I got the freespace cleaner from a link OCZ forums. Tony posted a thread re trim work around for RAID. Can't find it just now, have a search on the OCZ forums you should be able to hit it.
Printable View
I got the freespace cleaner from a link OCZ forums. Tony posted a thread re trim work around for RAID. Can't find it just now, have a search on the OCZ forums you should be able to hit it.
PART 1 - THE ASUS MENACE
I recently bought myself an Asus S2E on ebay.
I read this thread before buying the board & decided to find out if the board really is as bad as all the negative posts in this thread (and the poor reviews on new-egg).
Received the board from ebay in mint condition, with all accessories included & flashed with the latest BIOS 1402.
I put my E8700 in with some G.SKILL DDR3-2400 (not on the QVL, designed for the P55 chipsets, with XMP profiles), the board recognised both the CPU & RAM and posted fine @ 9-9-9-24.
I went into the BIOS and tightened the RAM timings manually to 7-7-6-18;
http://members.iinet.net.au/~michaeldd/E8700_790i.jpg
Things were looking good - now for some overclocking!
PART 2 - ATTACK OF THE HOLES
I decided to do a quick FSB test with my E8700, capable of ~570 MHz on air cooling.
I started at 460 MHz but the board failed to post (the LCD poster kept saying 'CODE INIT'), so I increased the FSB past the apparent hole until it POST'd at 476 & went up from there, only to hit a wall at 520 MHz.
Increased PLL, VTT & vcore, still no go. Set P1 & P2 to auto and increased NB voltage, played with NB GTL, nada.
I was disappointed, anything above 520 MHz and the mobo wouldn't POST.
PART 3 - REVENGE OF THE RAM
I decided to move onto some RAM overclocking - this is where the problems really started.
I inserted 2 x 2GB of RAM in the blue slots & set the CPU FSB to default (1333 QDR) and incrementally increased the RAM (dividers allowing) up to a final 1700 MHz with loose timings and P1 & P2 on Auto.
Anything above 1700 MHz & I would either get no POST or errors in memtest86+ (v4.10).
Increasing the NB voltage above 1.46v caused the number of errors to increase into the tens of thousands, yet the NB temperature seemed to be ok, idling at ~60°C and increasing to ~63°C under load.
I tested each RAM stick in my friend's X58, no error's - it had to be the 790i chipset.
Again i was disappointed...
I set the FSB to 1600 QDR & linked the RAM, the board POST'd ok. Everest showed the chipset latency to be ~55ns, unfortunately P1 & P2 had no effect on chipset latency at this low FSB frequency.
I decided to increase the FSB in an attempt to find chipset straps & decrease the chipset latency but the board couldn't post above 1600 FSB!!! WTF!?!
I discovered the size of the FSB hole... no, more like a freaking black hole, from 1601 up to 1904 !!!
I was seriously starting to believe all the bad things I had heard about the S2E...
The next step was a BIOS flash, I thought that maybe 1402 sucked, really really sucked...
Used EZflash to flash 0101, the board POST'd ok but couldn't recognise the CPU model due to missing microcode - tested the FSB anyway, no change.
Tried several more BIOS - still no change!
I couldn't believe it - I had purchased the worst Asus S2E ever made.
PART 4 - A NEW HOPE (DON'T MESS WITH THE ORIGINAL, K GEORGE!?!)
I began scouring the internet for clues on how to solve the FSB hole issue & came accross someone with the same problem - he couldn't set the FSB higher than 1333 QDR.
This is the line that really caught my attention;
EUREKA!Quote:
Originally Posted by meisterlr
With some thorough testing I found that EZflash 2 was actually causing BIOS corruption and that Awardflash v1.33 fixed the FSB holes for me (opposite to what meisterlr had stated - but hey it worked!)
So I retested the max FSB - the board could now boot to 550 MHz (2200 QDR) !!!
http://members.iinet.net.au/~michael...90i_550FSB.jpg
PART 5 - THE NORTHBRIDGE STRIKES BACK
The weird NB voltage issue was eating at me - how could increasing the voltage from 1.46v to 1.48v cause such a drastic increase in chipset errors? and why did 1700+ MHz on the RAM have the same effect?
