I've got my sig working fine. Looser timings, but it shouldn't be a problem...
Printable View
I've got my sig working fine. Looser timings, but it shouldn't be a problem...
I have no problem with 8. My e8400 is up to 4Ghz. Prolly could go higher, just dont have the time. :censored:
are we talking about stock bios?
Uh mine came with 14 and it was fine with 4GB. Now running b16_04 (iirc)
Thanks for all the answers to my questions about switching out a 680i for a IP-35 Pro without formatting my hard drive. No problems at all.
I'm currently running at 3744mhz and memory is at 4-4-4-15 2T, but I have a few issues with running 4 X 1GB sticks of memory. I'm currently running in slots 1 & 3, the memory will not run in 2 & 4 at all. Installing the memory in any of the slots, it feels very loose and the sticks do not snap in like other boards I have used in the past. Has anyone else experienced this problem?
The vdroop seems excessive, I've been running for a week with not a single crash or BSOD. Fast! ORTHOS will not run regardless of the vcore. Two cores will run for a few hours, but two cores will crash within a minute. I'm watercooled and the temps max out at 50C on the two cores that run. Any suggestions would be appreciated. I could run 4 X 1 GB at 4-4-4-12 2T on the 680i and ~3800mhz, but always had some kind of issue with the motherboard starting, even at stock 2400.
Great thread, I'm impressed by the amount of knowledge gained by sharing so much information.
Thanks again!
Would I be correct to believe if Core Temp (latest version) is reporting high temps ranging to TJ Max then it's correct, and not a beta bios issue as it supposedly reads directly the DTS? Just double checking before I reseat the heatsink....
Hi all!
Just completed the heatsink mod as per Ace-A-Rue outline!
http://i92.photobucket.com/albums/l3...5022008076.jpg
http://i92.photobucket.com/albums/l3...5022008074.jpg
http://i92.photobucket.com/albums/l3...5022008073.jpg
http://i92.photobucket.com/albums/l3...5022008072.jpg
If anyone is interested, I have started a worklog in the computer cases section of the forum! It can be found here: http://www.xtremesystems.org/forums/...d.php?t=177866
Regards
:up:
http://img178.imageshack.us/img178/3...rentwy0.th.jpg
what should I do to optimal my system, not necessarily overclock it.
advance thanks
ok, after tinkering with uGuru, it became official to me that I do not know what I'm doing
http://img177.imageshack.us/img177/8...0609zl9.th.jpg
Hey guys, I'm having a bit of a problem with a new machine I built using this board. I got the IP35 Pro along with a Q6600 (non g0, not sure what the stepping is), a Zalman 9500A CPU cooler, 2x2GB GSkill PC6400.
My problem is that I cannot seem to get the CPU temperature low enough to be stable. According to the bios it idles at around 60C with the Zalman. I tried reseating it and got the same results. I also tried the stock Intel cooler multiple times with slightly higher results (around 63C). I'm using AS5 and each time I take the heatsinks off it looks like there is good contact.
Normally I would not simply believe the bios temperature monitoring but every time I try to install Windows XP Pro it fails at random points with blue screens and file errors. I tried multiple hard drives and DVD drives with the same result so I eliminated those as culprits. I've also tried the latest bios (16 B09) which does not help. I've also booted with a diagnostic CD which has PC-Check on it and it fails the stress test. I've used memtest as well and the ram tests without error, so I'm pretty sure that's not it.
Another thing I should mention is that the CPU cooler doesn't even really get warm. The motherboard heatsinks get plenty warm and those show normal temperatures in the bios. This makes me think that there is poor contact between the actual cores and IHS on the CPU since I seem to be getting fine contact to the IHS.
Any ideas before I RMA the CPU? Is it even possible that it's the board's fault?
Thanks in advance.
Sniggle,
I'm beginning to suspect there's a heatsink contact issue with some. Here's my story for last night:
--
...I don't believe the ZeroTherm Nirvana NV120 is compatible with the ABIT IP35 Pro. I spent five hours last night reseating it about a dozen times, cleaning thermal paste each time. It is an absolute nightmare. I began reseating it because my load temps would reach TJMax. OK, so I reseated it and it was obvious it hadn't been properly seated. Fine. Reseated it, issue persisted. After finally tightening the bracket connecting the mounting screws to the cooler, it had sufficient contact for a 45 idle - 75 load. I removed because when OCing to 4GHz it'd overheat and shut down and tried reseating it - when I removed it, I noticed that it STILL didn't appear to have proper contact.
