Curse

malevolence · Oct 15, 2009, 10:00 AM // 10:00

Hello, yeah its me again....

I made this topic because is not related to the hard disk issue I was having in my previous topic.

Now, the issue is with my motherboard XFX 790i Ultra with the latest bios P08.

So far for me this issue started when flashed the bios to P08 (trying to fix the hard disk issue), and since I have this bios P08 the motherboard doesn't like my RAM anymore , well , actually it does, but not when all 4 memory sticks are installed.

I will explain what I did so far;

-I discovered that my motherboard doesn't like to boot when my 4 memory sticks are installed on the 4 memory slots , giving the annoying error C1. NOTE: Sometimes it does boot! buuut giving a blue screen when starting windows.

-If I install only 1 memory stick on the first memory slot (slot 0) , the system boots and not blue screens etc. It just does fine. I tested 1 memory stick in all 4 slots (slot 0, 1 , 2 and 3) and the slots seems to work fine , as the system and windows booted without blue screens or error messages.

-I tested each memory stick with memtest86 (bootable latest version) option 1 (1 cpu core , as I select the multi core and the program crash?) and tested every memory stick, 1 by 1, for long 3 hours each and no errors . NOTE: Also I cannot install 4 memory stick and run memtest86 (if I didn't get the c1 error), memtest86 crash.

So for some reason , since I flashed the bios to p08 I am not able to put all my ram....

Here are the timmings ,FSB and volts for my ram :

FSB 1600 9-9-9-24 T2 1.9v as recommended by Corsair (the manufacture)

With the previous bios p07 I didn't have this issue at all , it was working fine always , even with the 4 RAM installed.

What I want to do now , is to flash the bios back to P07 to make my system stable again, but I cannot find the bios P07 anywhere , on teh xfx website is only the P08

, as I have a ticket open with them , I asked where can I get the P07 Bios but I didn't get an answer yet.

If someone here has the P07 bios , please send me a link or email me the bios. Email me to [email protected]

Thanks.

tw1tchdp · Oct 15, 2009, 10:36 AM // 10:36

Didn't read entire thing but your motherboard came with disk should have earlier bios to install maybe?

Kenzo Skunk · Oct 15, 2009, 12:17 PM // 12:17

You can check memory from the starting menu of many linux distributions, straight from booting your comp from cd. This looks like two options: fault mobo or memory.

Leigh The Legendary · Oct 15, 2009, 12:28 PM // 12:28

Have you tried reinstalling windows?

Sometimes if there are significant changes in hardware (ie you updated the bios which means some things may have changes) windows hardware abstract layer will have problems. A simple reinstall of windows should show if there truly is a problem or not. Just back your data up and blow it away

malevolence · Oct 15, 2009, 12:51 PM // 12:51

Yeah , good points I forgot to mention.

I installed windows vista after my bios upgrade , I don't rememer if I installed windows with my 4 RAM modules installed , mmm, I think I am gonna try that now, and no backups needed is a fresh installation anyways ^^

malevolence · Oct 15, 2009, 01:34 PM // 13:34

Alright , finally manage to boot with all my ram installed. I lowered the volts on the bios to 1.7v instead of 1.8v. I am running Prime95 64bits version now to stress the pc and test stability

. Lets see how it goes.

Quaker · Oct 15, 2009, 05:49 PM // 17:49

I was going to suggest that you set the RAM timings to "auto" or "from SPD".

malevolence · Oct 15, 2009, 06:38 PM // 18:38

alright , prime95 crashed , cannot run it for even a minute lol.

Ec]-[oMaN · Oct 15, 2009, 06:49 PM // 18:49

I've tried searching on google also but all that comes up is people in your exact situation saying that p08 bios is really flakey with ram, and people asking for a download link to the p07.

By any chance you think this xfx board and the evga boards are using the same bios, people note almost the same thing and the EVGA boards using P"" bios tags also... same chipsets....

malevolence · Oct 15, 2009, 10:59 PM // 22:59

Quote:

Originally Posted by Ec]-[oMaN

I've tried searching on google also but all that comes up is people in your exact situation saying that p08 bios is really flakey with ram, and people asking for a download link to the p07.

By any chance you think this xfx board and the evga boards are using the same bios, people note almost the same thing and the EVGA boards using P"" bios tags also... same chipsets....

uh good question , but I don't know if that works. I am trying to get the link from customer support now , lets see.

