• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Need general advice overclocking AMD 1090t [BSOD]

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Rooshy Pooshy

Registered
Joined
Apr 18, 2012
Alright guys, I have my AMD 1090t BE overclocked to 3.8 ghz @ 1.425v and it seems to be pretty stable. My temperatures are ~20c idle and ~50c under 100% prime95 load after 10 minutes. I've got my CPU-NB at 2600mhz with 1.175v. It was being pretty stable but now when I run prime95, after about 10 minutes I will get a BSOD that says "a clock interrupt was not received on a secondary processed within the allocated time interval" WHAT COULD BE CAUSING THIS? :bang head
Does my CPU-NB need to be OC'd more? More voltage on CPU or NB? Do I need to OC my RAM? (no ****ing idea how to do any RAM overclocks btw :cry: "
 

Attachments

  • Cpu z and CPUID.png
    Cpu z and CPUID.png
    57.5 KB · Views: 157
up your ht link I had my NB frequency the same as my HT Link. I think I used 2400Mhz.. I don't remember whether or not my voltage was increase on NB but for 3.8 I was able to go to like 1.4Vcore.
 
up your ht link I had my NB frequency the same as my HT Link. I think I used 2400Mhz.. I don't remember whether or not my voltage was increase on NB but for 3.8 I was able to go to like 1.4Vcore.

I thought you were supposed to keep the HT link at or very near 2000 mhz? Also, do you think my BSOD could be caused by a slow HT clock?
 
Try leaving your NB stock and see what happens with just volts and multi on the processor. I had stablity problems when I started messing with the Hyper Transport speeds iirc 2000 what you should be trying to stay at. I think there is a phenom II overclocking guide on the forum somewhere.
 
Try leaving your NB stock and see what happens with just volts and multi on the processor. I had stablity problems when I started messing with the Hyper Transport speeds iirc 2000 what you should be trying to stay at. I think there is a phenom II overclocking guide on the forum somewhere.

I have read through every Phenom II guide I could get my hands on lol :thup:
I thought that when the CPU is being overclocked you have to OC the CPU-NB to keep up with it? I'm probably wrong. I will try your suggestion and report back! :attn:
 
Give your CPUNB voltage a bump to 1.225 and your HT Link frequency a bump to 2200. The main problem is your CPUNB v was too low, I think. The Thuban core CPUs seem to like the HT Link frequency to not get too far apart. Early on advice was to keep the CPUNB frequency and the HT Link frequency the same. And that seemed to okay for some of them but in time other Thuban overclockers reported that the HT Link couldn't keep pace with the overclocked CPUNB but found it did handle a small overclock just fine and that helped with stability.

The other think to be aware of is that the Thuban core CPU core temp sensors are often calibrated poorly, giving readings that are about 10c too cool. So, your core temps may actually be closer to 61c than 51c and you may be getting heat-related instability. What are you cooling this thing with?

Also, when you post pics of CPU-z, please include the "Memory" and "SPD" tabs.

And it would be helpful to include some detailed info about your system. You can refer to mine if you wish for an example of what we're interested in. When you have a few more posts you will have access to the "Sig" feature and it is common courtesy to put the system info in your signature.
 
Last edited:
Give your CPUNB voltage a bump to 1.225 and your HT Link frequency a bump to 2200. The main problem is your CPUNB v was too low, I think. The Thuban core CPUs seem to like the HT Link frequency to not get too far apart. Early on advice was to keep the CPUNB frequency and the HT Link frequency the same. And that seemed to okay for some of them but in time other Thuban overclockers reported that the HT Link couldn't keep pace with the overclocked CPUNB but found it did handle a small overclock just fine and that helped with stability.

The other think to be aware of is that the Thuban core CPU core temp sensors are often calibrated poorly, giving readings that are about 10c too cool. So, your core temps may actually be closer to 61c than 51c and you may be getting heat-related instability. What are you cooling this thing with?

Also, when you post pics of CPU-z, please include the "Memory" and "SPD" tabs.

