[Team.AU] Gigabyte GA-EP45T Extreme - P45 DDR3 Monster

Just got my hands on the pre-release version of Gigabyte's OCing Extreme board.
This is the board containing as much of the goodies we've been asking for that they could pack in.

http://img.photobucket.com/albums/v3...IMG_3933-1.jpg

After about 30 mins of familiarising with the BIOS and taking photos, i jumped straight into some FSB clocking :up:
And just like what we've come to expect from various P45 boards, this one clocks like a beast too :eek:

645fsb on stock air, no tweaking (only 1.35V cpu, and "2.66D" RAM ratio).
This CPU would never budge past 540fsb on X48 chipset.

http://img.photobucket.com/albums/v310/T_M/IMG_3976.jpg

Here's some juicy photos :clap:

http://img.photobucket.com/albums/v3...IMG_3937-1.jpg

http://img.photobucket.com/albums/v3...IMG_3941-1.jpg

http://img.photobucket.com/albums/v310/T_M/IMG_3944.jpg

http://img.photobucket.com/albums/v310/T_M/IMG_3947.jpg

http://img.photobucket.com/albums/v310/T_M/IMG_3949.jpg

http://img.photobucket.com/albums/v310/T_M/IMG_3951.jpg

http://img.photobucket.com/albums/v310/T_M/IMG_3952.jpg

http://img.photobucket.com/albums/v310/T_M/IMG_3953.jpg

http://img.photobucket.com/albums/v310/T_M/IMG_3961.jpg

http://img.photobucket.com/albums/v310/T_M/IMG_3962.jpg

http://img.photobucket.com/albums/v310/T_M/IMG_3963.jpg

http://img.photobucket.com/albums/v310/T_M/IMG_3965.jpg

Will update further with more FSB clocking results :yepp:

yep looking forward to this one, i think its going to be the awesome 3d platform for e8600 and single card benching =)

Nevermind E8600, with an E8500 and this kind of FSB you can easily max out the chip finally :D

Ok, that's it, I'm sending an e-mail :D
Thanks Tim, premature PCB and/or BIOS or not, it certainly looks good :D

Not premature, just pre-release.
AFAIK they are still working hard on RAM clocking.

Do Want! Sending an email out myself, DDR3 + P45 = win!

wow.. this board looks sexy :D
waiting for some Good results :up:

just got mine today, but the BIOS is D8 ???

Congrats mate!!!
Have a DDR2 "Extreme" version with me and that did 630 easy. Need good ram to get higher now.

Benny you sure you got the DDR3 version?
That BIOS code looks weird, drop a line to the usual guys for a BIOS update.

645FSB by auto! looking sexy board for FSB OC :up: wait to see more update Tim
can't wait for tried it :D

very nice

i am starting to have a detailed look at the ddr2 version of the board in Intel section :D :up:

Quote:

---

Originally Posted by T_M

Benny you sure you got the DDR3 version?
That BIOS code looks weird, drop a line to the usual guys for a BIOS update.

---

yes definitely DDR3 version.. and i just got the F3A BIOS. D8 ussually BIOS for pre-testing, very early... put E8400 to test it now :rolleyes:

btw tim it's a little unclear
did you boot into windows at 645MHz FSB?
how stable was it?

Nope, havent even installed windows yet.
This was all done in 1hr after unpacking it.

At 666fsb, the board failed totally.
At 650fsb, the board got to POST but froze at detecting drives.
645fsb POSTed and went into BIOS fine.

ok i see

i can do 630Mhz on Gigabyte P35-DS3P but windows is a different story

i really hope we can get this kind of windows stability (i have a feeling it should be ok as bob already pushed his for validation similar to your clocks :up::up::up:

let us know how your RAM clocks on it as well :D

maaan these boards are sooo nice ... i will do some tests with my Extreme board tonight :)

I think this will be my next mobo when it comes out around here. Looks awesome. Any ideas if the Corsair 1600 C7DHX will play nice with it?

really nice board Tim, and fsb clocks like everybody expected, but how about RAM clocking? Asus P5Q3 Deluxe has serious problems about this point...
btw, whats for that monster heatsink on the first pic? NB?

Quote:

---

Originally Posted by Achill3uS

really nice board Tim, and fsb clocks like everybody expected, but how about RAM clocking? Asus P5Q3 Deluxe has serious problems about this point...
btw, whats for that monster heatsink on the first pic? NB?

