Just got QX9650 ES O0

Specification :

Processor QX9650 ES (thanks to Intel Indonesia)
MB Gigabyte GA-X38-DS5 (F4B BIOS) know as X38-DQ6 by CPU-Z
VGA Gigabyte GeForce 8800GT 512 MB
Memory Team DDR2-1200 2 GB Kit (thanks to Foxhound)
HDD WD Caviar 320 GB SATA II
PSU Gigabyte Odin Pro 1200 Watt
Cooled by Thermalright Ultra Extreme 120 + Thermaltake 12 cm Smartfan
Memory Cooled by Corsair Memory Cooling




On air result @ 4.4 GHz vcore 1.72 volt

80K++ on 3DMARK 2001 SE



45K ++ on 3DMARK 2003



25K ++ on 3DMARK 2005



16K ++ on 3DMARK 2006 (CPU clock only stable @ 4.63 Ghz)




Super PI 32M @ 4.3 Ghz (update PI 32 @ 4.4 Ghz will publish soon)




to be continued.......

This post has been edited by Lodewijk: Dec 9 2007, 10:34 AM

4.4Ghz on air?? nice

Great review.

wow QX9650 ES 4.4ghz..

but the vcore applied made me

don't worry, penryn cool enough . Can you imagine Q6xx using that vcore on air ??? hahahaha

This post has been edited by Lodewijk: Dec 8 2007, 11:26 PM

d@ng it.. those word would bring me to throw my kentsfield

tha's why la... at first... wow... 1.7vCore, that's too much, how hi/lo the idle/load temp?

it's not really 1.72 volt needed for 4.4 Ghz.

Here's the situation :

- X38-DS5 vcore option above 1.6 volt increment from 0.0025 change to 0.05v (1.65, 1.7, 1.75, 1.8 etc)
- Vdroop on this mobo around 0.03, and real vcore detected by CPU-Z/Everest difference with BIOS setting around 0.03 too. Example : you set 1.65 volt in BIOS, the real vcore you get is : 1.65 - (0.03 + 0.03) = 1.59v
- When i need 1.68 volt for CPU, i can't set the vcore 1.7, because the real vcore applied will be 1.64. So i must go 1,75 volt on BIOS, get real vcore 1.72 and vdroop up to 1.68 - 1.69. This is the real vcore for the CPU.

About idle temp around 31 - 35 degree and full load reach 47 - 50 degree

Hmmm.... but still high vcore wor...... what's the default vcore?

I wonder whether these new 45nm proc can stand high vcore or not.... you know, narrower space, increase risk of electromigration, etc....

BTW, the temperature is cool!!!

Default vcore detected is 1.25 volt Time will answer your question about that.... i can't answer it now...

I see. Could you give some OC capabilities of this proc on each vcore setting? E.g. Stable OC at 1.3v, 1.35v, 1.4v, 1.5v, or something like that...... wanna know how they OC on lower vcore....

Edit : Oh ya, what's your vdimm setting?

This post has been edited by kmarc: Dec 8 2007, 11:57 PM

okay if i have time i'll do that... since i want to see this baby doing with dry ice first.

My dimm setting for all the benchmark 2.55 volt.

1.7vcore idle temp around 31c?
but the retail price of QX9650 is killing

Q9450 for most people.

Yeah, my next target is the Q9450...

Best bang for buck....

nice clock benny~! btw while the penryn seems to be a very2 good ocing chip, i wonder why it seems like clock per clock it has same performance as the current quad core..

maybe the extra l2 cache is too overkill .. currently...

if so, people who is not intends to oc extremely, might not have benefits migrating from kentsfield to penryn..

btw i hope the retail one also can hit as high as the ES.. or perhaps better than that!!

1.7 vcore and that kinda temp? jesus imma throw my kentsfield now lol

yap Q9450 best bang for buck... but there's a problem if you want to OC it...the multi is too low.. you will need very very good MB to reach higher clock with it.

---

Added on December 9, 2007, 8:16 amthanks irenic

Retail QX9650 is a monster !!! Can do 4 Ghz only with 1.25 -1.35 volt...and i saw somebody can reach 4.7 GHz with only 1.68 volt.

