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

Overclocking the Q8200

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
Hi, I'm new to this forum and I'd like to thank everyone ahead of time for posting up their info/comments on the q8200.

My Q8200 runs extremely hot and I'd like to confirm if it's me or something else causing the problem. If there are any other recommendations or things to consider any advice would be appreciated.

Here are my specs:
CPU: Q8200
Mobo: DFI JR P45-T2RS (with latest bios firmware)
Memory: 4gb OCZ DDR 800mhz (so-so ram)
Video: Sapphire 4870 512mb
HD: WD Black 1tb
Case: Antec Veris Fusion v2 (has 2 fans blowing air out right by hsf)
HSF: Scythe Ninja Mini with fan attached; Artic silver 5 thermal compound
Fans: 2 on case and 1 on hsf blowing cool air at cpu. 1 fan on NB chipset

Tests run thus far:
CPU: Q8200 @ stock speed @ stock voltage (cpuVcore: 1.18; nb core: 1.18; that's what SmartGuardian reads out)
Temp @ idle: Bios: 26c; SmartGuardian: 28c; CoreTemp: 45/46/45/45 (those are the 4 core temps that it reads out)
Temp @ load: Bios: N/A; SmartGuardian: 34c; CoreTemp: 57/55/57/56 (4 cores)

CPU: Q8200 @ 400fsb(2.8ghz) @ stock voltage (cpuVcore: 1.18; nb core: 1.18; that's what SmartGuardian reads out)
12 hour occt burn in is completely stable
Temp @ idle: Bios: 29c; SmarGuardian: 30c; CoreTemp: 51/49/49/50
Temp @ load: Bios: N/A; SmartGuardian: 37c; CoreTemp: 62/59/61/61

CPU: Q8200 @ 425fsb(2.9ghz) @ (cpuVcore: 1.21; nb core: 1.21; that's what smartGuardian reads out)
OCCT burn in is UNSTABLE and aborts after 12mins
Temp @ idle: Bios: 30c; SmartGuardian: 30c; CoreTemp: 54/51/51/50
Temp @ load: Bios: N/A; SmartGuardian 44c; CoreTemp: 71/69/69/70

Background note: I've built many pc's in the past, but it's been about 6 years since my last build though. For this PC, I have never tried to install with the stock intel hsf. I've reinstalled and reseated the HSF (scythe) 3 times. I've made sure that the pushpins on hsf are "tight" and I assume it's made good contact with cpu. At load, the hsf does not get particularly hot. The back of the mobo where the cpu sits does not get particularly hot either (maybe 38-42c). On touch, the hsf heatpipe close to the cpu do NOT feel particularly hot (definitely NOT 60c). At the top of the HSF (furthest away from cpu), the fins and heatpipe are not at all hot (they are ambient temp).

Other thoughts:
1. Personally, I'm ok with average overclocks. There's nothing I'm going to do that requires me to bump it to 3+ghz or 3.4ghz. However, I'm not happy with the temps that I've getting. Load at 2.8ghz should not be at 60c. I'm afraid it's going to decrease the lifespan of the cpu.
2. I suspect that since my hsf (scythe) isn't getting particularly hot, it's not dissipating heat properly, but everything looks ok on visual inspection. The pins are in properly, everything looks level, and the pushpin on hsf look/feel tight (I had to put quite a bit of pressure for the last pin to click).
3. I also suspect that my cpu is actually running quite cool (unfortunately, I don't have a separate thermal diode), but it's getting erroneous readings from coretemp/realtemp. Additionally, the reason the CPU crash at 425fsb was most likely due to the low voltage being applied to cpu/nb cores. (just a theory)
4. Any recommendations would be great. I don't want to be shotgunning in the dark, doing the same things and expecting different results.

Thus far what I can't understand:
1. Why is my smartguardian program (comes with DFI board) reporting such low temps CPU: 37c, when Coretemp is reporting 60c's?
2. Why is Coretemp (and RealTemp) reporting such high numbers?
3. Why is stock/default voltage for my mobo on q8200 so low?

Thanks ahead of time,

Joey
 
There is nothing wrong with the temperatures of your Q8200. Either the sensor on this chip is buggy or it senses high temperatures.
If you touch the heat sink which you have, you wont find it that hot. So this cpu reports high temperatures. Dont worry.

Smartguardian tells you the temperature of your CPU, as in inside the case.
Coretemp tell you about the temperature of your DYE (Q8200)

Stock/default voltages are low because you have a 45nm chip my friend, they are designed to run on lower voltages.

Hope this helps.
 
Thanks for the heads up on where the temp monitors were reading the temps.

1. Are the temps that's on your system (idle: 50c and load 68-70c) normal for you? Are they still at this temp or have you done anything to tweak them down?
2. How does your hsf feel similar, heat-wise, to mine? (I know this is more subjective than objective, but the problem is I think my probes/sensors are out of wack)

Assuming my temp sensors are wack, I'll resort to the feeling the hsf/back of mobo method for temp sensing. I'll see how high I can go before it crashes again.