To make a long, very painful story short, if you mount the heatsink with the fan facing the front or rear of the case, then the side bars on the mounting bracket will come into contact with the two capacitors above the CPU socket, and I think that's where proper contact is being prevented. If you face the fan down towards the video card, then the fan hits the northbridge heatsink. If you face the face up towards top of the case, well, that's how I last tested it and it's getting crap temperatures (60-90)
With a cooler like this with an E8400, since it's comparable to a TRUE 120 I'd expect then temps between 30-50 at stock, at MOST.
--
...so for heat/cooling issues, I wonder if the two capacitors above the CPU socket are hitting other heatsink screw mount brackets and preventing proper contact. If you remove your heatsink, does the thermal paste appear to have spread properly?
Each time I remove either heatsink (the Zalman or the Intel) there seems to be fine contact. Additionally it seems as if the Intel heatsink should work regardless since all boards need to accommodate it. I've been applying the AS5 in the recommended fashion as posted on their site (a line across both cores).
Is it even possible that I would have gotten a CPU with poor contact between the IHS and cores, or is Intel's QC too stringent for that to be true?
what does Uguru report your cpu temp at?
coretemp is a software code that interpolates what the DTS is putting out...SO...you have to look at that angle too.
normally, i haven't seen more than a 13-15C difference between Uguru and coretemp, that is why i asked about Uguru CPU temp.
Hmm, maybe it is a G0? I was under the impression that the characters 'G0' were in the stepping, which mine does not have. My stepping is SLACR.
I just removed the stock heatsink for the third time and there is perfect contact. As with before, the chip was idling at 62C as reported by the bios and the beatsink was barely getting warm.
This is all at stock speed and voltage. I can't even seem to undervolt it with this board.
What would be a good 4GB kit to run with this board with tight timings?
I'd like to try and get DDR2-850 speeds @ CAS 4.
Is there anything out that might work for me?
Thanks!
what settings would be good for 8 x 400 for the voltages besides the cpu? Ill start at 1.4v for cpu vcore and work my way down........
here is your 'G0' SLACR: http://processorfinder.intel.com/det...px?sSpec=SLACR
what bios do you have installed?
my recommendation is to clear CMOS with battery out; unplug your PSU; push the start button to unload the caps and then put everything back together in 5 to 10 minutes....power up and enter the bios to set optimized defaults...save and now go into windows to see the temps...if you are running stock speed at default values, speed-step and C1E will kick in to reduce your vcore, multi and temp....let us know how that works out for you.
i just pulled this up at newegg for cas4 memory...the patriot stuff likes it has good ratings from previous buyers: http://www.newegg.com/Product/Produc...me=4GB(4+x+1GB)
Reset the CMOS, loaded optimized defaults, and bios still reports 63C idle. I think I'm going to chalk this up as a bad CPU and RMA it with newegg...
take a look at these 2 x 2gb kits: http://www.newegg.com/Product/Produc...2508081&name=4
Update! Im now at 3816 screens below
http://img148.imageshack.us/img148/629/cpuvh9.jpg http://img148.imageshack.us/img148/9688/memsc2.jpg http://img151.imageshack.us/img151/3575/mem2ji0.jpg
http://img148.imageshack.us/img148/9008/voltyp8.jpg
Voltages are still low, I ran orthos overnight and memtest 8 hours stable. Im pretty happy with this setup!
Sniggle, the 16.09 Bios seems to show high temps with Q6600's. Try the 16.03 Bios.
INTEL Chipset Device Software drivers 8.6.0.1007 (non-official)
http://www.station-drivers.com/forum...hp?p=8381#8381
Direct link:
http://www.station-drivers.com/telec...ivers.com).exe
Note: may require manual updating of each System Device, if existing older drivers are already installed.
just a little update my e6750 is up to 3.5 stock cpu volt the board started with ill have to update my bios sometime for more ram options but the flashing procedure looks tricky dont want to edit the file wrong and brick my board... :p:
Anyone got some suggestions on values to start with on a Q6600 G0?
