Q9550-s Stable at 450 FSB x 8.5 .. can i get more ?

apoppin

Lifer
Mar 9, 2000
34,890
1
0
alienbabeltech.com
i got the new 65w q9550 e/s from intel and i am trying to get it as high as possible without frying it

it was rock solid at 3.6Ghz at default VCORE but now i am needing to add voltage

450x8.5 is a "pass" with 1.2875v and temps as recorded by Core temp in the mid 30s at idle - reasonable

however 456 fails with 1.29375v
:(

dare i go higher
-what is "safe" ?

[there is no safe, but you OCers may know what i mean]

DDR strap is 333 so i am keeping the RAM slower than usual while i look for the CPU maximum

The CPU i test against, my e8600, got to 4.25Ghz ,, it actually stabilized at 4.3Ghz but 4870x2 runs too hot to keep it there

rose.gif

 

Acanthus

Lifer
Aug 28, 2001
19,915
2
76
ostif.org

Elusiv1

Member
Jun 29, 2005
129
0
0
My C1 needed 1.3625 for 4Ghz stable, i had a problem trying to find a sweet spot after 3.83ghz with this chip until i found the right GTL's and Northbridge voltage..
 

Elusiv1

Member
Jun 29, 2005
129
0
0
My memory is also running 1:1 on the 333 strap, i'm using the 2.00B strap multiplier. So right now my memory is @ DDR2 942 in the bios, the tighter timings got me stable at 4ghz..
 

apoppin

Lifer
Mar 9, 2000
34,890
1
0
alienbabeltech.com
Originally posted by: Acanthus
Originally posted by: apoppin
Originally posted by: Acanthus
according to the intel spec sheet, you are safe up to 1.3625v with E0s

this is the new "s" q9550

http://processorfinder.intel.c...tails.aspx?sSpec=SLGAE

damn, i guess you are right

ok, i am trying 455 at 1.3

i guess the main thing is to watch the temps

4.0 Ghz or bust !!
:D

Lemme know your results, im thinking about going Q9550 or Q9650.

Well, it is gonna be a 50 page review .. split into several parts
- you don't want all the results posted here :p

But my OC is good, so far .. i guess i was even "undervolting" it with great results at 3.6Ghz

So far it *passed* 459 FSB .. so 3.9 Ghz .. it is catching up on my e8600! and i am testing at 460 where it is also stable
- i am wanting to see where it needs the next voltage boost. The last one - to 1.3v came at 457FSB [and it is one 'notch' up at a time] .. i am *testing* it, see

- i am REALLY hoping for 4.0Ghz as it would make the perfect comparison for a quad-core vs dual-core shootout. i am using a custom 16-gaming benchmark suite; mostly DX10 games at 19x12 and 16x10 with 4870x2/4870 and GTX280
rose.gif


it is taking a [long] while as i stopped to OC my 9550 and i am inching the FSB +1 at a time .. and noting temps and results
- it is just starting to get warm .. mid 30s on all 4 cores
-- so far .. i spent over 50 hours with it since Thursday night - including overclocking e8600 and benchmarking results with three video cards at 4 and 4.25Ghz
:Q


Thanks for all your overclocking suggestions .. i think i will get very close to 4.0Ghz as it is still not getting hot

461 got the BSoD, so i am going up over 1.3v
EDIT:

it is 2.2v on the PC8500, right ?
- it 924Mhz so it should be fine

... a little higher on the voltage?
:evil:
 

Acanthus

Lifer
Aug 28, 2001
19,915
2
76
ostif.org
I push 60c loads on this 65nm beast.

Ive had it at this OC for over a year now, and im out of the safe voltage range entirely.
 

francisA

Member
Dec 2, 2008
153
0
0
Go for it!!! :)

I had it at 475x8.5 (4.038GHz) @ 1.296v actual. VID is 1.250v. That's stable 16hrs Prime, 2hrs OCCT and 30 IBT. I didn't go any further on any stability tests as I personally think it's sufficient for what I use it for.

I'm currently testing 485x8.5 (4.123GHz) @ 1.328v actual. So far it's been 2hrs OCCT, 20 IBT and 10hrs Prime stable.

Going for 4.2GHz next.

I tried to shoot up to 500x8.5 for 4.25GHz @ 1.344v, went into XP, got normal stuff ok but hung when I did Prime just after 4mins LOL :)
Did 500x6 but it did the same thing, albeit running Prime a little longer but hung nonetheless.

My G.Skills are only rated DDR800 so not sure if that's what's causing the problem. But I'll up the fsb higher slowly until it hit that wall.

