Printable View
100C is the official setting for 45nm CPU's (AFAIR its same for both Duals and Quads) but I do know for a fact its 100C for a 45nm Wolfdale Dual Core.
That info was publically disclosed by Intel months/year ago.
All Apps were updated ages ago for this info inc Everest/CoreTemp and RealTemp.
i always use 100c for tjmax.
Yes 100C is the figure released by intel which many people choose to ignore, for whatever reason - perhaps tjmax 95 gives more accurate readings, or perhaps they prefer to use the same value as that of the QX9650. I'm really puzzled by the 77C load temp you get because my chip, presumably lower binned (being a 9550), never reaches 60C with almost 1.4 vcore. :shrug:
Congrats on that 536x9 andressergio!!! Nice work :up:
thanks mate !Quote:
Originally Posted by OC Nub
Andressergio Is Very Good Men
That's Great FSB For Quad
What's Ur Cooling In High FSB??
Can U Take Shot Ur Rig And :) Maybe Be Mistake For Change Mobo To DFI GigaByte Is Best For Quad
Great job andressergio. I think you will make it to long 5Ghz before I do.
High voltages aren't THAT scary.:up:
Oh mate sorry i forgot to thank you that also sent me some templates :( damn y updated it ;)Quote:
Originally Posted by thebluemeanie1
thank you !!!!
Cheers
Sergio
Good job Sergio, I'm glad you sorted things out :up:Quote:
Originally Posted by andressergio
Thank you mate it just needed a pump on vcore and some tweaks on ram :up:Quote:
Originally Posted by Jor3lBR
cheers
Sergio
Thank you ViViDQuote:
Originally Posted by ViViD
here's the pics posted as i had to re build my loop for some 2 stupid orings that came faulty on the BFG card
http://www.xtremesystems.org/forums/...d.php?t=227904
Cheers :up:
Sergio
You guys really make me want to try water cooling after seeing the pictures. Nice looking rig sergio.
thank you mate and keep in mind that in not using the best rad and also sharing the loop with the VGA, and still have great tempsQuote:
Originally Posted by OC Nub
Cheers
Sergio
Wanna show you some shocking pics from what goes with new Batch of q9650 coming out in good old Germany::eek:
4050MHz with only 1,152v:clap:
http://www.abload.de/img/q9650_4050_115zpbi.png
4320MHz with 1,248v:clap:
http://www.abload.de/img/q9650_4320uovd.png
More not tested, because of cooling with air and temps excied 70° above 4450MHz @1,3v.
So long
nel-p alias nelly
do not forget this: http://www.xtremesystems.org/forums/...d.php?t=206237Quote:
Originally Posted by andressergio
:D :D :D
congrats sergio! :clap: finally 4.8 :toast:
Thank you Napalm !!! will need to pump some volts to run the other test but this batch is HOT !!!Quote:
Originally Posted by NapalmV5
Cheers :up:
Sergio
Congratulations mate 1,15 is very nice !!!Quote:
Originally Posted by Nel-P
Cheers
Sergio :up:
Well on my way to 4900 :D i just tested the flavour for a moment but coudln't SPI, for now i just got this
http://img188.imageshack.us/img188/9766/spi16.jpg
I bought some 2x1GB Crucials Ballistix old batch double sided so i will test with them tomorrow and see how it goes.
Cheers all :up:
Sergio
http://img91.imageshack.us/img91/2329/4ghzstablized.jpg
100% IBT stable. Reduced voltage to 1.3125v (in BIOS), still stable. Probably can reduce more.. going to try it tomorrow.
VID of 1.2750v.. L84xB stepping.
I hold the record for highest VID @ 1.2750v (1 week old CPU).
But I have learned it means nothing on newer 45nm CPU's, does 3GHZ @ 1.03v.
I have not mounted CPU correctly as ran out of good TIM and had use some rubbish stuff that was hard to work with and its applied badly (proper TIM is on route).
Mobo is an Asus Maximus II Formula Rev A2, the heat pipe has so far not being remounted with new/better TIM so although it runs cooler than some older batches it will no doubt be badly TIM'd.
I'm stopping for now at this due to better TIM on heat pipe may let me do more on 1 notch lower volts as 1 notch higher was needed from 470FSB to 475FSB (suspect heat makes it need more).
Ignore VCore etc for now all were set high for testing FSB so I know if it fails its due to FSB not overall MHZ etc.
I have tested it to 3.6GHZ and 4GHZ and will try find lowest voltages possible once new TIM applied.
All cooling fans are are 50% for testing so PC is quiet, bellow is idle but I have Primed/IntelBurnTest'd/MemTest/Vantage stressed it and it passes.
http://img37.imageshack.us/img37/1594/newsfq.jpg
congrats AgentGOD now you finally got a Q9650 :up:Quote:
Originally Posted by AgentGOD
cheers mate
Sergio
So it appears that I got a "bad" quad. If I try to lower the vcore any further down... it will fail IntelBurnTest. And the general conclusion here is that 4 GHz @ >1.28v = mediocre. Is this true?