It boots on 400*9 atm with voltages as, but crashes prime within a few seconds. I know the cpu can deal with 400*9 as I've had it on another motherboard before.
CPU : 1.45
DDR : 1.9
CPU VTT : 1.23
MCH : 1.29
ICH : 1.09
ICHIO : 1.55
CPU REF : 67%
It'd be fun to push it further than that as well of course ;)
Hard to tell without having a component sig like below. Is the RAM stable at 400? What is the VID of your CPU? You might just have a bad clocker. Or perhaps the vdroop is costing you the voltage you need...
Yeah I'll get around to making that sig ;), the memory is ballistix 8500 chips, and I've had em in a 400fsb system before, the only thing that's been replaced is the motherboard. The vdroop is quite nasty though.
Thanks, it's on 1.44 now, and that helped considerably. Toying with the voltage and I'm on test5 now, which is a large improvement from a test1 crash :p
I'm reaching my temperature limit though, 68C atm, but the arctic silver isn't completely set yet either. So I figure I wont get it any higher than 3.6, but I'm happy with that anyway.
//edit, at those voltages I have PWM1 overheating :/
@Ace-a-Rue
plz need your advice , i'm now using beta 16.b04 is it worth to move to the new beta 16.b09 bios ?
thanks
regards
Anyone with 2x2gb gskill 1066 PK in this board?
Just saw Ace-a-Rue's PWM mod post and realized that's what I need, I should have the spare screws and bolts in my collection, but I dont have any AS ceramique left. Now I have normal AS5 and whatever crap came with my thermalright ultra 120 extreme.
Would I be better off using AS5 or using the stuff that's already there, just making sure it's got proper contact?
Yeah I was thinking something like that, which is why I asked first, thanks a lot :), gonna order AS Ceramique now :)
//edit, when removing the middle heatsink, there's a white pad around the core, I kept that on, should I remove that as well and use ceramique there too? Or just keep it on the core itself.
Hey guys just wondering, how much does tref actually effect stability and ram speed.
My ballistics would change the tref depending on divider on this board and speed I think. My team extreme wont, its allways at 2600t no matter what.
Is this a spd thing or does the mobo change it. I didnt see any option any where in our bios.
Good advice, the 16B09 is really only for E8x00 users.
Ace, you heard anything about that latest bios? Looks like it failed some QA checkpoints... :(
not sure what's causing the delay...i will try to find out if any new word is out...on monday, i was told by bingo13 that abit was working on refinning the wolfdale code.
Just a quick question about going to 500 FSB and above on this board... I've been currently running stable at 512FSB x 7.5, w/ 1.3 volts on my E8400 - I'll be posting full voltage specs soon when I reach my final speeds. I've read elsewhere that using 500 FSB or above can actually decrease performance by forcing a northbridge strap....
Would anyone know if this is still the case with this board and what the details would be? Asking first as the testing will take an hour out of my few available each night for testing :rolleyes:
man, i just got a Gateway system from CompUSA with a Q6600 in it for 350. Turned out to be a B3 :( The kicker is, ive got it running orthos at 3.4Ghz, set at 1.5v in the bios, it reads 1.47v in the OS, and when running Orthos drops to a whopping 1.4v. Is this much droop normal? Thats insane! At least its running stable though even with the droop at 3.4Ghz. Just looked and the PWM is at 93C under load according to HW Monitor!
vdroop is bad with this mobo :down: Especially with Quad cores only fix is the vdroop physical mod
yea, no kidding. And man oh man do those B3's run hot. 66C under load across 2 cores and 62C on the other 2.
VDroop mod works quite well though, It took me about 20 minutes of work to complete it on the mobo, another 30 or so to take it all apart and put it back together. The results were great for $2 in parts and an hour of my time.