No doubt about it, you can still go higher with the Q9550-s
 

apoppin

Lifer
Mar 9, 2000
34,890
1
0
alienbabeltech.com
i got it .. it is 470 FSB at 1.3375v .. or 3.995 Ghz
- the temps started to rise a couple of degrees starting with 469 FSB where it needed a notch more voltage

that is where i am going to test it vs e8600 as it leaves me "wiggle room" if there is later instability to up the voltage a bit more

i don't know if it will go much higher .. there is no advantage with an engineering sample anymore
- and now the voltage is at 1.3375 where i am beginning to get nervous
:eek:

of course i will test it a bit more, but may settle on 4Ghz for 24/7

that is a decent OC for a quad core and a decent match for 4870 Crossfire-X3 graphics
rose.gif


and i go for "gaming stable" .. if i can play ALL of my games for hours on end with no issues .. for days
- it is "solid" imo
 

Gillbot

Lifer
Jan 11, 2001
28,830
17
81
first thing I do is set mine to 1.4v core, 1.38-1.39 vtt and see how high it'll go, then I tweak the voltages down from there.
 

apoppin

Lifer
Mar 9, 2000
34,890
1
0
alienbabeltech.com
Originally posted by: Gillbot
first thing I do is set mine to 1.4v core, 1.38-1.39 vtt and see how high it'll go, then I tweak the voltages down from there.

that is why i go the other way :p
.. didn't you have really bad luck with your quad's OC ?

i started at 3.6G .. then inched up all the way .. it *appears* to be OK at 471FSB with 1.3375v but i may have thermal issues with 4870x2 or CF-x3 .. i got my e8600 to 4.3 but backed it down to 4.25 for practical reasons

there was ZERO measurable difference between 4.25Ghz and 4.3G with any GPU - 4870, x2, CFx3 or gtx280

and now i got to run 16 benches; make sure it is stable and watch my temps
rose.gif



you guys think 4.0Ghz with 1.3375v is OK for 24/7 .. it is idling in the middle-upper 30s in a warm room
??


471 is stable .. 472 requires even *more* voltage .. i don't think i want to go there
 

Gillbot

Lifer
Jan 11, 2001
28,830
17
81
I run mine 24/7 at 467fsb. I think I could get more but my mobo has limited options and is more suited for clocking a dual core. With out GTL ref voltage adjustments, it's a crap shoot playing with settings and hoping for th best.
 

geokilla

Platinum Member
Oct 14, 2006
2,012
3
81
I never changed anything when overclocking my E7200 besides the vcore. I don't think you guys need to change anything either...

@apoppin. I think 1.35V under OCCT load would be good enough for a Q9550s. Anymore and I'd worry about frying the chip. Just make sure that the temps are at or under 65C.
 

TidusZ

Golden Member
Nov 13, 2007
1,765
2
81
Originally posted by: francisA

No doubt about it, you can still go higher with the Q9550-s

I dunno, mine's an E0 Q9550 without the s, and I got 4.25 stable without much effort at 1.392 actual. Got another q9550 no s in the mail right now too, building for my brother. I'd bet dollars to donuts I get it to 4+. Will let you know.
 

apoppin

Lifer
Mar 9, 2000
34,890
1
0
alienbabeltech.com
Originally posted by: TidusZ
Originally posted by: francisA

No doubt about it, you can still go higher with the Q9550-s

I dunno, mine's an E0 Q9550 without the s, and I got 4.25 stable without much effort at 1.392 actual. Got another q9550 no s in the mail right now too, building for my brother. I'd bet dollars to donuts I get it to 4+. Will let you know.

it's like a "who cares" for me

it all becomes bragging rights after a certain point. What practical difference does it make - especially if you are a gamer?
- there is a "sweet spot" even for CrossfireX-3 and it is around 3.6 to 3.8Ghz. After that, we wait for faster GPUs and move to Core i7 :p

i ran my e8600 at 3.99 and then again at 4.25 and there is a very tiny FPS difference in 13 games i tested; it is unnoticeable in most cases - that is partly what i am writing my review on.
rose.gif


Originally posted by: Gillbot
I run mine 24/7 at 467fsb. I think I could get more but my mobo has limited options and is more suited for clocking a dual core. With out GTL ref voltage adjustments, it's a crap shoot playing with settings and hoping for th best.

Finally, i had to move my voltage up a notch and my FSB down a notch .. in fact i dropped it two so i can match it *exactly* with my e8600 at 3.99Ghz .. if not, i still have to go up ANOTHER notch for 10mhz ,,

so again .. this has turned out perfect for me

@apoppin. I think 1.35V under OCCT load would be good enough for a Q9550s. Anymore and I'd worry about frying the chip. Just make sure that the temps are at or under 65C.

well, it is at 3.99Ghz - 469 st 1.344 .. so that is 1.35v to get 10 more Mhz; it is symbolic and i achieved it and now it is back to where i am comfortable with it [and it fits perfectly - by serendipity - to match the 16 benchmarks i just ran on my e8600 also at 3.99Ghz so i don't have to rerun them at all!!]