Naa, I seen average of about 1.25v-1.28v for most on here.
The post the other day with 1.15v on air is exceptional.
Increase PLL Voltage and it should let you lower VCore, try 1.55v PLL for safe start.
What's the maximum safe CPU PLL Voltage? I have no experience with Yorkfield, only Kentsfield :PQuote:
Originally Posted by humeyboy
Well too much PLL Voltage like VTT Voltage (FSB Termination Voltage) can kill a CPU.
Your Mobo has a default PLL Voltage of 1.50v, it turns Yellow at just above 1.60v (Text in Bios and LED on Mobo if you set it for PLL instead of VCore).
Set it to 1.55v which is very safe and try lower VCore.
I have read peeps saying different limits for safe PLL Voltage on these CPU's but 1.60v should be still ok.
I tried bumping the PLL to 1.60v, and I still couldn't lower the vcore without losing stability in IBT :(
Bummer must need the VCore then, PLL overall is for higher MHZ with lesser Voltage but you may be needing more VTT for your current FSB.
Also the GTL's are very important to stableness and you may be able to lower voltages then.
I had no luck with my Quad on this Mobo at even 400-425FSB till I set the NB GTL and I have found my CPU GTL's and NB GTL are stable at + values, the NB wont boot at - values.
With my above Voltages (ignoring VCore and PLL as not important to what I am talking about) I have set all 4 CPU GTL's and the NB GTL to +40.
I calculated this not guesswork, easy since both VTT and NB are set to same Voltage.
The P5Q Deluxe uses ratios instead of +/- values (ex: 0.630x, 0.670x) for GTL references.
So how would to guys recommend switching over to water? Any good complete setups, or build piece by piece?
If you want good water cooling, you should build one yourself.Quote:
Originally Posted by TooFastForYa
AgentGodQuote:
Originally Posted by AgentGOD
as you have a high vid probably you'll need more vcore that any other chip but if temps are not so high (i mean on 24/7) not IBT then its ok...also its important the GTLS on the CPU and MCH so the VTT
But on a 1,25+ vid you will need mostly more vcore...
Cheers
Sergio
Hey guys, Long time No see.. How is everyone.
I Finally got a EP45.. time to see what this Q9650 can really do on a good board instead of my 750i.. lol.
Question, Does ANYONE have a Q9650 they'd like to Sell. I'm looking for a used Q9650 to buy, as I'd like to have another one. I might get a Q9550 since the E0's are now $219 New at the egg. but I want the Q9650.
Also, I might be buying one from Ebay, or another forum if I can find what I want. So I'm wondering if any of you guys like Napalm, and Andressergio, or others that have followed this thread from when it had 30 pages.. Could tell me from what they have seen to be the best batch #'s and VID's for the Q9650. I know almost all the 45nm Q9650's are good performers. but I've noticed that the lower VID ones clock higher, but run warmer, and the higher VID's still clock good, but run cooler. But this could be wrong..
Any info on Batch# and VID's for the Q9650's would be great. as I want to get another soon.
Check Deals thread: Q9550 for $170 @MicroCenter :up:Quote:
Originally Posted by xguntherc
Nooooo don't... I got one of the newer E0s and it's nothing like what you see here - 1.4v for 4GHz and no less than 1.3v for 3.6GHz or I get blue screens and random restarts. :bs:Quote:
Originally Posted by xguntherc
If I'd known this before buying it, I'd have gone for a preowned Q9650 from the outset.
And 1.15v for 4GHz stable is just crazy, seriously :eek:
I know, I don't have a damn microcenter near me or I'd have got one already.. but thanks!
Not quite, theres a guy on here that has a 1.30vid, cant remember his name right now.Quote:
Originally Posted by humeyboy
Can you post a template of your full settings? I have a Dlx also.Quote:
Originally Posted by AgentGOD
I know this belongs in the liquid cooling section however, the guys that are running q9650's on water... what is your setup? I have no idea what to order. Im on vacation next week. My plan is to order parts tomorrow and start on it this weekend if possible. I would like something good but not toooo expensive, around $300.00 tops. Is it even doable for that price?
I pieced my kit together but this kit is basically what im using and under 300.
http://www.petrastechshop.com/pecoel.html
Nice, thanks Hoss. is this where you ordered from?
http://img3.imageshack.us/gal.php?g=image006vsh.jpgQuote:
Originally Posted by Hoss331
Ive ordered some from Petras, Sidewinders, and FrozenCPU. :up:
You may not be able to go much lower with that VID but try, 1.54pll (if I recall 1.50 would overvolt), 1.26vtt, both cpu gtls at .635, nb gtl you can probably still leave on auto for now but try .665 if you want. See if that stays stable then try to lower vcore a little more.Quote:
Originally Posted by AgentGOD
I tried as high as 1.60v PLL.. it didn't help reduce vcc.