I hope this will help people in the future.

There is nothing wrong with the temperatures of your Q8200. Either the sensor on this chip is buggy or it senses high temperatures.
If you touch the heat sink which you have, you wont find it that hot. So this cpu reports high temperatures. Dont worry.

Smartguardian tells you the temperature of your CPU, as in inside the case.
Coretemp tell you about the temperature of your DYE (Q8200)

Stock/default voltages are low because you have a 45nm chip my friend, they are designed to run on lower voltages.

Hope this helps.
 
Im using an after market cooler.. that is the Asus Silent Knight 2.. Which is a pretty good cooler. With room temperatures around 20 the Q8200 @ 3.3ghz 100% load does not cross 60. Where as idle is around 45-49.
On stock HSF i have a +10, i.e 55-59.
 
Thanks to SamSaveMax and Steven and All! Now I am using your result... For my build, the core temp is higher, idle with 52 52 42 42C, full load with max 75 72 67 67C. Can't pass long time P95 small FFTs. (Only passed once 20mins) Probably due to my G-751 Thermal Compound not apply correctly (not even or too much or too less). Another issue is that the Ram is running only at 960MHz, I will try to run it on around 1066MHz.

Thanks again!

Edit1: (480x7)
It seems that I can't run the ram at 1152MHz even with 2.220v ram voltage.
The cpu voltage can be lower to 1.2625v to run at 480x7.

Edit2: (480x7)
I reapply the thermal compound, now the core idle@ 51 51 41 41C, almost the same as above.
then full load P95 small FFT, below is the result:
71 71 65 65 (3mins)
76 73 71 72 (6mins)
78 75 72 72 (10mins)
79 77 73 75 (14mins)
79 77 73 74 (19mins)
79 76 73 74 (25mins)
79 77 74 74 (30mins)
freeze at 38mins with 79 77 74 74

So, after 13~15 minutes, the temperature was stable but the system still freeze at 38 minutes.
Please help!

Edit3: (460x7)
One question, if my Q8200's VID is min 1.1500v max 1.1875v, then in BIOS, I should set the CPU V Core to a much lower value than 1.2625v, such like 1.2125v. Is this right?
Right now, I am running on 460x7 with 1.2125v core, 1.30v MCH and testing.
74 72 69 70 (15min), looks like I set the V Core too high before. Now the temperature is a lot lower then it in edit2.

Edit4: (460x7, CPU VCore 1.20000v, MCH 1.20v)
Looks like I was right about my question in edit3. Here is the result when I set VCore to 1.20000v
53 53 44 44 (idle)
72 71 65 65 (3min)
72 72 65 65 (6min)
72 72 65 65 (11min)
72 72 65 65 (20min)
72 72 65 65 (30min)
72 72 65 65 (50min)
72 72 65 65 (80min) .... stabled!!!! LOL

Note: There is one good thing about GA-EP45-UD3P, which is if you set the CPU VCore to 1.20000v in BIOS, the Loaded Core Voltage in CPU-Z or SpeedFan is 1.20v! (I didn't go to see the idle Core voltage) There is no vdrop here. In SamSaveMax's ASUS P5Q, we need to take the vdrop into account.
 
Last edited:
I am going to settling down at 460x7, 1.19375v CPU VCore, 1.20v MCH and Ram is running at 1104MHz, 2.10v, 5-5-5-15

One question, in Prime95's Torture test, I can only pass Small FFTs. There will be quick error in the other two tests, In-place large FFTs and Blend. Should I work on it or it is fine? Thanks.
 
You are probably pushing your ram too hard. They might be rated for 1066, but your pushing them higher then that. Try running memtest or lower your multi by one, or back your clock off by like 10FSB and run Prime again. Small FFTs is CPU intensive, while Large FFTs and Blend include ram in the stability test. Some ram can be overclocked, and some cant. I know yours israted for 1066 but anything over is considered iceing on the cake.
 
You are probably pushing your ram too hard. They might be rated for 1066, but your pushing them higher then that. Try running memtest or lower your multi by one, or back your clock off by like 10FSB and run Prime again. Small FFTs is CPU intensive, while Large FFTs and Blend include ram in the stability test. Some ram can be overclocked, and some cant. I know yours israted for 1066 but anything over is considered iceing on the cake.

You are right, it is the ram. I can't make it over 1066 with 5-5-5-15.
Now, it runs on 470x7 CPU, 470x2 RAM, stabled... OK!

Thanks!
 
I tried to upload a ss Cant seem to get the file size small enough.
Ill edit my sig.
Ive had to up my 500fsb voltages slightly to hold tRD9
490fsb is much easier :).
Figures in breackets are for 490fsb.
For nb temp I drilled the hs and put in an external read sensor.
nm heres one : How not to get 24/7 stable settings.
1287d1247704842-beware-validation-mediaonevalidation-bs.jpg


heres a not very good pic of the inside
1156d1244976943-ud3p-north-bridge-tim-scan0002.jpg
 
Last edited:
Looking good man!