---the only thing i do not understand

why is the "S" 9550 priced $100 more than the regular one - it isn't like you are saving 30w 24/7 or that it overclocks much better than the regular one

:confused:

anyone care to explain the pricing rationale


thanks for all your comments .. i got what i wanted
:heart:

 

apoppin

Lifer
Mar 9, 2000
34,890
1
0
alienbabeltech.com
Originally posted by: Elusiv1
I though there were only two steppings, E0 and C1??

yes

but intel raised the price and lowered the vcore on the Q9550"S"

look at the main site's latest article :p

so tdp is now 65w instead of 95

i now have a e8600 for sale that does at least 4.25Ghz
:D


anyone know why "s" is so much more expensive ?
 

SickBeast

Lifer
Jul 21, 2000
14,377
19
81
Originally posted by: apoppin
Originally posted by: Elusiv1
I though there were only two steppings, E0 and C1??

yes

but intel raised the price and lowered the vcore on the Q9550"S"

look at the main site's latest article :p

so tdp is now 65w instead of 95

i now have a e8600 for sale that does at least 4.25Ghz
:D


anyone know why "s" is so much more expensive ?

My guess is that they are marketing it toward server builders who need something that does not generate alot of heat, and uses very little power. Quite often in a rackmount server you're forced to go with a low-profile HSF and you need a CPU with cool runnings. Server chips tend to be sold at a premium.

It's quite possible that they take these chips right from the center of the die in order to be able to validate them to run at a lower voltage. It's quite possible that intel would have otherwise sold these chips as much faster models with a higher voltage spec.
 

Elusiv1

Member
Jun 29, 2005
129
0
0
G damn, and i've been fighting with the voltages on this C1 for the longest time!! I could probably hit 4.25 stable with lower voltages on my UD3P with that thing, i'll never know because i don't plan on swapping anything out at the moment.. My next upgrade will be an i7 or whatever the hell else they come out with in 2 months!!lol i8 or i9??
 

apoppin

Lifer
Mar 9, 2000
34,890
1
0
alienbabeltech.com
thanks SB
-that makes sense that it is aimed directly at OEM, not the enthusiast. i didn't think they necessarily overclocked better, and as always, it is the luck of the draw. And of course, i did not got for Max OC, as i do not want to overvolt; my 3 GPU cores generate a lot of heat and it would probably OC a little better in a cooler room, case and climate. The CPU cooler is fine; heat does not seem to be an issue. Every time you come to a roadblock, a little voltage boost would stabilize it; i like that about this CPU - a lot.

yes, this is clearly my last Penryn upgrade .. but i am waiting for the next gen of i7 MBs and more reasonably priced DDR3 that actually makes a performance difference over DDR2 that is not theoretical in most cases or on synthetic tests
- what i really want to do is bench i7 vs penryn just for gaming

until then, imo, 4.0Ghz quad is a perfect match for today's fastest videocards; i am doing Q9550s' benching - right now - at e8600's same clock [and at 4.25Ghz, +260Mhz higher] to compare quad vs. dual core gaming at 19x12 and 16x10
--i will see you guys later. i feel confident in my OC without frying my CPU. Thanks!

rose.gif
 

taltamir

Lifer
Mar 21, 2004
13,576
6
76
i got the new 65w q9550 e/s...
450x8.5
isn't the whole point of getting an engineering sample to have an unlocked multiplier so you can OC it by raising it without increasing the vcore?

Anyways, check the temps on your northbridge. I found them to be problematic in my case.
 

apoppin

Lifer
Mar 9, 2000
34,890
1
0
alienbabeltech.com
Originally posted by: taltamir
i got the new 65w q9550 e/s...
450x8.5
isn't the whole point of getting an engineering sample to have an unlocked multiplier so you can OC it by raising it without increasing the vcore?

Anyways, check the temps on your northbridge. I found them to be problematic in my case.

not any more

http://www.techarp.com/showarticle.aspx?artno=407

thanks for the tip; this ASUS P5e deluxe has some extreme passive cooling on the NB, i believe and i make sure there is some air circulation - 3 HD4870s tend to roast my case anyway
:sun:
 

Acanthus

Lifer
Aug 28, 2001
19,915
2
76
ostif.org
I have that same passive cooling system with heatpipes on my P5K deluxe.

Temps do get relatively high, but there are no stability issues.