Im not saying it will, just that 1.50pll and auto used to overvolt. Try the gtl settings.
Tried it, still spat out errors at 1.2875v.Quote:
Originally Posted by Hoss331
I'm running the Apogee GTZ with a Feser TFC 360 Rad, and a D5 Vario pump.. that is one of the better setup's unless you want a better pump, or a better cpu clock like the Heat Killer 3.0, but for Solid performace and decent price, the GTZ and the D5 and a Feser rad are a solid bet. you could go cheaper with a TC PA120.2 or 120.3 radiator, but i really like the feser, or get a XSPC RX Rad as well. good luck.
Also, will some of you Q9650 owners post your Batch #'s and VID's. as I'm about to get another one and I'm trying to choose a good one by batch number. I'm wondering if the older (First released chips) will be better, just like the first Q6600's are better than the latest ones that have high VID's. as I was offered a L844B Q9650 for a GREAT price, but it's a recent batch number, I guess there's no way to guess the VID tho.
I also got offered a L845B, you guys know of some best chance batch numbers?
Mines an L845B, does 4 GHz @ 1.3125v 100% IBT stable. VID is 1.2750v.
oh wow... so you have a L845 and it has a 1.27 VID.. scary, sounds like it still clocks just fine tho. 4 Ghz with 1.31v is fine with me.
hm.. well I got offered a L844 for $260 shipped with he thinks is a 1.25 vid. so i might take that, not sure yet tho
Take no notice of VID, seriously - the best Q9550 I've seen on the forum has exactly the same VID as mine (1.25) and it can do 4GHz on 1.21v, and 4.25GHz on 1.31v. Clearly the VID value is :bs: in many cases since these CPUs can run stock probably at or below 1v.
I've got an odd problem with mine....it's @ 4Ghz at 1.30825v in BIOS (1.290v actual), and it'll pass OCCT Burn Test, SuperPi, IntelBurnTest for HOURS - yet games still cause it to hard crash - what gives?
I had exactly the same problem with my E8500 @ 4.4GHz - solid in every stress test, but games made the system crash. I resolved the issue by lowering to 4.3GHz.
It's pretty much a given that 3D apps stress the system in an entirely different way and of course the usual software (OCCT SPi Prime etc.) does nothing for 3D stability, so it must be a PCI-E thing, whether bus frequency or the GPU itself.
Thing is, it didn't do it with my E7200 @ 4Ghz, although that was at 421x9.5, rather than 445x9.
Try prime95 Blend test.
If you are successfully passing hard stress levels on IntelBurnTest but failing Prime95 blend test, it is an indication that your memory/northbridge is unstable at your current settings. I personally use four instances of Memtest for Windows for dead accurate memory testing.
Now I'm convinced I got scammed on ebay - ten days after buying the Q9650 the seller won't contact me and nothing has come in the post. :down:
So I just bought a brand new & sealed one from a reputable retailer. The FPO # is L845B636. What are the chances it'll be better than my Q9550 E0 which needs >1.4v for 4GHz? Does anybody have a suckage Q9650? :D
Just thought I would post some BIOS settings for the Gigabyte UD3 series. I was hitting a FSB wall around 400, but after tweaking the various voltages a bit, it runs stable at 500FSB.
http://www.burr-oc.com/wp-content/up...4-1024x640.png
http://www.burr-oc.com/wp-content/up...3-577x1024.jpg
http://www.burr-oc.com/wp-content/up...3-577x1024.jpg
theres no a suckage/bad q9650.. even the worst one out there can do 4.5ghz ;)
the suckage is your mobos (anus that is)
and that goes to three of you: LennyRhys/AgentGOD/Kai Robinson
get ep45 if you guys want 500x9 out of your q9650
I'd rather not have to pay out another $200 (equivalent) for a motherboard that may, or may not help.
I'm not looking for 500x9 either, 445x9 is my limit.
Quote:
Originally Posted by xguntherc
Xguntherc: there are some batches that are a tad better then others, especially when your buying a new Q9650 (small box). But in a good board and the right settings they will perform great like NapalmV5 said. Look at my chip and Sergio's I had both on the same board and achieve loads of good results with them.Quote:
Originally Posted by LennyRhys
Lenny: I guess you haven't read the entire thread but there are loads of Q9650 better then this one you pointed out.
Here's Mine:
Q9650
FPO/BATCH L820A930 (old ones)
VID 1.2125V
Q9650 @ 4500Mhz 1.32vcore / 1.34vtt / 1.5pll 1.4mch:
http://i43.tinypic.com/34huzhe.jpg
Q9650 @ 4700Mhz 1.39vcore / 1.4vtt / 1.5pll 1.6mch:
http://i42.tinypic.com/ae1dol.jpg
http://i40.tinypic.com/uc1s3.jpg
exactly.. mobo more important than vid
Jor3lBR - congrats on your 7 gold cups :toast: awesome!