---

i thought this was the chipset itself that was problematic rather than any specific board manufacturers.........
i guess we'll find out soon when he posts some results :)

Wow ,really nice mobo's look and FSB. :)
There's a lot of news about voltage ( VPLL for example ) and memory settings .
You put *.**D ram setting ? I can't boot with it on the X48-DS5 ,only *.**B for take 593MHz.
You started @600MHz ? It's for that ,maybee .
I wanna see 7** of FSB . I think it's possible :D . I hope receive quickly my EP45DQ6 or EP45-DS5 with my E8400 ( 645MHz on Commando ).
Waiting for some very good results like usual from you ;)

really nice mobo. i am curious how it can clock DDDR3

Clocking DDR3 has been a problem for all companies using the P45, so yeah it will be interesting to see how it goes this week.

Nice looking board is DDR3 clocking raising a few issues, is that just poor mhz clocking or bad timmings as well i have seen early pl level problems with DDR2 + P45 does anyone have any more info on this?

Really good mobo Tim! Good luck for ddr3 clocking..

Setting CPU Clock Drive = 800mv, CPU Clock Skew = 50ps, RAM = 2.00D allowed me to POST fine at 655fsb.

http://img.photobucket.com/albums/v3.../Picture-1.jpg

In Windows im getting my best CPU-Z validations when I boot with the same settings above but 560fsb, then when in Windows change Performance Level in Memset from 3 to 4 (something weird going on with these values):

630fsb using setfsb:
http://valid.x86-secret.com/show_oc.php?id=380419

Nice results man! Cant wait for you to really punish this thing under some LN2 + 3dmark action.

Any idea of the release date on it?

630 validation in windows is promising

looking good there Tim.... i stuck on 600 Mhz FSB with E8500... :( http://valid.x86-secret.com/show_oc.php?id=380419

The performance level thing is a weird one, when left at Auto memset reports 3, but if i set 3 in BIOS it wont boot.

Quote:

---

Originally Posted by T_M

The performance level thing is a weird one, when left at Auto memset reports 3, but if i set 3 in BIOS it wont boot.

---

Yes.. i experience the weird problem on PL... mine shows PL = 1... :p:

weird
looks like they've done some strange stuff with bios

on biostar P45 memset works fine

have you tried setting to PL that it would work with .... for example at 600MHz FSB you should be able to run PL-9 or looser if chipset is struggling ....with better cooling or chipset PL-8 was not uncommon at such high FSB on P35

Yeah, i tried PL=13 on both DIMMS and it came up as 13 in memset.
But when i booted at Auto, and set 13 in memset it froze....

Quote:

---

Originally Posted by dinos22

weird
looks like they've done some strange stuff with bios

on biostar P45 memset works fine

have you tried setting to PL that it would work with .... for example at 600MHz FSB you should be able to run PL-9 or looser if chipset is struggling ....with better cooling or chipset PL-8 was not uncommon at such high FSB on P35

---

When i set PL loose i.e from 1 to 3 it's still working, but above 3 it will give BSOD instantly. So..it's really weird, and not impact at all to reach more higher FSB.

Quote:

---

Originally Posted by T_M

Yeah, i tried PL=13 on both DIMMS and it came up as 13 in memset.
But when i booted at Auto, and set 13 in memset it froze....

---

:ROTF::clap:maybe it was marijuana day during this bios writing procedure :D

they do work very hard though :yepp:

Is there a price tag on this board yet?

Tim can you please report on memory testing you've done so far

CAS5/6/7/8 clocks :up:

Memory testing so far:

8-8-8-20 2T-55:

400fsb
2.00D fine
2.40B fine
3.xxB fine
4.00B fine

420fsb
4.00B fine

430fsb
4.00B fail to POST

have you tried CAS7 and 6

Not yet

Pulled apart the cooling on the board and soldered on some volt read points.
See pictures below.

It appears that leaving PLL set to auto results in huge volts pumping into the CPU PLL when running high FSB, at 560MHz im reading 1.89V.
Turned it down in BIOS to 1.6V set (1.63V reading) and its stable just the same, so it seems too much PLL being set by BIOS at the moment.

