i am gonna try this again wolf[IMG]httpQuote:
Originally Posted by theYipster
i am gonna try this again wolf[IMG]httpQuote:
Originally Posted by theYipster
So I took my PC down to the place I brought it from and explained that I wasn't able to get the ram to operate at specified settings. If you recall I have been concerned about bad ram hindering my OC ability. Well I picked it up today and they had all 4 dimms running at the advertised 1066. They did this by enabling Sli-Ready setting to Expert, overvolting the DIMM setting to 2.20v and manually setting the timings in the advanced dram area to 5-6-6-15-2T. So then is the ram ok? I'm still not sure.
I tried once again to get the E8400 up to the levels others have gotten with the P7N .... without success. If the ram is in fact not bad then I guess the CPU or mobo must be. I guess i'm SOOL. I will wait for a official BIOS update, until then I am suspending all OC efforts. A shame.
My Q9300 OEM came in today, got it installed and BIOS 1.0 recognized it fine. Temperature reading is not okay at the moment though, this could be due to the programs (CPUIDHW, Realtemp) not reading correctly or me not screwing down my heatsink tight enough. I will reseat it this evening and see if there is any difference.
https://home.comcast.net/~adamsbr2/m...ocker/9300.jpg
It's weird because after running 3DMark I rebooted to check what the BIOS CPU temp was and it was in the high 30's. So I have a feeling I seated the heatsink correctly. Problem is I want somewhat accurate temps before I increase Vcore for overclocking. So for now I plan to see how far I can get on default Vcore with the first increment set in BIOS and probably 1.3 NB. Any suggestions on where to start with the CPU GTL and FSB Term?
I wouldn't worry too much BKA. The first E8400 users experienced the same issues regarding temp readouts. I'm not sure if their issue was with the chips themselves, but my guess is that it's was due to monitoring software not correctly supporting the chip. I suppose it's the same issue with the Quads, and that it'll be resolved in future software updates.
Thanks Yipster, any suggestion on the GTL or FSB TRM to get started or should I just play around with the setting? I read your earlier comments on the Q6600 settings for these values. Maybe I'll start around there somewhere.
I was just looking at the detail for that 3DMark score and the CPU is recognized as a 4 Core, Intel Pentium III. LOL
Bka,
Expect the Yorkfields to stabilize under different voltage needs. You might want to try 1.425 for the NB and 25 for FSB Term. I've found those to be stable throughout different vCore and GTL ref combos. As for actual vCore and GTL Ref needs, you're pretty much trailblazing here.
Btw, what's your chip's VID in Core Temp?
Hi everyone,
Quick interjection here: As I've noted, it's very important to know your chip's VID when overclocking on this board due to the way vCore is set in the BIOS. The easiest and most popular way to get this information is from Core Temp, which until now was quite a pain to get working in Vista x64.
A new version (0.97) has just been released and it support Vista x64 natively. No more 3rd party batch script enablers or running the program in disabled-driver-signing mode.
Download here: http://www.alcpu.com/CoreTemp/
I'll have to check when I get home from work. I do have the new version but haven't really been using as CPUIDHM was giving the same temps plus my GPU temps but I'll check when I get home and report back. Thanks for the suggestions!Quote:
Originally Posted by theYipster
EDIT: I do know the BIOS was reporting 1.20,1.28,1.36 in BIOS setting, fluctuating between the three.
Please correct me if I'm wrong because my knowledge and understanding of this is somewhat limited. Isn't VID a non-changing value, programmed into each individual cpu that tells my bios what the 'stock' voltage is for my processor?
Loonym,
You are correct.
CoreTemp reports 1.0375 as the VID, CPU-Z report curent Vcore at 1.128-1.136, CPUID as 1.13-114.Quote:
Originally Posted by theYipster
CoreTemp and CPU-Z also report it as a C1 revision not M1 like TigerDirect listed for the CPU description.
I've never heard of an M1 stepping. I believe C1 is correct. Do you have your chip overclocked or do you have speed step (Intel EIST) enabled? At stock settings, cpu-z should be reporting vCore below your VID. Note that Speedstep / EIST can cause an incorrect VID reading (reporting a VID that's too low.) If you suspect this is the case, turn off Speedstep and read the VID again. Unfortunately, I don't have a comparative basis at the moment as I haven't read up on Q9300 chips.
No, I disable Speedstep/EIST from jump. MSI does report a M1 stepping on the supported CPU's for this board, although both C1 and M1 are listed as "Under Testing". Only the C0 is listed as supported by BIOS 1.0.Quote:
Originally Posted by theYipster
Another thing I noticed, although the BIOS has it listed as 7.5 for the ratio, the highest that can be adjusted is 7.
EDIT: Intel only has the M1 listed in their Core2 Quad processor list. Hmm.
I really never thought it was significant enough to mention but I doubt the boards 'true' support for the .45 line. i.e. the ability of bios to set the initial voltage of the cpu correctly at boot.
Okay, so I have screwed up somehow. So I tried a simple overclock with no luck getting into Windows. So I decided to try the beta BIOS 1.1B4 but noticed that if I disabled EIST it drops the adjusted CPU Ratio down to (7x333) for 2.33 core clock with all auto settings. And the option for "Adjust CPU Ratio" only has 7 or 6. No matter what other setting I use the only way to get back to a 7.5 ratio is to enable EIST. I know this wasn't the case with the 1.0 bios because I double-checked it after you mentioned it.