Budwise, if your PWM temps are getting that high I'd put a small fan on the PWM heatsink or better yet, take a look at Ace's signature and follow the heatpipe mod link. Because you have a quad core, the PWM area will be under more stress and generating more heat.
yea, i already did the heatpipe mod. A fan may be necessary. I dropped the Vcore down a hair and continue to test. So far PWM is at 80C...
curious..what kind of paste did you use on the PWM part of the heatsink?...did you just re-use the plastic push clip with spring tension or did you get the stated bolts and nuts?
i got the bolts, but reused the thermal pad for the PWM part. It seemed like it would be a bit messy to have thermal paste all over them. My biggest concern is really the vDroop. Going from 1.47 to 1.40 is just nuts. Ive never seen a board droop THAT badly. Im considering looking at another board for that reason alone.
i would be concerned with 80C PWM temp...the TIM pad is not doing its job very well...have you put a fan on the PWM heatsink?
are you measuring it with a DMM...or...are you looking at what is set in the bios or are you reading what Uguru in windows desktop tells you when it is under a load?...i presume the above stated droop is under a load, correct?
what is your idle Vcore from Uguru and also from CPuz 1.44.1 (do not have uguru and Cpuz opend at the same time)
agree!:up:
do the vdroop mod, what I set in bios = actual idle and load volts measured via dmm within .001 to .004v
yea, its under load with dual orthos. If set to 1.5v in the bios it idles at about 1.47v in windows according to CPUz and HW Monitor. Both also show that when put under dual orthos the vCore drops to 1.4v and sometimes 1.39v. I dropped my OC down because of it to 3.3Ghz and temps have dropped a little and its 10 hours orthos stable. Im just a little annoyed that you have to give it so much vcore just so it doesnt bottom out under load. Because of the such bad droop you have to set your vCore .07v higher at idle than you should. Ive never done any soldering on electronics like this so I think the vdroop mod would be little risky.
unless you are playing heavy cpu demanding games, there is no need to keep your vcore elevated for orthos type running.
i use wPrime 1024M stability bench to find the everyday vcore to do my daily chores....that vcore will be less than needed to support Prime 95, Orthos or OCCT benching.
as far vdroop, it is built into Intel's design for motherboards adorning their logo...some mobo makers add an additional circuit to compensate for vdroop...but...that increases the cost of the board.
Here is a new temp program for intel core processors: http://www.xtremesystems.org/forums/...d.php?t=179044
unclewebb wrote the program and I think he did a very good job.
will not work in vista64 even with disabling driver signature enforcement.
Yea, thats too bad. Maybe there will be a 64 bit driver for it in the near future for the 64 bit users.
Abit IP35 Pro audio question - I'm trying to record "internally". In other words, trying to record audio (using Adobe Audition) from a web site. The problem is that I only see 3 audio recording devices - Mic, Digital line in, and Analog lines in. Is there a way to set it to record from the internal source? Thanks.
when replacing the thermal pads on the heatsinks, the center heatsink has a grey extra pad going around the outer part of the chip, should I replace that as well and use AS ceramique on it? Or should I leave that, and only use ceramique on the center of the chip?
Ron 61
[off]can you host the "realtemp.zip" somewhere?[/off]
it is right here: https://home.comcast.net/~captain-is...d/realtemp.zip
I just did the heatpipe mod this weekend, and I left the padding on that goes around the outside of the PWM. Seems like it was there the even out the base with the bottom of the chip. Not sure. My temps went down a bit with the mod - probably 7 degrees C? Didn't use nylon locking nutz - hope that won't be a problem. Boy, it was tough getting those northbridge bolts threaded!
So here is an annoying quirk. It seems that after bringing Vista 64 out of sleep, the CPU fan header runs my fan at full 12V speed regardless of the uGuru settings (I have it set to run at 30% for below 40C and 100% at 55C).
If I turn off the computer, turn off the power supply for a few seconds, then turn it all back on, the problem goes away and the fan works correctly until I put it to sleep again. Any fixes? I am running BIOS 14. Must be some sort of BIOS bug, no?
you must mean the NB heatsink, which has that white square pad surrounding the chipset...i took mine off so i could use 400 or 600 paper to finish bottom of the NB heatsink...then, i put electrical tape over the electrical contacts protruding up around the chipset to protect them from being grounded (shorted) by the heatsink.