Im not a big fan of heatpipe assemblies, so i took an old P35 cooling system, chopped off the heatpipes, and stuck them onto this P45 board :D
SB heatsink needed slight modification to the mount holes, seems about 1-2mm wider spaced holes than the P35 boards.

http://img.photobucket.com/albums/v3...Bheatpaste.jpg

http://img.photobucket.com/albums/v3...oheatsinks.jpg

http://img.photobucket.com/albums/v310/T_M/P45.jpg

http://img.photobucket.com/albums/v3...wHeatsinks.jpg

http://img.photobucket.com/albums/v310/T_M/Voltages.jpg

http://img.photobucket.com/albums/v3.../Voltages2.jpg

DDR2 board read points would be able the same i assume?


can you tell me something
do all the voltages set properly from bios
if you set all the values do they all read similar volts to what you've set in bios

my board refuses to boot past 1.5v NB at high FSB and it requires me to keep my NB volts lower than Termination volts (VTT) which is no good as that would mean that for me to go past 1.5vNB i have to shove around 1.54v VTT and any more volts in NB would result in higher VTT......i'm not going to kill my CPU just to stabilise it at 600MHz FSB

i am also testing a biostar board and they have now unlinked NB and VTT volts and i can select 1.6v NB and 1.35v VTT and be able to get 32M SuperPi stable up to 615MHz FSB with PL10 which is pretty good and fast too

the linking of vtt and nb is a deliberate thing, the biostar engineer told me that the board faces 'instability and risks' if we seperate these voltages too much, so we might need to get gigabyte to do the same thing with their bios

When set on Auto, here's a quick tabulation of what PLL and Termination volts do (measured as above) during OC'ing an E8400:

FSB -- PLL --- Term
333 = 1.570 / 1.211
400 = 1.690 / 1.212
450 = 1.794 / 1.524
490 = 1.894 / 1.525
500 = 1.895 / 1.630
550 = " / "
600 = " / "
620 = " / "

My readpoints for PLL / Term read approx 0.03V higher than BIOS set and what ET6 shows

OMG dude
1.63VTT and 1.89PLL
that is crazy

woah that is crazy seems all P45 are doing some form of auto = high CPU PLL / CPU VTT/FSB volts even Asus P5Q series from what I read...

If it's an intel spec thing, then does that mean high CPU PLL /FSB volts is okay for 45nm cpus on P45 compared to P35/X38/X48 ?

Holy cow, that PLL is nuts. I can't imagine high PLL on P45 is any different than X38/48. Dead CPU FTL.

Well, this CPU is taking it without skipping a beat on stock air idling around 45C

Wonder how long it will last at those voltages.

I'll let you know :)

according to some ppl running 1.5v VTT will kill your CPU in a day lol

so does that mean 1.63 will take a few hours

actually it would be an interesting test since you don't seem to care if you kill the chip hehehehe

i have a feeling it won't be as bad as it's been made out to be

p45 for oc'ers look awesome but those volts :shakes::shrug:

Quote:

---

Originally Posted by dinos22

i have a feeling it won't be as bad as it's been made out to be

---

Dont let chicken little rule your world :up:

Quote:

---

Originally Posted by T_M

Dont let chicken little rule your world :up:

---

:up::D

anyone use the Gigabyte DDR2 cheaper x48 boards curious how they perform as well?

That doesnt really belong in this thread dude.

Quote:

---

Originally Posted by Glow9

anyone use the Gigabyte DDR2 cheaper x48 boards curious how they perform as well?

---

Its called the search function ;)

First round with a bit of LN2 @ -100C
Efficiency was not an aim, hence the RAM timings.
But i know for sure it ran the same MHz at 6-5-5-14 1T but i ran out of LN2 for higher multi.
Pumped more vCPU than i needed, and manually set VTT and PLL to 1.72V (measured) and 1.92V (measured) respectively.
Vdroop is a big problem when running 32M, but 1M hardly anything.
NB cooled by P35 heatsink on air.

Wanted to test what a real benching system might handle, hence two sticks of RAM and two cores.

http://img.photobucket.com/albums/v3.../585223MHz.jpg

http://img.photobucket.com/albums/v310/T_M/6351M.jpg

good start mate

cool :) so VTT 1.72V and PLL 1.92V wow looking very high :eek:

Wise words from Sam...

vpll > 1.70V will make it degrade then die
vtt >1.45V is 100% guaranteed death for quad 45nm

nice board and clocks to:up:. did 2.4v of pll to a e6300 on dice and it didnt like it lol. it was fun while it lasted:rolleyes:

Good Start mate :)

Awaiting my board!!