This post has been edited by Lodewijk: Dec 9 2007, 08:16 AM

wad i can say...

A few reviews showed that the wolfdale is, on average, faster than conroe by about 5-10%. Can't remember about kentsfield vs yorkfield though.... IIRC, the most benefit would be those apps that take advantage of SSE4, especially video apps....

Oh ya, forgot about the CPU multipler. Hmmmm.... actually not a problem for me as I'm not an extreme overclocker. Currently using the E6750 with 8x multiplier too....

Okay final test... game movies ! load some demos on it.. get movies

sorry im blind

This post has been edited by alieamin: Dec 9 2007, 03:45 PM

yeah the difference is very big when comparing kentsfield's b3 vs yorkfield..

wow so low vcore .. i guess next year 4ghz gonna be a 24/7 standard for all ..

yup the SSE4! forgot bout it.. nice feature and surely benefits those who works with video editting ,,

This post has been edited by irenic: Dec 9 2007, 04:45 PM

4ghz on ~1.3 vcore seems too fake to believe with the current mainstream hardware.

and imo the Q9450 shouldnt be a problem to most so long they arent extreme overclockers, im using 355x9 and the default multiplier is 10 and the core extreme mainly unlocks those higher up multipliers that arent used by most

eitherway the oc results are just too impressive be it achievable clock speeds, voltage requirement, and most importantly temperatures.

with all the current clock speed advancements on intel's core chips, one would think that they are returning back to their old P4 clockspeed race? cuz the main thing thats really advancing since dual/quad cores are clock speed. wonder if they could come up with some kinda tech to ultilize the multi core procs even if the application doesnt ultilize it. eg:
program only uses 2 core, the processor's 10 cores would then divide up the load between all 10 cores, without assistance needed from applications/OS.

newegg's selling 1.3k usd



btw,

mind to test 1M super pi?

Dry Ice short test with Gigabyte GA-P35T-DQ6 and Team Xtreem DDR3-1600 2 GB kit.

This CPU not too good, since it need more and more voltage....enjoy the pictures.

3DMARK 2001 84K




Super PI 1M < 10s





Super PI 32M < 10m

really soon 4ghz on air will be standard liao

just to ask, can all the 4 cores goes at that speed ?? but it seems promissing...... hehe

q9450 has how many multipliers ?? 333*7.5 izzit ??? go 4 ghz u just need 533 fsb hehe........ 1066 ram would do fine but mobo i dunno lar

mine q6600 stuck at 3.3 due to core 2 and 3........ pump more then 1.45 already still not much improvement over dual orthos, orthos over the core 2 and 3 often fail, but strangely super pi all this goes find

Current Intel CPUs can't throttle each core separately. That means that all the 4 cores are running at that speed.

The Q9450 has a 8x multiplier, just like the E6750.....

SuperPi doesn't test for CPU stability like what orthos does.....

8 ar? i tot 8.5 lol...this mean 800mhz ram wont do well?

this review is poisoning me from getting abck to use my coming proc to boil egg again

but too bad it's too cold to boil an egg.

nice benny... still pushing it i see... congrats on succesfull indo oc event.. and congrats here as well

thanks bro.... i think if you got the QX9650 you can beat my score easily

i doubt benny really push this proc to its limit.. :-"

So it's possible, but what about OCCT / Orthos ? and yes, this is what we call overclocking -- spending huge amounts of money to get little or not real-world benefit -- WTG but I'd still have to see stability

That 5 GHz processor alone draws 465W

This post has been edited by wodenus: Dec 16 2007, 10:19 PM

I think most people these days just post a screen completing PI and some benchmarks program. As a conclusion, people are posting up what's the maximum Ghz the CPU can boot on and not the actual stability of it

I seriously respect people who can spend an extra Rm82 a month just to look at some really good numbers I hope we get 6 Ghz soon.. the processor alone must cost something like Rm100 a month to run 24x7


This post has been edited by wodenus: Dec 16 2007, 10:09 PM

sweet man 1M super pi under 10 seconds