Since you are pushing the FSB so hard you should run Large-FFT first. I'd also drop the CPU multi to 6 while testing this out. The FSB is going to be the main hold back for you. Once that is stable you can up the CPU multi, and use Small-FFT and vcore to stabilize the CPU which shouldn't be hard as the chip has plenty more in it (it just can't handle the high-FSB and the low multi limits its potential).
 
@kick: If you look closely, your screenie above shown worker#2 had failed.
Both Speedfan & Realtemp confirmed that as well.
 
Last edited:
Ahh sorry some of you may have missed the tongue in cheek part of it..
I prefaced the screenie with "how not to get 24/7 stable settings.

The above screenie I took from a post I made on another site warning of the dangers of cpu-z validation.
Ie I could have validated those settings and it wouldnt be "real".. I think I could validate up to around 530 fsb on air if I wanted to risk burning out the cpu /mem/mbrd.

Yeh cant miss the worker failing.. it was just a .. be carefull on others validation post. My sig figures are what I use most of the time.
I can get 507 stable for IbT /prime/ for an hr or two.
I couldnt be bothered runing it longer its just I like to be able to test run 5 or so mhz above what Id normally run

After 8 months or so .. on my machine there isnt a lot of change on the original posters 3.5Ghz on air . Yeh I/anyone could probably run 3.55-3.6 with highish voltages without errors but maybe having to gimp memory down to cas 6 timings(depending on sticks).If u have to do that tho youre gonna drop overall pc performance.

Ok heres a 1hr run at 505

How do we know its a 1hr run ! you didnt stop the test..
Take my word for it :) + students of prime 95 sequencing should be able to work it out oh and pc wizard temps should be transposed and the 12v measuring is way off as usual.
1278d1247536028-q9550-ud3p-part-deux-quest-500-fsb-5051hr.jpg


Were at the end of the line on chipsets and lets say generally 45nm quads dont like much above 500/2000 fsb on air..thas ok my 8200 blows away all standard clock quads :)

ok heres 490 with low(ish) voltages.

1209d1246442155-q9550-ud3p-part-deux-quest-500-fsb-2hr490lowcpuv.jpg


same test but just after prime is stopped
1210d1246442707-q9550-ud3p-part-deux-quest-500-fsb-2hr490lowcpuvstop.jpg


also on small /large Ftt they arent a problem the only difficult ones for me are on blend.
For occt peeps like Sam..yeh I run that too generally just the 30 min tho.

My sig 14 hr prime has an average room temp of 25C .. a bit hot but it will run it.
 
Last edited:
Hello, I'm new to this forum and I've been following these posts for a while and decided to purchase the Q8200 till I get around to getting myself an i7. anyway when I get around 3.40 it'll post but Vista will not load it'll either freeze or show the bsod then restart. Any ideas? Here's my specs

Asus P5k
G,skill DDR2 1066 ram timings are 5.6.6.18
Asus ATI Radeon x1950pro
HSF is a Scythe Andy Samurai Master with a 120mm fan on it.

Here's a screenie of my CPU-z where it's currently at.
 

Attachments

  • Untitled.jpg
    Untitled.jpg
    181.3 KB · Views: 1,618
just as a qwik guess .. your cpu core voltage is very low even with vdroop,cpu will work at v low volatages 400-450+ after that you may have to bump it hard,anything above 470 as you get near 500+ it seems to need more n more.
Depends on your cooling too.
 
Hello, I'm new to this forum and I've been following these posts for a while and decided to purchase the Q8200 till I get around to getting myself an i7. anyway when I get around 3.40 it'll post but Vista will not load it'll either freeze or show the bsod then restart. Any ideas? Here's my specs

Asus P5k
G,skill DDR2 1066 ram timings are 5.6.6.18
Asus ATI Radeon x1950pro
HSF is a Scythe Andy Samurai Master with a 120mm fan on it.

Here's a screenie of my CPU-z where it's currently at.

I would say your at about the limit for that P35 board. You're not going to get much more than 460FSB stable, and good job on getting there.

That's one thing that sucks about these low multi quads. The FSB gives out way before you reach the full potential of the chip, and thus you reach a limit, and giving a lot more vcore does absolutely nothing. You can play w/ vNB, GTLs, etc for weeks, and maybe get a few more FSB MHz...but not much.

You'll be able to go a little higher on a P45, but it isn't worth it. Especially if this is just a hold-over for i7.

I say be happy w/ what you've got, and maybe even try to lower any voltage setting you can while remaining stable. Maybe tweak out the RAM/NB timings if your up for a challenge.
 
just as a qwik guess .. your cpu core voltage is very low even with vdroop,cpu will work at v low volatages 400-450+ after that you may have to bump it hard,anything above 470 as you get near 500+ it seems to need more n more.
Depends on your cooling too.
Thanks for the reply kick! Right now I have it at 3.30 at voltage 1.400 seems ok so far.

I did have it running at 3.40 but again I couldn't even get to the green bar. It also did this at 3.35 and higher. I did raise the voltage with no luck.
 
Back