Nice board, but crazy voltages there....

i'm awaiting your high VTT endurance test results :rofl:

Can you log how many hours at what voltages just in case the chip does succumb? I'm hesitant in sticking my QX9650 into the P45T-Extreme sitting here before I do my 4850 + X48T benching :eek:

good start T_M !!!.....the voltages are nuts !

Quote:

---

Originally Posted by Oliver

Wise words from Sam...

vpll > 1.70V will make it degrade then die
vtt >1.45V is 100% guaranteed death for quad 45nm

---

lol you shouldn't say this Oliver

btw T_M, have you tried the new F3B BIOS ? Hope it can fix 2T/1T and gain the FSB. I don't have time to test it now. You can download the BIOS here

Going to use my 65nm chip to start, will check out the F3B BIOS.

Quote:

---

Originally Posted by TheKarmakazi

Its called the search function ;)

---

Seems relevant, I don't know too many other review sites that don't compare motherboards. Why pay for a premium board if it I dont have something to compare it to? This seems like a damn nice board

Since when was this thread a review site? Please take your X48 queries to an X48 thread (search!)

I completed one run at 540fsb (8.110s) but cold bugged it foolishly (-110C)
After that i couldnt repeat it.

Time log.... hmm... i'd say last night the chip ran at those volts or similar for about 1.5-2hrs (love how little LN2 it takes on dual!)
Add to that all the air time ive spent at the auto settings > 600fsb.

Nice run dude. I'm awaiting by board. I have P45-DS3R here. sucide screen was at 653 fsb with air cooling. used E8400. The CPU still lives after 2,2 vpll and 1,54 CPU termination. it was on air.

Try opening up the Termination, it seemed to be a bottleneck on FSB for me

Tim
keep your NB volts close to VTT volts
and you will get max FSB

NB volts slightly lower.... like 1.5NB >> 1.54VTT sort of difference

Will give it a shot on air, no more LN2 runs until after July :(

haha i love tim, not afraid to throw the volts to get the job done, well done very promising mate,

ill be using your thread as a baseline when i recieve my board later this week... hoping for some high fsb 32m actions...

Looking very promising this board too ,Tim what PL were you using to boot in Bios when you ran the cpu on LN2 ? If you remember.
Keep pushing it bro ,:up: Poor E8400 :rofl: , CPU-Z screenshot shows 2.0v:eek:

Quote:

---

Originally Posted by T_M

First round with a bit of LN2 @ -100C
Efficiency was not an aim, hence the RAM timings.
But i know for sure it ran the same MHz at 6-5-5-14 1T but i ran out of LN2 for higher multi.
Pumped more vCPU than i needed, and manually set VTT and PLL to 1.72V (measured) and 1.92V (measured) respectively.
Vdroop is a big problem when running 32M, but 1M hardly anything.
NB cooled by P35 heatsink on air.

Wanted to test what a real benching system might handle, hence two sticks of RAM and two cores.

http://img.photobucket.com/albums/v3.../585223MHz.jpg

http://img.photobucket.com/albums/v310/T_M/6351M.jpg

---

EVIL :D

Great work T_M.

I 2 have Coldbug at -110

Migth just be the caps, anybody have a clew what to change them with ?
-110 SUX :D WANT TO GO LOWER :D

EDIT: Btw if any readings with oscilloscope is needed please say so.
I have access to every measuring tool in the world here at work!

PL was Auto all the time, and in Windows after booting at 620 it was reading in Memset as "4"

Great work Tim, get an 8500 in that board next! 650FSB on 8500 get some really nice spi!!

Thanks for the response bro :) , Auto PL in Bios worked for you.
Keep Pushing it Tim , great job.

Thx for sharing.

Will give the board a spind again in the weekend :)

I m testing EP45-Extreme but DDR2 version.

And after few boot , i saw that that board needs alot of VTT to boot to 600mhz !