I decided to take the cooling off and examine the board - maybe something was damaged in transit.
This is what I discovered;
http://members.iinet.net.au/~michaeldd/790i_nb_tape.jpg
A piece of plastic tape sitting between the NB chip and the NB cooler!
What better way to cook your NB than to increase the voltage with ZERO contact to the heatpipe.
I wonder how many left the Asus factory like this???
Removed the tape, replaced the TIM & re-mounted the cooling - idle temp 53°C @ 1.46v - load temp 54°C @ 1.46v - much better!
PART 6 - RETURN OF THE OVERCLOCKING (ok, so I'm lame - you don't need to tell me)
Having dropped the NB temp by a whopping 10°C I decided to re-test the max RAM frequency again.
Result, 2x2GB DDR3-2074 9-11-9-24 2T;
http://members.iinet.net.au/~michael..._ddr3_2074.jpg
I know the bandwidth/timings are crap but being that the board wouldn't even POST with the RAM above 1700 MHz in the first place, this is a big improvement - and the best part - no chipset errors!
Just for sh%ts & giggles I stuck another 2x2GB of the G.SKILL DDR3-2400 into the white slots to see if the NB can handle 8GB, it's currently running at DDR3-1824 without errors;
http://members.iinet.net.au/~michael...p2_enabled.jpg
http://members.iinet.net.au/~michael...led_49.7ns.jpg
I can see how easily people would become frustrated and give up on this motherboard.
As for the plastic packaging tape on the NB, well, Asus QC failed, plain and simple.
Very good effort and job there, remembers me of the 680i days :p
Persistance wins (mostly) in the end !!
Thanks Leeghoofd :).
For Trigger5521 & Sangrak (and anyone who's interested), here's a ss of my 4 x 80GB Intel G2 SSD's in RAID0 on the 790i;
http://members.iinet.net.au/~michael...CQ_ENABLED.jpg
chino
I freakin love your mini review of the board. you rock sir. excellent job.
You make me wanna go spend $500 right now on buying water cooling parts and start overclocking my board LOL
Interesting BIOS thing, will definitely re flash my board with AWD the day I decide to come back to this board :)
Btw, which BIOs let you go p to 550mhz FSB?
Hey slim thanks for the positive feedback - I actually got a fair amount of info from your posts, it's good to see you still keep up with this thread.
I can get to 550 MHz FSB on 1402 (but also some earlier BIOS such 0402, 1301 & 1203) but I didn't test that many with my E8700 before I dropped my quad in.
My QX9650 goes to 476.25 (1905 QDR) on my 780i and I have equalled that on the 790i;
http://members.iinet.net.au/~michael...6.25_x_8.5.jpg
I dropped the multiplier to 6 and disabled the 2nd core which allowed me to push the FSB up to 487.5 MHz (1950 QDR).
http://members.iinet.net.au/~michael..._487.5_x_6.jpg
Looks like I might need some Vcore, VTT, PLL and GTLVREF adjustments to make the 2nd core stable above 476.25 MHz.
[Edit] I was interested in comparing the 790i with the P45 for vcore requirements at high clock speed and as usual the Nvidia board requires more voltage.
Here's a ss of the E8700 @ 4.762 GHz, with the voltage set to 1.50v (LLC disabled), the P45 can do this with ~1.46v;
http://members.iinet.net.au/~michael..._4.762_ghz.jpg
http://members.iinet.net.au/~michael...hz_everest.jpg
Some reasonable bandwidth here, with the chipset latency coming down to 45.3ns (P1 & P2 enabled).
4X raid 0!!!!!! I love it, you sir have the xtreme sickness.Quote:
Originally Posted by chinobino
thx for posting bench. Also good job and feedback on OC'ing. Where were you when the board 1st came out we could have used you skills.
Haha, cheers Trigger!Quote:
Originally Posted by Trigger5521
I bought the 780i instead because DDR3 was so damn expensive at the time.
Recently my gf's top PCI-e slot died on her Asus A8N Premium, so I gave her my 780i and bought the S2E and some non-QVL RAM.
