Hi Nick, sounds like a plan to me.Confirm memory remap is enabled.I am Sp1 with vista and current with all updates.I am using 182.50 WHQLlet me know what now, I have ordered the 2 new sticks of ocz ddr3 1600mhz, should be here in a day or so. I think on balance best to stick with what I have and get it optimised and set up right, BTW, do you think my card running at 90 degrees under load is not helping? Looking on the net this seems fairly common, but its pretty hot to me!! I was looking at getting one of the new arctic coolers for it, says they will run it at 65 degrees under load.thx, Mark

Good... yes, 90 is OK on the 280 but it is high Mark. Can you please open this tower and run it open (side off) and confirm the same temp on the video is present. What software are you using to pull this temp? It sounds to me like it can

Hi Nick, I used Everest for the temp values, and the side of the case was off already, as I had swapped out the card to get the serial number for possible RMA. In the bios, values for both the parameters you wanted me to look at are expressed in format 0.57x - 0.53x for CPU GTL reference and 0.49x - 0.67x for the other one. I will let you know here as soon as I have the new sticks and slot them in. My case is a silverstone Temjin TJ07BW, its loaded with fans, all the watercooling stuff and triple fan rad is in the bottom compartment. I have all my hardrives in a caddy with fan assisted cooling on it, I run the fan on the GPU at 95% all the time. Thanks again for the time you have put in to helping me out, its much appreciated, rgds, Mark

Hi Nick, I used Everest for the temp values, and the side of the case was off already, as I had swapped out the card to get the serial number for possible RMA. In the bios, values for both the parameters you wanted me to look at are expressed in format 0.57x - 0.53x for CPU GTL reference and 0.49x - 0.67x for the other one. I will let you know here as soon as I have the new sticks and slot them in. My case is a silverstone Temjin TJ07BW, its loaded with fans, all the watercooling stuff and triple fan rad is in the bottom compartment. I have all my hardrives in a caddy with fan assisted cooling on it, I run the fan on the GPU at 95% all the time. Thanks again for the time you have put in to helping me out, its much appreciated, rgds, Mark

Just to be sure Everest is right.. ( I never trust one source especially with newer hardware) use EVGA Precision Toolhttp://www.evga.com/PRecision/If you dont want to sign up to get it there.. get it here insteadhttp://majorgeeks.com/EVGA_Precision_d6021.htmlVerify the temps are indeed what Everest says..And I do not run things like Everest when I use a systems. I keep the footprint very low by using software what handles my card clock, fan speeds and temp monitor at the same time.. RivaTuner is what I use however it can be tricky to set up. Verify the temp of that card with precision toolspeaking of card clocks... are you overclocking your video card? If so I would like you to not do that during software tests of your CPU clock and when/if you do clock that card please be careful about exceeding its ability.. you can see stability problems from that just as easy as you can see them with the system itselfAnd do ot use Everest for CPU temps under load tests... you need to make sure of that and RealTemp or Coretemp will provide that data

Hi Nick, an update for you: memory will arrive today, so, i need to know wether to apply those BIOS settings you posted before or after I install the new sticks. Shall I try blue slots first ?Re card clocks, i have not touched it, it is the factory overclocked card. Re the temp on the card, I put the case side back on and ran EVGA precision ( nice tool BTW!) and used the OCCT GPU test. 3 funny things - firstly, I used the same test the other day, and the test screen was showing 123 FPS, after 30 minutes my temp on the card was 91. Last night when I ran it, frame rates on the same test were about a third, and after 30 minutes it blue screened. But temp never went over 60.Incidentally, do you know what the errors it detects if you run it in error check mode for the GPU represent? I tried that mode as well and it picked up thousands of errors.....However, I have carried out 2 fairly long flights in FSX, and report no problems there, no crashes, smooth flights and nice scenery ( no blurries etc )Anyways, i wait for your further instructions, best, Mark

Hi Nick, new sticks are in, bios settings as advised, running OCCT as we speak. Prior to installation, I ran through all your system tweaks and made sure all were carried out ( except the defrags, I have set stealth defrag for every time I boot and all drives report zero fragmentation) However, despite 2 successful and stable flights last night, after that FSX took a huge time to load and I had the graphics anomolies again. These anomolies also carried on over onto the desktop, after I switched FSX off. Went away after a reboot. riorto the flight, I ran AlacrityPC again to ensure that there were no services stopped that should have been started. Last night I also used the frame rate limiter utility. I have not used both of these apps since we started this thread.Update: OCCT just stopped after 10 minutes with an error message "error detected on core 2"OK, so what does this mean and what do I do now?thx, Mark

Hi Nick, new sticks are in, bios settings as advised, running OCCT as we speak. Prior to installation, I ran through all your system tweaks and made sure all were carried out ( except the defrags, I have set stealth defrag for every time I boot and all drives report zero fragmentation) However, despite 2 successful and stable flights last night, after that FSX took a huge time to load and I had the graphics anomolies again. These anomolies also carried on over onto the desktop, after I switched FSX off. Went away after a reboot. riorto the flight, I ran AlacrityPC again to ensure that there were no services stopped that should have been started. Last night I also used the frame rate limiter utility. I have not used both of these apps since we started this thread.Update: OCCT just stopped after 10 minutes with an error message "error detected on core 2"OK, so what does this mean and what do I do now?thx, Mark