Hey guys and girls,
I could not come up with any answers this weekend to this problem and was hoping some of the folks on this board can help out. My computers main use is a home theatre pc. On the weekend, I got a great deal on two 5.1 surround systems (spherex xbox 5.1 system). The inputs can either be optical or digital coax for full surround.
Here is my problem, I am using the optical s/pdif out for the setup in my living room where one of the 5.1 systems is and it's working great. My other 5.1 system is for the computer/gaming room. Is there anyway to hook up the 2nd 5.1 system in my computer room? I am thinking I need some type of analog (from the ip35's 7.1 outputs) to s/pdif dongle/splitter/adaptor or something?? Any ideas? Here is a picture of what I have to work with on the sound system.
JJ
This program is pretty cool... reports my E8400 at 4GHz as 35c, which I think is more accurate than 45c considering the ZeroTherm NV120 I have on it. The XSBench is also pretty nifty :) Would be neat if there was also a "submit" button to upload it to a server for comparisons
question... is there a pencil mod for this board or do i have to soldier resistors in?
I think your gonna have to do some soldering. This thread talks about all the volt mods for this board
http://www.xtremesystems.org/forums/...d.php?t=152233
Just wanted to give my thanks for many of the posts on this thread. Made a huge impact on my decision to go with this board, and everything has been running smooth!
no beta update this week as long as CeBit is going on...i was told by Bingo13 that the bios engineers go to CeBIT to answer those pesky performance questions on the newest chipsets:D
Well, there is such a thing as an optical splitter. It looks like this : http://www.cablestogo.com/product.as...=102&sku=27027. But 2 rooms worth of toslink cable might get a bit expensive.
compared to my previous post and after doing the WinXP Tweaks...
am I finally doing this right::confused:
http://img507.imageshack.us/img507/9369/3408pi0.th.jpg
I dunno...this is kinda scary now...any advice or hints what I should do...first time I OC a cpu
http://img530.imageshack.us/img530/2086/3408pb3.th.jpg
Well, I finally got my board to work with all Four 2GB sticks! It would work fine with either pair for a total of 4GB memory, but wouldn't even boot with them all installed. I had tried just raising voltages and relaxing timing, but no dice. Then after a suggestion on looking at timings in MemSet and further tweaking I finally got it working. I tightened many of the memory timings, and relaxed tFRC. The timings I currently have (from MemSet) are as follows, which can be compared to the earlier timings (which were the "AUTO" settings in BIOS) which are also listed. I intentionally tightened the first few timings to the specified timings for my memory, and intentionally lengthened tRFC since a poster suggested that it might be one of the ones to play with.
I also raised MCH to 1.33 volts and ICH to 1.09 volts. (I had previously tried just the voltage increases and they alone were not enough to get all the sticks to work, so I suspect it was the timing changes.) Timings were set in BIOS and read out via MemSet.
Of course, Vista wasn't happy and insisted that it be reactivated, which necessitated a phone call to Microsoft. The "nice" lady there seemed to want to imply ("how many pc's is your software running on?") that I wan't legal, but gave me the new key that I needed.
Readings from MemSet: the 1st timing listed is what is now working. 2nd timing listed is what didn't work with 8GB, but worked fine with just 2 sticks.
Cas# Latency (tCL) NOW @ 5 old timings which wouldn't work 6
Ras# to Cas# Read Delay (tRTC) 5 6
Ras# to Cas# Write Delay (tRTC) 5 6
Ras# Precharge (tRP) 5 6
Precharge Delay (tRas) 14 15
All Precharge to Activate 6 7
Refresh Cycle Time (tRFC) 50 (0nly one raised) 46
Performance Level / Read Delay (tRD) 8 9
Write to Precharge Delay 14 15
Write to Read Delay (tWTR) 11 12
Activate to Activate delay (tRRD) 3 3
Read to Write delay (tRD_WR) 8 8
Read to Precharge delay (tRTP) 3 3
All Precharge to Refresh delay 6 7
Refresh Period (tREF) 2600T 2600T
Command Rate (CR) 2T 2T
I think I'm going to settle with 498x8 as that seems to be providing me with the best performance/stability island - details in the pic:
http://i271.photobucket.com/albums/j...stemStats3.jpg
GOOD SHOW!...you were persistent and it paid off big time!:up:
as to having to get a new key from M$, they have a scoring system which is part of your install...the scoring system is tallied and stored as a digital signature...from there on, it keeps tabs on changes to your system hardware...if you change the original point tally by 3 points (total points are 10), then you need to get a new key...thing to keep in mind, the network card is worth 3 points where as hard drive and memory is worth one point...what i am not sure; adding and subtracting multiple hard drives and multiple memory sticks count as only one point, two points or more....probably, your swapping out memory sticks set the M$ monitoring bell off (big brother keeping track) forcing you to call in.:yepp: most likely
this is how M$ keeps track of mobo changes...possibly, as their thinking goes, you are trying to install the OS on a secondary computer, which is not allowed.:)...that is what they are trying to prevent.