EDIT:

Just got it ! here it is http://www.xfxsupportb.co.uk/XFX_790i_P07_ISO.zip

SO I am going back to P07 yay !

malevolence · Oct 16, 2009, 11:23 AM // 11:23

lady's and gentleman's

I flashed bios P07 , my memory was at 1.8V , system boot, windows boot. Tried to run Prime95 in 2 or 3 minutes crash. So I went to the bios , set the mem volts to 1.9V (I was always using like this). Went to Prime95 , it was better this time, it crashed after 2 hours, so still something is wrong here , moving the volts of the mem to 1.9v worked ,but not for a long run.

Something that I was looking is how *hot* my cpu is running... it max was 72º after 1 hour of 100% load, I think that is too much, the funny part is that I touch the sink , and is cold nothing hot, I think something is wrong but I don't know, so I took the cooler out , cleaned the cpu , then I applid the "OCZ Freeze Extreme Thermal Conductivity Compound" that is brand new , just got it last week. mounted the cooler and same results , is hot hot hot

, at idle is like 40ish , and when is at 100% load it goes easily up to 62º and after an hour is like 70-72 , ahhh by the way my cooler is this one: Xigmatek HDT-S1283.

I am running my CPU at stock/default , no OC here... volts are managed AUTO , so I am not changing anything in the bios related to the CPU.

My questions are now,

or the Xigmatek HDT-S1283 isn't working ?
or the OCZ Freeze Extreme Thermal Conductivity Compound isn't good at all ?
or there is something wrong on the installation of the cooler ? (I cleaned the old thermal compoun from last week , and applied a new one.)
or is my CPU broken ?

malevolence · Oct 16, 2009, 01:10 PM // 13:10

Wow, this review say it all, I am having problems with the cooler , thats for sure, they tested a Xigmatek HDT-S1283 with the same CPU I have the Q6600 and they have far far lower temps than me.

http://www.hardwarecanucks.com/forum...-review-5.html

They said this too :

Quote:

Plastic pushpins are not what you call a great long term installation method when you are dealing with a cooler which is above their maximum specified weights. When Intel engineers designed this system, they specified 550 grams as the nominal maximum weight it would be able to support. Unfortunately, even at 550 grams these little pushpins tend to flex a lot (as any malleable plastic material will) which means that you will get inconsistent mounting pressure on the CPU as time goes on. Even though we say this, you should remember that uneven pressure shouldn’t break your chip or anything drastic like that (far from it in fact) but it does mean that part of the CPU will not have great contact with its CPU cooling solution. This in turn will lead to increased heat as the cooler will be unable to remove the entire thermal load away from the CPU. With lower TDP (Thermal Design Power) chips like our e4600, this should not be an issue as this particular cooler is designed to be ultra efficient, however it will be interesting to see how it fares on the hotter q6600.

Elder III · Oct 16, 2009, 03:10 PM // 15:10

if you still have the stock cooler (assuming it came with one) you should try that.... even intel's "so-so" stock coolers will cool your CPU way better than that. Also, make sure you get an even and thin layer of thermal paste on their.... you don't want too much.

As far as Prime95 goes, considering the notoriety of that motherboard when it comes to being picky about memory and Bios settings, anything over 2 hours is sufficient. You could also try Everest Ultimate (free trial edition) for their Stability Test, it stresses the CPU 100% load (as well as some other things) and will show you if you have a stable system, without torturing it like Prime95 does.

malevolence · Oct 16, 2009, 03:24 PM // 15:24

cool Thanks ELder III ,

now the temperature , I went to the OCZ website and applied the thermal paste exactly as they recommended and it lowered my temperature a bit.

Thats a good idea, I should try the intel cooler and see how it goes.

I will update.

Snograt · Oct 16, 2009, 04:03 PM // 16:03

If my eVGA nForce experience is anything to go by, every "P" BIOS from P01 to P08 has been a P0S.

malevolence · Oct 16, 2009, 06:41 PM // 18:41

lol snograt hehehe thats true ^^

Ok , time for an update , after the good application of the OCZ thermal paste , I ran Prime95 for 3+hrs without crashes or errors. and my temperatures were:

Load: 100% all 4 cores

-Core#0: Low: 42ºC High: 54ºC Load:100%
-Core#1: Low: 42ºC High: 54ºC Load:100%
-Core#2: Low: 40ºC High: 49ºC Load:100%
-Core#3: Low: 43ºC High: 51ºC Load:100%

Well , my temperatures are better , but , how come that I see a review with the same Cooler I have and the same CPU I have (even OC'd) and they have a much lower temperature ?

