Jump to content

ccone

Members
  • Content count

    42
  • Joined

  • Last visited

About ccone

  • Rank
    New Member
  1. One more thing... not sure if you tried this but... To use all 4 DIMM slots you have to disable the CPC 1T setting... (Or set CPC to 2T, nto sure off hand how it is exactly stated in the BIOS). AFAIK, this is true of all NF4 boards. CC
  2. According to Anandtech's mobo review, they didn't report having to change the voltage to get the same timings out of a 4 DIMM setup as a 2 DIMM ... just changed the CPC to 2T. (Of course, they may have just loaded optimized defaults and truested it to make the appropriate changes) http://www.anandtech.com/mb/showdoc.aspx?i=2465&p=12 CC
  3. The reason you were having issues before flashing to 510-2 is that the original BIOS that shipped with your board (probably 310 or 125) did not fully support the Venice and San Diego cores. When I tried running my rig on 310, I would see memory errors all over. A simple flash up to 510-2 and resetting my memory timings fixed everything for me. Granted, I use different memory, but it just shows what a difference a BIOS that fully supports the processor can make. CC
  4. If done properly, BIOS flashing is nothing to be scared of. Just make sure you don't have any memory errors befre you do it, and don't try to flash from within windows. That bootable CD with all the BIOSes on it really is nice. Once flashed, load optimized defaults, and manually set some of the important memory stuff... then get the memory in and flyin. :nod: And before booting into windows with all that memory in and tweaked out, run mem-test to make sure you're still error free. A bad memory config will corrupt a windows install pretty darned easy. C C
  5. Benefits: 1) A working system. 2) A working system. 3) You'll be able to use dual-channel, CPC, Bank interleave. 4) A working system. Seriously though... The 310 BIOS that probably shipped with your MB does not fully support Venice or San Diego cores. You'll get *much* better performance with a BIOS that does. You'll also see far fewer odd problems. *EDIT* OK, I'm a dunce... just noticed your sig... there is no 'probably' to it... you have the 310 BIOS that shipped with the MB Also, just because the install seems to be going, doesn't mean windows will be stable. If you have memory issues, that install will get corrupted and you'll end up having to do it over again. CC
  6. I'm not sure about which exact BIOS will be best for you. One thing to remember is that using a Beta BIOS will void your warranty. I'm not sure whether 612-2 is a beta or not, but I'm sure that 623-2 is *official*. That being said, I am not too concerned about using a beta BIOS (as my siggy indicates I'm using one). I have had no problems with them. The BIOS choice is more dependant on your memory than PSU. Unfortunately, I don't know which BIOSes are best for UTT stuff. I'm sure that if you search the forums here, you'll find that info. Actually, 623-2 would be newer than 612-3 ... those first 3 numbers are the date that the BIOS was put out... 623 = June 23rd. 612 = June 12th. As for a matrix, that would be very helpful, but I haven't seen one. Just search this forum (you could also check the BIOS Factory forum in the Overclocking area). Not sure about which slots will work better for you. That depends partly on the BIOS being used and on the memory. If it works, it works. Using 'AUTO' on voltage can be problematic. I've had better results setting it to what my information tells me is necessary. When looking through the forums you will come across this IMPORTANT piece of info: Using the 4V jumper can cause big problems down the road. I'm not well read on this, but my understanding is that this config causes some excessive heat to develop where that 4V is being brought down to the Voltage you are using. Without some active cooling on the top-right section of the board, this will cause board and/or memory failure over time. I'm not at all certain about this, but look for any/all problems that have been found relating to this setting. DEFINITELY - if you do not plan to set your memory voltage above what is available by default without changing that jumper... don't change it. It only gets you something if enable it AND change the BIOS settings to something over the cutoff. There is a set process to follow to get those additional voltages into the menu. They will not show up if you change over to that other jumper position while the voltage setting is 'AUTO'. Not sure about this... the manual will tell you what those settings are for. Not sure about your memory timings and voltages. The Search function is your friend.. Look for others that use your memory and someone will have some timings up that work. Again... I *strongly* recommend that if you don't need over 3.2V that you don't set the 4V jumper up for the extra V's. 4 sticks *should* run. It can be tricky to get them going this way. You will (at the very least) have to disable the CPC 1T function in the Genie menu in the BIOS. I dunno... whatever works? If you have access to one, you could use a temp probe to see what the memory temps are at when under load. Interesting. It should properly recognize the processor... It'd be cooler if it called it a FX-69 :nod: FYI the 623-2 and 623-3 BIOSes are almost the same... the only difference is the memory tweaks that are there. Any time you see a BIOS that is in the same series 623 in this example, that is the case. Usually there is a XXX-1, XXX-2, and XXX-3 .. each of them is tqwaked for a different kind of memory. CC
  7. 1) Any more specs on your PSU? What are the 12V rails rated? (might have only one 12V rail... but whether 1 or 2 are there, needs to total 27A for non-SLI setup.) 2) Since you have not flashed BIOS, I would start there. Take out 1 memory DIMM, leaving one in slot 2... flash to the latest BIOS, then load optimized defaults and make any other config changes you need. I suggest flashing with only 1 DIMM because I suspect you are experiencing memory errors and using 1 DIMM puts you in single channel mode which is much more stable. This is only until you get a BIOS in there that fully supports your Venice core. 3) When that is done, enable MemTest in the Genie menu in the BIOS, put your other DIMM back in and power up... it will enter Memtest and start testing your memory. Make sure you get clean passes through all tests (preferably for several hours). If it shows errors in any test, you will need to tweak your memory timings. Most likely causes of these errors can be too little voltage or some value left at auto which works better when user defined. 4) When you can get clean passes using Memtest, go back into the BIOS, Genie menu and set it back to disabled.... next time you boot, you should be able to install Windoze. CC