I thought I set the Core Voltage at 1.3 in the BIOS but, it seems to be read by CPUZ @ 1.28, don't know if that's the same one we're discussing.
Remember, I never OC any PC before so this Abit IP35 is taking my virginity away:rofl:
BTW, where can I get these 40mm fan? and how do I secure them on the board once I install the board inside the case?
As, always thanks to everyone for sharing their settings and experience.
I'm learning so much.:up:
Thanks for the input. The s/pdif splitter is closer to what I am after. The problem I have with that is that the sound will be the same on both speaker sets. I would like the ability to have media center running in the living room on it's own sound, and me playing a game on the other 5.1 system (from one computer still).
The optical and analog ports will both produce sound from the motherboard audio. Since my living room setup has taken the only s/pdif output from the onboard sound, I am left with the 7.1 analog still. Is there not an adapter of some sort (can't find for the life of me) that will take the analog output from the htpc and convert it to coax or optical digital for me??? Looking forward to some more feedback here!
Toslink cost not an issue with ebay. I have 50 foot dvi's hdmi's and toslink now added to the mix for cheap cheap cheap on ebay!
JJ
I lowered the Core Voltage to 1.25 I believe...see if you can make a recommendation:
http://img254.imageshack.us/img254/9887/3408vc5.th.jpg
advance thanks
with the E8400, it is about the same as my QX9650...to get a 1000 mhz over stock, i've seen about 1.28-1.30v actual vcore...remember, this is not locked into concrete since every cpu is different when it comes to overclocking...
this below overclock (your image) is using to much vcore, in my opinion...try lower vcore for 3.4 mhz
http://img530.imageshack.us/my.php?image=3408pb3.jpg
yeah I think that's when I was using the uGuru to "play with"...but now I'm doing everything in the BIOS....I'll go back to 1.30v again, let it run for 8hrs and see anything would pop up.
Then I'll probably bring up the Bus Speed...just don't know what's the next number I need to enter...what's next to 334.0 MHz?
Is there a website that tells the next scale number?
That part I didn't find a link to how to scale....
I'm just trying out one parameter at a time.:)
Advance thanks...
Found this on [H]ardforum:
maybe I'll got 400MHz...if that 333MHzX9 runs stable for even 4hrs.Quote:
Originally Posted by firas
stay 1:1 FSB to memory while you adjust the other parameters.
MCH, ICHIO and Vdimm play roles in attaining max overclock besides vcore.
try MCH of 1.4x; ICHIO of 1.65 and Vdimm of 2.1 to 2.2 to attain 3.6 mhz using 9 x 400 FSB (CPU VTT set to 1.3)....you can adjust up or down from there to see if it helps or hinders.
use Uguru to adjust your voltages...cpuz vcore is fairly accurate for vcore readout...don't forget that cpuz conflicts with uguru when both are visible on the desktop...just minimize Uguru when you want to open cpuz.
once you got your final overclock then adjust your FSB to memory ratio to see how high you can get your memory...don't forget max memory speed means more stress so voltage increases might be necessary to hold it stable.
good luck!:)
http://img264.imageshack.us/img264/4...c3d8ds9.th.jpg
can't wait to go home:up:
just changed my hsf to TRUE from zalman 9500..... my cpu temp went down but my pwm went up... think my 9500 was blowing on to the hs for the pwm... what size fan should i get to mount on to my pwm and how? with zip ties?