I am just catching up here with what you postedDo not use OCCT for video card testing.. its only to be used for CPU stress testing. For video card I set up loops of 3DMark and let it run then check the temp graph after 5 or so loops. Sometime I will run moreIf a OCCT failure shows then that means the OC settings need to be adjusted. CPU Voltage, DRAM voltage, NB voltage need to be looked at. I need to know exactly what you are set up to right now in the BIOS based on the list above and what the CPU temp showed during the stress testsStealth defrags do not work the same as name defrags and such background tools should never be enabled. Preboot defrags are not doing the same job either but regardless that is not causing these errors.Anomalies showing in the desktop tends to suggest either a hardware failure or a overclock failure. Once the system is stable and error free through testing if those show up on the desktop it could be an indication the video card is defective. I have also seen power supplies going bad bring on errors like that under stress

also.. please show me the CPUz 'CPU' and 'Memory' tabs

Hi Nick, heres the screens you requested:temp never went over 50 degrees during the stress test ( all 10 minutes of it) Incidentally, while waiting for your post, i just ran up FSX, and check out the difference in the screen rates tonight and last night. The pliatus shot is from last night before the changes I listed above and the new memory. The T-45 is just now. Config files settings unchanged. Spooky Eh? I know frame rate counting is not what its all about, but that is a hell of a difference in performance.....mind you that has been a symptom all along, a crash and then storming performance after...let me know what settings you want me to change, cheers, mark

I am just catching up here with what you postedStealth defrags do not work the same as name defrags and such background tools should never be enabled. Preboot defrags are not doing the same job either but regardless that is not causing these errors.

Do you mean i should NOT have O&O set to do a stealth defrag on boot? I have it set to do all the drives on boot. I thought it would be a good thing? ( there is so much to learn here... )Oh, and since the reboot, as you can see, no graphics troubles either....

OK, assuming that temp came from REALTEMP or CORETEMP and never went over 50c under a full load (which sounds kinda fishy for a quad), lets find out if this a CPU or memory instability issue.. raise CPU VOLTAGE to 1.32 and retest OCCT 1hr CPU testIf that crashes OCCT I want you to lower CPU Voltage back down to 1.28 and set the DRAM VOLTAGE to 1.95 and retest OCCTLet me know the resultsStealth defrag at boot is not needed.. the only 2 defrags that are needed is the complete name and the offline one a month or soIf I did not post it in my list.. its not going to deliver the result

temps were from real temp nick and duplicated by OCCT readouts. Just changing the settings now and running again.... and dont forget, I have a pretty good watercooling setup....

temps were from real temp nick and duplicated by OCCT readouts. Just changing the settings now and running again.... and dont forget, I have a pretty good watercooling setup....

I just want to be sure.. this is clocking 101 over the internet so my concerns are with your hardware when I see things that may not add up and I just want to double check What you are looking for here is which component needs help... the memory or the CPU (or both)This is just step 1.. you have memory latency optimizing to deal with and that may change the ballgame again requiring voltage trim

Hi Nick, OCCT crashed at 10 minutes with cpu at1.32. Changing Dram to 1.94 (1.95 not available as choice, only 1.94 & 1.95) will report backps, what do you mean by "this is clocking 101 over the internet"? (sorry for my lack of technical knowledge...)in all these tests, i am seing core 1 at 50 and the other 3 at mid to low 40's.....EDIT:update, 20 minutes into OCCTwith the dram volyage changed and it has not crashed so far!! Looking promising...

Hi Nick, OCCT crashed at 10 minutes with cpu at1.32. Changing Dram to 1.94 (1.95 not available as choice, only 1.94 & 1.95) will report backps, what do you mean by "this is clocking 101 over the internet"? (sorry for my lack of technical knowledge...)in all these tests, i am seing core 1 at 50 and the other 3 at mid to low 40's.....EDIT:update, 20 minutes into OCCTwith the dram volyage changed and it has not crashed so far!! Looking promising...

Clocking 101 means its a beginners clocking courseIn the states someone may say.. this is driving 101, flying 101, etcDuring the load test.. check your CPU load level.. bring up the monitor from the task manager and see if all 4 cores are at 100%If 1.95 DRAM and 1.28 CPU works then we will lock it there and the next step will be to trim latencyPlease download this software: http://www.tweakers.fr/download/MemSet40.zipunzip and run memset.exePlease post the GUI that comes up

Hi Nick, looking good after nearly 1 hr of occt with cpu back to your original setting and dram up to 1.94. I need to go to bed now, or the missus will kill me, catch up with your thoughts tomorrow, and, i know I have said it already, but thx again for your help best, Mark

Hi Nick, OCCT completed without errors or problems, here is the screengrab from Memset:speak soon, will try running 3dmark a few times while i wait to hear from you re next steps, thxEDIT: I cannot get 3d mark to even run, it freezes after 5 seconds, gonna try swapping out the card for my old GTX8800ultra, as obviously something screwy here.... BTW, FSX and black shark run, but not 3dmark, strange... will report after card swap