  8. In general with any striped RAID setup (RAID 0, and RAID5 for example) ... the more disks the better the performance. (though the differente is probably not noticeable without using a benchmarking tool past the 2nd disk for a RAID 0). The logic is this: In a striped array, the data being written to disk is split up and written to many disks. The writing and reading of the data can be done simultaneously on each disk... so if 1 disk can write/read 1 MB per second, a single disk will take 1 second to write/read that data... 2 disks would only have to write/read .5 MB each... thus cutting the time required by the array in half. As a note, this only applies to the time that is spent waiting for each drive to handle the data... not the overhead added by the controller, OS, system bus, etc. Bear in mind that a RAID 0 array has no hardware redundancy built in and no method to recover from a failed disk. Even if you have a 10 disk RAID 0 array, if any 1 of those disks fail, the array is destroyed and completely unrecoverable. Obviously, as you add disks, you increase the probability of having one fail, so the bigger your RAID 0 array, the greater risk you run of losing all data on that array. If RAID 5 were an option for any of our integrated controllers, it adds parity information into the array as well. This takes up space (roughly one hard disk's capacity), but allows you to keep running with one disk failed/missing and to replace that disk at any time without losing any data. As a result, RAID 5 arrays must have at least 3 disks and the more the better. Since you lose 1 disks capacity to the parity information, the higher the nuymber of disks, the lower the percent of that total is used for parity. A RAID 5 array handles data reads as fast as a RAID0, but data writes a little slower because when writing it has to compute a parity stripe that can be used to reconstruct the data for a missing disk in the case that one fails. Probably more info than you needed... but It's nto often I can talk about stuff I know (there isn't much that fits that description ) CC
  9. OceanSeas said it already... I thought I'd repeat, since I didn't see any response: There are FOUR power connections to be made: 1) 24-pin main ATX plug. 2) 4-pin 12V power plug right next to the 24-pin plug. 3) Floppy power connector between the CPU and PCI slots 4) HD power connector near the IDE connectors. Look at the green dots on the pitcure in this post: http://www.dfi-street.com/forum/showpost.p...76&postcount=95 I would tend to agree that flashing to the appropriate BIOS will go a long way towards solving your issues. CC
  10. I'll repeat what oliveryochest said. DEFINITELY -- Flash the BIOS up to either 510-2 or 623-2. (Make sure you only have one of those memory DIMMs in, since I think you're getting memory errors with both in. After the BIOS is updated, you should be able to run with both sticks in) I have first hand experience of the difference the BIOS level makes. Upgrading from the 310 to 510-2 took me from instant huge amounts of memory errors in mem-test to flawless. Also: In the BIOS, check the settings being used. Make sure you set the memory voltage up to 2.8V or 2.9V. Also locate the recommended timings for that OCZ Gold. Leaving settings on 'Auto' is not always the most stable. I"m not sure about the timings I'm using, but mine is working fine. CC
  11. Hmmm..... W = A * V .... sooooo.... using simple math (not my strong suit).... Each GFX card is reported to use just over 100W on the 12V rail (actually, that may have been the 7800's that I saw that number for, aparently, 6800 Ultra's are higher)... 100 / 12 = ~8.5 so we're using 17A to drive the video. Each core in the X2 takes about 60W ... 5A each... so without touching any of the other components, we're sucking down 27A (out of the 33) on your 12V rail. Not sure what other devices take that same voltage... chipset? other PCI Devices? HDisk? I know that you've pretty much decided it is memory and not power... but a lot of times insufficient power results look alot like memory issues. If you have a really beefy PSU layin around, you could try it. Worst that can happen is it doesn't work. CC
  12. Reformatted, eh? yeah.. that cleans everything up. In your multimeter testing was that when running a demanding 3D game? or just at a windoze desktop? If not when running something demanding, you'll see a much higher load when gaming (of course, you probably knew that)... Wait... I might be able to find the theoretical amp draw on all those components... I'll check on that. One question I have not seen asked/answered... what BIOS are you using? I figure that since you ran at one point, you have flashed to 510 or later (preferably 623 with that X2). Rubbing forehead... hrmph back to my usual state *stumped* cc
  13. After uninstalling the chipset/video drivers... have you run a driver cleaner? I'm thinkin that maybe the BSOD's you are seeing are either a result of residual files and registry settings from the initial failure, or perhapes some windoze corruption.... Bringing me to what could have failed you in the first place... runnin a 2X and 2 7800GTX's off that 520W PSU... just a wild guess... could you be pullin too much juice for it to keep up with? Since we know the memory is fine and was for that initial test run... I'm not sure what kind of Amperage is taken by each GPU... but I'm sure it is quite a bit... add in the amperage taken by each of those cores... just a shot. CC
  14. ccone

    First time builder - LOST

    You have the 310 bios then... That is probably at least a part of your problem. Flash up to one of the 510 (beta) series or 623 (official) series of BIOSes ... there are 3 of each and which one depends on the memory... unfortunately, I'm not sure which in the series would be best for that Winbond UTT :confused: ... I'm using 510-2 as you can see in my siggy, and it works great for me... AG swears by it for general use too, I think... but that is a place to start. Got side-tracked there.... the 310 does not fully support Venice/San Diego cores, I think that may be what you're running into. Good luck CC
  15. As for the performance degradation of using an 8X PCI-E slot instead of 16X .. To my knowledge, no one here has tested with the new 7800's .. but the 6800 cards were not using all the bandwidth available even when restricted to 8X. This means that with the 6800 series at lease we are not seeing any loss of performance using the 2nd slot (as long as you put the jumpers to SLI mode) CC
×