I am measuring the temperature with the software Core Temp version 0.99.5 , is this program accurate ?

PS. I will try out Everest Ultimate later

Elder III · Oct 17, 2009, 03:09 AM // 03:09

Well for reference sake I have a computer sitting on my test desk right now with an AMD 945 Deneb Quad Core OC'd at 3.4ghz and it only hits 54C after running awhile with 100% CPU load.... stock cooler and stock thermal "pad" - AMD does have better stock coolers than Intel imo, but nothing compared to the Cooler you bought.... either your case is not getting good airflow or you have a semi effective cooler. Personally though I wouldn't worry about it if that's the clock you want to maintain... those temps aren't in any danger zone or even close and you will rarely if ever hit 100% CPU load in regular useage.

malevolence · Oct 18, 2009, 04:10 PM // 16:10

Quote:

Originally Posted by Elder III

Well for reference sake I have a computer sitting on my test desk right now with an AMD 945 Deneb Quad Core OC'd at 3.4ghz and it only hits 54C after running awhile with 100% CPU load.... stock cooler and stock thermal "pad" - AMD does have better stock coolers than Intel imo, but nothing compared to the Cooler you bought.... either your case is not getting good airflow or you have a semi effective cooler. Personally though I wouldn't worry about it if that's the clock you want to maintain... those temps aren't in any danger zone or even close and you will rarely if ever hit 100% CPU load in regular useage.

Thats not the clock I want, I want to OC this cpu to at least 3.2ghz , I bought a new cooler , so I will try out on tuesday I think. (plus the coolermaster HAF case)

Lord Sojar · Oct 21, 2009, 05:55 PM // 17:55

Hmm, glad you figured out it was a thermal issue. My initial guess after reading your first post was actually faulty memory controller, but next in line is temperature issues when it comes to Prime95. 3+ hours is acceptable for Prime95 on Torture mode.

If you want to clock it to 3.2GHz, you may want to check the airflow in your case. Easiest way to do that is to simply take an incense or something that generates lofty smoke that is easy to see, and put in front of the intake on your case. Watch how the smoke moves using any of the following light sources: blacklight, laser pointer refracted using the internal case metal, led flashlight. See if a lot of the smoke is being thrown out the back, or if it is lingering in the case for long periods. Then, you can decide which route to take in regards to improving airflow in any given direction. You want negative pressure in the case, meaning hot air is being forced out the back rapidly, and not staying in the case, creating positive pressure.

malevolence · Oct 23, 2009, 10:37 AM // 10:37

Quote:

Originally Posted by Rahja the Thief

Hmm, glad you figured out it was a thermal issue. My initial guess after reading your first post was actually faulty memory controller, but next in line is temperature issues when it comes to Prime95. 3+ hours is acceptable for Prime95 on Torture mode.

If you want to clock it to 3.2GHz, you may want to check the airflow in your case. Easiest way to do that is to simply take an incense or something that generates lofty smoke that is easy to see, and put in front of the intake on your case. Watch how the smoke moves using any of the following light sources: blacklight, laser pointer refracted using the internal case metal, led flashlight. See if a lot of the smoke is being thrown out the back, or if it is lingering in the case for long periods. Then, you can decide which route to take in regards to improving airflow in any given direction. You want negative pressure in the case, meaning hot air is being forced out the back rapidly, and not staying in the case, creating positive pressure.

Excellent advice Rahja

I will try that just for fun at the moment

Well, my xigmatek cooler wasn't working properly , probably the plastic pins were not making enough preasure to the sink and the cpu and therefore was overheating.

I got the new cooler on tuesday and yesterday I ran Prime95 StressTest for 9+ Hours , and the temperature was very good comparing to my old cooler , it was stable for 9 hours ! (cpu at stock settings) and the max temperature reached was 45 degrees lol , thats great for 9+ hours I guess. And the idle temperature is 25/27 degrees

Next week I will get the coolermaster HAF , and I will test the system again.

I am still having some issues with the memory, I tested (with memtest86) every memory module on every memory slot for hours and all modules and slot seems to work perfectly.

Memtest86 crash if I have all 4 modules (or more than 1 module) installed at the same time , even if windows boot and prime95 was stable for 9+ hours with 2 memory modules installed. Is that normal ?

Anyway I still need to test all 4 modules with prime95 (the 9+ hours stresstest was done with only 2 modules in dual channel configuration)

I will update