread spectrum, EIST and Turbo Mode. - With this setup I was able to reinstall Win 7, all the updates, and start downloading and installing drivers - then it locked up again. - Upped memory voltage to 1.55V and carried on. Also decided to install the 1.30 BIOS. After installing drivers I successfully calculated my Windows Experience index and ran Futuremark Vantage to completion (suspect score of P7800 is a little low, but at least it ran) - Installed the remaining 6Gb of RAM and was able to repeat the Windows Experience and Futuremark tests on the same settings with 8 Gb. - Tried to run IntelBurnTest v2.3 and it crashed after just 10 seconds or so. - Upped the memory voltage again to 1.60V. System will now complete IntelBurnTest v2.3 fine. Currently on my fifth pass at Maximum with no probs. Temps peaked at 73'C per Real Temp 3.00 and I'm showing a speed of just under 39GFlops (no idea if that's good). Fingers crossed I've narrowed it down to memory voltage. Looking online, others with Gigabyte boards have had problems with the same RAM kit as me - to the point of it not even posting. About to reboot and try re-enabling some of the features, like EIST and Turbo mode.
Agreed. I might have to do some forensic work on the chips. The part number is definitely for a 1.5V kit. The 3-channel set of the same dimms was reviewed on Toms Hardware, and was the main reason I bought it - had heard of problems with Corsair and OCZ. Have had bad experiences with Kingston in the past. Know my Crucial is 'cheap' RAM, but as I'm not overclocking it should be more than enough. Latest update - I re-enabled EIST, Turbo, and all the clever 'phase switching' stuff on the MSI board, and with the full 8Gb install. CPU burn test crashed out after about 18-minutes, but that really is a torture test. It also crashed with a polite error box and didn't take Windows with it - left running overnight and desktop still worked. Might try giving it a few extra millivolts. Assume if I keep the voltage below 1.65V I'm not doing any harm? Can run 3DMark 06 a treat also. My last build crashed out in that due to graphics issues on an ATI card so still have nightmares about that. This time my stability probs are definitely CPU/RAM-related.
nah ive got my modules @ 1.66V cause my x58 UD5 doesnt do 1.65V only 1.64V. sould be perfectly fine. 1 question is though that my memory modules are DDR1600 @ 1.65V but DDR1333 @ 1.5V just wondering if your modules could be DDR1333 @ 1.65V and maybe DDR1066 at 1.5V ? most likely not but just a thought
It's a good suggestion, but apparently it's DDR1333 @ 1.5V. These are they: http://www.crucial.com/uk/store/partspecs.aspx?IMODULE=CT2KIT25664BA1339 The triple-channel kit was reviewed on Tom's here: http://www.tomshardware.com/reviews/6gb-ddr3-x58,2372-3.html "Rated at DDR3-1333 CAS 9, Crucial’s most impressive specification is its industry-standard 1.50 V operating voltage." That said, what do you buy from Crucial if you're running an AMD or Intel X48 set-up that takes the higher-voltage DDR3? These are the only standard 2x2Gb kit they do. Is my tweaking the RAM voltage compensating for anything else that might be on the low side? There seem to be so many different voltages to tweak in the BIOS compared to my old Abit AB9-pro. If it's stable at 1.65V though I'm not touching anything!
http://www.ultimatebootcd.com/ Boots from the CD and will let you stress and test pretty much every major component within the PC - worth a look if that's what you want to do
Yep - passed 6 hours worth of tests without a single error on 'Auto' (1.50V), but the system fell over installing Windows. Crazy. Currently set at 1.60V which is giving pretty good stability, with all my other changes reversed (although memory forced to run at 2T instead of 1T). I have Memtest86+ v4.00 running whilst I am at work, so should get a nice solid 9hours worth of results. Going to try 1.65V later to see if I can get CPUBurn to complete a few hour's worth of tests. Might try some Prime95 whilst I'm at it as know opinion varies. At 1.5V CPU Burn locked the system in seconds on a 'standard' test. At 1.6V it manages over 15 minutes at 'Maximum', completing several passes. Even the bit-tech review of the GD65 didn't complete the Prime 95 test, so I guess there are still future improvements to be made for this board. Hoping the next few BIOS updates iron things out - was the same for my Abit AB9.
With memory at 1.65V the system can now run Intel CPUburn for 10-passes on max without crashing. - which which 8gb RAM is well over 80 minutes. CPU temps get up to about 78'C, but machine is still stable. Memtest86+ 4.00 still checks out fine too. Have been happily installing programs, drivers, games all weekend - Crysis runs on high with AA enabled, and played a few hours of HL2 EPs 1&2 with maxed out settings. Weirdly, Prime95 does produce errors on the torture test (blend mode) on cores 1 & 3 after about 10 minutes, but I have no other visible signs of instability. Not really sure why that might be, but think I'm going to leave well alone since the system is working great besides that. Need to transcode some video, so will see how that goes as a 'real world' test. Impressed by the performance so far under 64-bit Windows 7. Everything is instant. Reckon I'll notice it even more when my SSD arrives.