Edit :

yes it loves VTT :lol:

http://valid.x86-secret.com/show_oc.php?id=384142

http://valid.x86-secret.com/ccpuz.ph...4142&nocache=1


Edit :

CPU in air and NB also :D

Memset locked ???

How about this M/B Ram-OC?
Can it 2000+Cas7 stable

No, DDR3 RAM clocking on P45 is still a serious problem

Quote:

---

Originally Posted by Pt1t

I m testing EP45-Extreme but DDR2 version.

And after few boot , i saw that that board needs alot of VTT to boot to 600mhz !

Edit :

yes it loves VTT :lol:

http://valid.x86-secret.com/show_oc.php?id=384142

http://valid.x86-secret.com/ccpuz.ph...4142&nocache=1


Edit :

CPU in air and NB also :D

Memset locked ???

---

yeah it's locked for me too

how much VTT did you use EEEEEEK :D

I am all over this board!

Wonder if this place actually has it: http://www.alwayslowest.com/AL/index...s&part=2869838

Anyone have any clue when this is going to be out?

-Collin-

Quote:

---

Originally Posted by T_M

No, DDR3 RAM clocking on P45 is still a serious problem

---

Worse than P35? Hit DDR3-2000 at 7-7-7-20 on my Blitz Extreme. Think this board will be better?

-Collin-

If and or when the kinks are sorted out it should be capable of those standard sorts of OC's.
But for now I cant even get more than 800MHz very stable, and theres serious divider issues.

Quote:

---

Originally Posted by NinjaZX6R

Worse than P35? Hit DDR3-2000 at 7-7-7-20 on my Blitz Extreme. Think this board will be better?

-Collin-

---

i am pretty sure that p45 ddr3 will rock. but when? nowadays 1333 is faster than any higher freqs. weird. i don't think its mfc's bios related rather intel related.

BIOS version f3b is the first to make this board DDR3 capable :D

First RAM testing:
900MHz 7-6-5-15, 2.1Vdimm

http://img.photobucket.com/albums/v3...007-6-5-15.jpg

Quote:

---

Originally Posted by T_M

BIOS version f3b is the first to make this board DDR3 capable :D

First RAM testing:
900MHz 7-6-5-15, 2.1Vdimm

http://img.photobucket.com/albums/v3...007-6-5-15.jpg

---

:banana::banana::banana::banana:ing sweeeeeeeeeeeeeeeeeeeeeeet

Just doing some quad FSB testing and 460 was stable with no tweaking.
Pushing for 470 now.

Tim, how about 2T/1T issue ? It seems no changing in F3B BIOS.... but this BIOS change a lot in memory performance i think

Yep still no change between 1T/2T, even if you set it in BIOS.
Memory performance of 4.00B is good, but at the same settings with 3.20B or even 2.44B it fails to post.

Bumped up vMCH manually to 1.35V (read), locked the VTT and PLL at the same settings as Auto 450fsb (1.52 and 1.795 read) and managed this:

http://img.photobucket.com/albums/v310/T_M/476fsb.jpg

Wow amazing stuff :)
I'm loving this board after endless problems with the EP45-Extreme, switching to this one and reading here on XS gave me hope.
Can boot windows at 620 fine, looking to push to FSB's like you have. (Still on F3a)
Excellent clocks T_M, this is mt P45 board of choice :clap:

why does memset lock me out of being able to adjust anything?

Great info in the thread mate!! :up:
I will give my E8400 and EP45T-Extreme LN2 action soon :D
Thanks for the BIOS too ;)

@ Thomas, what VTT were you pushing on EP45-Extreme?? My Xtreme board can do 570FSB boot with 1.36V and ET6 makes 590 possible nothing higher. Same chip does 590 in Ep45-DS3R and 610 with ET6/set-FSB :s

@ strange|ife - no idea man, no problems for me using the latest beta available on XS

Reset a few BIOS settings and got the board to successfully run 2.40B so played with some more FSB on quad.
Without much trouble got it doing CPU-Z to 515MHz on 4 cores, and 2M at 495MHz

board looks okay to me
using c8 but just seeing what it can do of course, not bad ram speed being reached on a kit of mine that isnt the strongest ever, but 1050 on the trams with 630 fsb and low pll and vtt used is alright with me

http://valid.x86-secret.com/ccpuz.ph...9389&nocache=1



http://valid.x86-secret.com/show_oc.php?id=389389