And it would be helpful to include some detailed info about your system. You can refer to mine if you wish for an example of what we're interested in. When you have a few more posts you will have access to the "Sig" feature and it is common courtesy to put the system info in your signature.

I will try all this when I get home later. Quick question, isn't the max CPUNB voltage supposed to be 1.2? I thought any higher than that was bad... Also, I don't think my problem is related to heat stability (Are the readings really that far off?) because I'm cooling my cpu with a hyper 212+ in push pull with two 69 CFM fans blowing toward my exhaust fan and with a 200mm exhaust fan above it. I will also post those other cpu-z tabs when I get home. Is my specific BSOD a heat related error?
 
I will try all this when I get home later. Quick question, isn't the max CPUNB voltage supposed to be 1.2? I thought any higher than that was bad... Also, I don't think my problem is related to heat stability (Are the readings really that far off?) because I'm cooling my cpu with a hyper 212+ in push pull with two 69 CFM fans blowing toward my exhaust fan and with a 200mm exhaust fan above it. I will also post those other cpu-z tabs when I get home. Is my specific BSOD a heat related error?

AMD states that maximum safe CPUNB voltage is, believe it or not, more like 1.5 but I would never put it that high myself. Stock is about 1.175. Most people find that 1.225-1.25 is about right for carrying the overclock.

If you look at your core temp at idle in your HWMonitor pic you will see that min is 19c. Unless your room temp is cooler than normal, that's below ambient. So, that should tell you something right away about the calibration of your core temp sensor. I would add at least 5-6c too it for a realistic estimate. On the other hand, your core temp is within 2c of your CPU socket temp. Usually there is more spread when you have a good aftermarket air cooler like you do. Usually we see more like a 10c diff between the two in that scenario with the socket temp the higher one. So, the core temp reporting may be more accurate than I'm giving it credit for. Just something to be aware of, that's all. Sometimes core temp sensors calibration is such that the temp reading under load is pretty accurate but the idle report is skewed, i.e. the calibration was done on the upper end of the scale.
 
Last edited:
AMD states that maximum safe CPUNB voltage is, believe it or not, more like 1.5 but I would never put it that high myself. Stock is about 1.175. Most people find that 1.225-1.25 is about right for carrying the overclock.

OK :D
Another question for you, what should I have my DRAM frequency set at? It is set at 800 mhz by default in the BIOS. Its this RAM http://www.newegg.com/Product/Product.aspx?Item=N82E16820231311. On a side note, how do I get my rig in my signature??

Edit: I got my signature set up. I'm an idiot :p
(can you see it?)
 
Last edited:
For "Sig" go to top of page and Quick Links and Edit Signature. Not sure how many posts you need to make to have access that. Someone said 5 and someone else said not until you have one gold star under your user name. Try it and see.

Before I can answer your question about the DRAM frequency setting you will need to post pics of CPU-z tabs: Memory and SPD. That will tell us what your memory is currently set to and what the manufacturer says its capable of and the recommended values for voltages and timings.
 
Here's the answer about the Sig from moderator I.M.O.G.: "It's just 2 posts in the hardware or software forums. Once they hit 2 posts, a "promotion" task runs every hour. So once they hit 2 posts, it could be up to an hour before they get the sig, depending on when the task runs next."
 
Here's the answer about the Sig from moderator I.M.O.G.: "It's just 2 posts in the hardware or software forums. Once they hit 2 posts, a "promotion" task runs every hour. So once they hit 2 posts, it could be up to an hour before they get the sig, depending on when the task runs next."

2 new threads or just 2 posts?
Also, I'm currently trying the settings you suggested, I have my CPU at 3.8ghz @ 1.4375v CPUNB at 2200mhz 1.225v and Hyper Transport Bus also set to 2200mhz. Will report back after prime95 testing. Memory and SPD tab screenshots incoming.
 
As promised here is all the CPU-Z information that was requested. Let me know if you need anything else from me :D

Starting prime95 blend test now. Wish me luck!
 