I have to admit that it is a lot easier to do troubleshooting when a product has been out for a few years and you can get a lot of information from a single thread like this one, plus choose from a large selection of BIOS :D.
I noticed that the Soundmax soundcard makes a 'pop' through my amp/speakers when I boot up.
I'm not too fond of this sound (and don't want to spike my speakers), can anyone recommend a good PCI-e sound card to replace it?
Kick ass!Quote:
Originally Posted by Trigger5521
Out of sheer envy of all that performance, I just ordered another Vertex LE. Should arrive sometime next week.
I haven't striped drives since my Athlon XP system many, many years ago. Is there a way to retain data on my first drive without having to reinstall all my software again? I just finished installing all my crap and I don't want to go through the pain again. I figure I may not have a choice, though.
Chinobino, Trigger and I use an X-fi titanium in the top 1x slot. Works well. The latest driver solved the crackling with SLI enabled and the sound is great.
If you are switching from no RAID to RAID0 then you would first need to ensure that your current Windows installation already has the drivers installed for the RAID array, then you could attempt to image your current installation with Acronis True Image or Norton Ghost onto another drive, then create the RAID array and attempt to restore the image, no guarantee that this would work though.Quote:
Originally Posted by Sangrak
Ahuh! I will have to have a look for one of those, thanks for the suggestion :).Quote:
Originally Posted by Sangrak
I just read that Creative is launching a new X-Fi card for the Fall. You may want to wait until then to pick another soundcard. The drivers will probably be buggy as hell (like all Creative drivers) when it first comes out, but it may be worth it.
I just got my second Vertex LE today. Hopefully, I'll have the time to stripe them this weekend. I'll post some benches as soon as I get them running.
Thanks for the heads up Sangrak - I haven't bought anything yet - was looking at some 'Creative Sound Blaster X-Fi Titanium PCIe Fatal1ty Pro's on ebay but haven't bitten the bullet yet.
Any ideas as to what features the new X-Fi card will add?
WOWWWWWWWW i cant believe my eyes that this thread is still alive, you guys are die hard striker fans.
Hey it's neo mike!Quote:
Originally Posted by neo mike
Are you gonna go ebay hunting for another board then? or buy your old board back off your neighbour? :D
You spent all that time making the heatsinks and didn't get to use the last board long.
Hi chinobinoQuote:
Originally Posted by chinobino
Man, I have the exact same setup as u, and Im going crazy since 2 years ago trying to get my S2E stable on both cores at 4.7 Ghz.
Need your advice sooo bad.
I finally bought a couple of months ago an E8700, but 4.7 stable seems impossible to me.
I have water, temps are pretty ok. I got it stable on prime95 for more 6 hours on both cores (manually stopped) at 4.5Ghz (500FSB).
I got it at 4.5 with
vcore: 1.3687
LLC: enabled
PLL:1.5
VTT:1.2
NB: 1.62 for P1 and P2 enabled on the ram (otherwiser 1.3)
RAM: 1.8 (1600Mhz CAS) just bought 2 new sticks with lower voltage.
SB: 1.5
Didn´t touch GTL REF volts at all (all on auto right now) Donno how to calculate it.
So, any advice u might gave me to get it stable at 4.7, would be greatly aprecciated. VTT, PLL, Vcore, GTL, guess I have to do some fine tuning there. My concern is GTL´´s cause donno where to start there.
Thanks in advanced
Franco
Hi Francko1979, sry I completely missed your post, here is a late reply...
4.7 GHz is most likely going take a fair bit more vcore.
As you can see in my SS, v1.504 vcore (LLC disabled) to be stable @ 4.762 GHz.
I have put up to 1.55v through my E8700 (LLC disabled) to run at 4.830 GHz (10.5 x 460) but the heat was unmanageable on air coooling.
I leave all GTLVREF's on Auto, I found that adjusting any of them made the PC fail to POST or it corrupted my OS install.
I used 'Auto' as a reference on my 780i to see what they were being set to as the FSB increased, I found that positive values helped, even on a 45nm CPU.
The problem with the S2E is that you can't see what 'Auto' is setting them to as there is no monitoring for the GTLVREF's :(.