Attachments

  • Updated CPU-Z Stuff.png
    Updated CPU-Z Stuff.png
    103.1 KB · Views: 134
UPDATE: With all of trents recommended settings I was able to run prime 95 for 15 minutes without a hiccup! :D
Temperatures were ~50-55c during the 1024k test, but once it reached the 8k tests temperatures went up to ~60c and I stopped the process just to be safe. Idle temps are still 20c or lower. Now, is there any reason to try to overclock my RAM? Do you think it would be possible to get to 4.0ghz? I think I might need more cooling power for that...
 
One piece of info you need to know to avoid confusion, CPU-z reports the ram's bus frequency which is half of the DDR3 transfer rate. So, you are currently running your ram at 800 mhz (400x2) which is actually a considerable under clock since the ram is rated for 1333 mhz. The Thuban core CPUs will usually be able to handle ram at 1600 mhz without a problem so you should be adjusting your ram frequency upward to run it at the full 1333 mhz it's rated for. Also, bump your ram voltage up to 1.55.

My main concern is your high temps. Was that 60c you quote a core temp or a CPUTIN temp? Core temps are the critical ones and we want to keep them from exceeding 55c under full Prime95 blend load. What is your room temp like? Do you live in a hot climate?

Please have HWMonitor open on the desktop before and during Prime95 testing. The HWMonitor pic you show looks like the temps are at idle and that not very helpful.
 
Here is a screenie of the same HWmonitor window, I left it open after that round of torture testing as well as playing the Witcher for about 2 hours. Should I even be that worried about those temps? I mean, how often will I realistically have all 6 cores at 100% load?
 

Attachments

  • HWM.png
    HWM.png
    28.7 KB · Views: 129
Oh and another thing, I believe I might have used too much TIM when I applied my hyper 212+. I was nervous about the grooves caused by the direct contact heat pipes so I got a little over zealous with the TIM. One small dot on the middle of the CPU and a tiny dot near each of the 4 corners of my heatsink :p
 
Well, we don't even know if your system is stable at the current settings and that is the value of running a long Prime95 blend test (at least 2 hrs.). It ensures the system will be stable under even the most demanding apps.

Your are sacrificing a significant amount of performance with your ram frequency so low as it is now. Your should go into bios and set your ram frequency to 1333 mhz and your timings to what they are in CPU-z "SPD" tab, JEDEC column #4 (9-9-9-25-34-2T). I thought I already said that? Put your ram voltage to 1.55. As far as the memory timings go, you only need to concern yourself with the 6 timings that show in JEDEC #4 plus Command Rate (CR) which should be set to 2T. When you go into bios and take the timings off of "Auto" you will see about 20 or so individual timings. All but the ones in JEDEC #4 you can leave on Auto as they are relatively forgiving and unimportant.
 
Well, we don't even know if your system is stable at the current settings and that is the value of running a long Prime95 blend test (at least 2 hrs.). It ensures the system will be stable under even the most demanding apps.

Your are sacrificing a significant amount of performance with your ram frequency so low as it is now. Your should go into bios and set your ram frequency to 1333 mhz and your timings to what they are in CPU-z "SPD" tab, JEDEC column #4 (9-9-9-25-34-2T). I thought I already said that? Put your ram voltage to 1.55. As far as the memory timings go, you only need to concern yourself with the 6 timings that show in JEDEC #4 plus Command Rate (CR) which should be set to 2T. When you go into bios and take the timings off of "Auto" you will see about 20 or so individual timings. All but the ones in JEDEC #4 you can leave on Auto as they are relatively forgiving and unimportant.

I will try all those RAM settings in a little bit. Even if my system isn't 100% stable (seems pretty close to me, but what do I know) it is wayyy more stable now than it was before. Do you think it's ok to leave my clock settings how they are now? The most demanding thing I do is gaming, so probably either bf3 or GTA 4 with graphics mods.
 
Back