My settings for 4.762 GHz (476.2 x 10);
vcore: 1.50625v (1.50000v was failing small FFT's on Prime, so I nudged it up)
LLC: Disabled <--- This board does some crazy things with LLC enabled and high vcore, so be careful if you leave it enabled
PLL: Auto (1.55v) <--- I have tried increasing this up to 1.74v but I still could not get over the 550 FSB wall
VTT: 1.4v <--- You may need less so start low and work your way up
NB: 1.44v (P1 and P2 enabled) <--- If I go above 1.46v I get errors in memtest from the chipset
RAM: 1.64v (overvolts to ~1.70v)
SB: Auto (1.5v)
The biggest problem is the increased heat from upping the vcore, being that you have water cooling you should hopefully be able to keep it under control :).
Also, you might want to consider using a lower FSB, setting the multiplier to 10.5 (448 x 10.5) and getting that stable first, then try to get 470 x 10 stable etc
That will take some stress off the NB and CPU.
Hey Guys need some ideas here. My OCZ gamerxtreme 850W power supply kicked the bucket last week. I did'nt want my system down for a few weeks waiting on the RMA so I bought the Corsair HX1000W modular PS figuring I will sell the RMA PS when it arrives.
My problem is the system freezes after resuming from sleep now that I have the new PS installed. WTF?? this doesn't make sense to me. I have tried so many things, BIOS re flash, new video drivers, disable X-fi sound card. Stock clocks no OC's. Increase PLL. Nothing works to solves this, it's driving me nuts because I need my puter to sleep and wake properly.
Shoot your ideas please. What would the power supply have to do with freezing on resume?? Thx for any help.
It's odd that a psu change would affect your system coming out of sleep, especially when it was working before. I never use sleep since my Vertex LE doesn't respond well to it.Quote:
Originally Posted by Trigger5521
This may not have any affect on your problem, but it's important nonetheless. Since the HX1000 is a 2-rail PSU, make sure you have your GPU's on separate rails. Here's a diagram which you may or may not have seen:
Hey Trigger, have you checked your Event viewer logs to see what's causing the crash?
I had the same issue, 'nvoclk64.sys' (Nvidia System Tools) stopped my gf's 780i from waking up from sleep so I removed it.
The default Windows 7 SATA/RAID drivers where causing my S2E to wake up from sleep with Aero disabled so I updated to the latest chipset drivers (15.53WHQL) and the problem has been fixed.
I did look at the event viewer all I'm getting is the system was rebooted unexpectedly. I have to power down after a failed resume. It's not trapping the problem as far as I can see. I just upgraded to 15.53 that was no help.Quote:
Originally Posted by chinobino
I'm going to try disabling system tools and see if that helps. If not I'm going to start disabling services one after another.
Hell YEAH!! Chinobino, renamed 'nvoclk64.sys" to .old and bingo see resumes no problem. Friggen awesome I want to buy you a beer.
Thx!!
Awesome sauce Trigger, glad to help :).
Nice! Glad you guys figured it out.
Gave up on the QX9650, it seems to dislike high FSB with 4 cores enabled...
Still chasing the Holy Grail of 500 MHz FSB on a quadcore, it's time to move onto the Xeon X3380...
X3380 ALL 4 CORES - FSB 491 MHz (1964 MHz QDR) w/8GB RAM
http://members.iinet.net.au/~michael...ll 4 cores.gif
http://members.iinet.net.au/~michael...-8-8-22 2T.gif
X3380 491 MHz FSB P1 & P2 Auto (Poor chipset latency & bandwidth)
http://members.iinet.net.au/~michael...uto 84.1ns.jpg
Close but no cigar.
Fails to boot into Windows above 1964 (will POST though).
BIOS settings for 491 MHz FSB;
1964:1637 (FSB:RAM)
Vcore 1.45V LLC enabled (1.488V)
PLL 1.60V (1.64V)
VTT 1.40V (1.45V)
VDIMM 1.64V (1.71V)
NB 1.58V (1.60V)
SB 1.60V (1.58V)
NB GTL -20mV
CPU GTL 0 -95mV
CPU GTL 1 -56mV
CPU GTL 2 -56mV
CPU GTL 3 -95mV
P1 & P2 Auto