So then I tried P05 with the same results if I disable EIST, no 7.5 ratio only 7 or 6. So I figured I'll just flash back to 1.0 and leave it at default for now until MSI gets another BIOS out there. But now its happening with 1.0 also. No 7.5 ratio unless EIST is enable and of course that removes the Adjust CPU Ratio option altogether. Not really a problem in Vista because I always have my power managment set to High Performance and Processor Power Management set to 100% in minimal and maximum, so it runs at (7.5x333) all the time but now the the auto vcore is at 1.200-1.208 instead of the 1.128-1.136 it was earlier.
So I figure I will test out by putting the E8400 back disabling EIST, do a few reboots and then replace the Q9300 to see what happens.
Also with EIST enabled now Coretemp shows a VID of 1.1125V which lets me know I'm not going crazy because I know it was disabled at first when I reported the 1.0375 as the VID.
So if swapping CPU brings back the option to disable EIST while keeping a 7.5 ratio, I will probably flash the 1.14B again and see what I can come up with at the 7 ratio just for kicks tonight.
It could very well be that MSI does not yet have a BIOS capable of handling the half-multipliers found on the Q9300, Q9550, and E8500 chips. If they don't, I am sure they are working on one, as the P7N Diamond is indeed advertised as 45nm ready.
You should e-mail MSI tech support about this issue, and you may want to start a thread on the MSI forums as well. The issues with EIST may also be a factor of missing / incomplete BIOS support.
In the mean time, I would suggest leaving everything at "Optimized Defaults." When MSI releases an update that addresses the half multiplier issue, you should then revisit your OC efforts.
Good idea, will do. In the process of putting the 8400 in now and if that gets me back to EIST disbaled with the 7.5 I'll leave on defaults for now.Quote:
Originally Posted by theYipster
Yes, it's always a good idea to make the people at MSI tech support aware of every issue. They are probably sick of me by now even though they are always friendly and polite. I have an ongoing dialogue with them and they have asked me to help them test a few things as well. They are the ones that are going to write the bios that will make us all very happy ;)
Well, with my E8400, I get choices of 6, 7, 7.5, 8. 8.5 and 9 under Adjust CPU Ratio, so go figure. I'll make sure to included all this in the email.
You doing a cmos clear with mobo power discharged when you make these cpu swaps?
Yup, because I have my normal settings saved (RAID and stuff I disable) in the user setting. So I only have to adjust the Cell Menu after loading those.Quote:
Originally Posted by loonym
EDIT: So after putting the Q9300 back in, sure enough EIST disabled and its auto set at 7.5, but still only 6 and 7 ratios were available. Leaving well enough alone for now. Going get some gaming in before I retire for the night. Thanks again for the suggestions Yipster.
I think the VID under core temp for different mobo also plays an role in determine the voltage
Under Core Temp 0.97. My E8400 is 1.0XXvolt for MSI P7N Diamond but it is 1.1XXvolt for my Asus Maximus Formula.
It might due to the different manufacturer has different calibration on the vcore even the cpu is the same...
So, I will use the monitoring software as a guide and reference only. Using multimeter is still the best way for reading more accurate voltage
Memory,
I have never encountered a chip whose VID differs based on the motherboard, but you may be right. I've never used the same chip in an nVidia board and an Intel board. If you have a multimeter, that is of course the way to go.
Well, I emailed MSI and they told me to Scram, Beat it, you idiot, its still under testing, didn't you read the CPU support list for this motherboard!!!!!
Just kidding but they did say that I would have to wait for the next BIOS release because they are still testing this revision.
Well, at least now you have something to look forward to :) Ah, the joys of being first with new technology.
Hey has anyone found a pencil mod for the vdrop cuz im droping about a 0.05 and its cause me to have high volts to keep stable
I really think they must be very close to releasing a new official bios to address whatever issues are currently present and to support the new wave of cpus from intel.
Do tell, do tell. Any insight there loonym?Quote:
Originally Posted by loonym
I guess just really call it a hunch. I've received a couple e-mails from MSI support taiwan asking me to load certain bios' and test specific thing like boot priority, eist functionality, seemingly silly stuff, but it feels good to know they aren't forgetting us and I'm sure I'm not the only guinae pig they are approaching in this way.
That's cool. Hopefully by this weekend so I can play around with it. I sold my E8400 so the Q9300 will be staying in for now. Although I have a Celeron E1200 that I can use if needed.
hi guys. first post here after reading from you for weeks. I just built my system and wanted to know if one of you could tell me which are the best settings I should use in the bios (see my sig for specs). for now, I'm using the load default settings except for the RAM which I had to put manually at recommanded timings and voltage by crucial. I don't want to overclock anything right now, I still have to learn a lot about it before trying. all that I want is to be sure I'm using the right and faster settings available. also shall I upgrade to 1.1b04 or perf 05 bios if I don't plan to overclock now ?
thanks in advance for your help ;)
gamin, If your not overclocking and everything is stable and working to your satisfaction I feel you should be happy. Very nice rig too :up:
Hello All,
Your thread helped convince me to migrate from an MSI Neo2 FR to the P7N Diamond. Overall, I would say I'm quite pleased with the beefier board.
I have, however, a nagging graphics problem: 3dmark06 skips a few frames every couple of seconds... just a little hop. My score has also lowered from about 13000 to just above 10000. All my hardware is the same aside from the new board and 3dMark06 was silky smooth with the Neo2. I've tried two drivers, 169.02 and 169.25. Also ran a diagnostic through rivatuner and everything seems okay.... getting a x16 PCI bus. I've even played around with the "PCI Latency Timer" under the advanced BIOS features to no avail. In fact 3dmark06 wouldn't load under the highest setting, 128.
I'm still running the shipping bios on the motherboard- v1.0. Is there a bios update that addresses graphical hiccups like this?
From reading this forum, it seems you guys know more about what's happening on the ground than the tech guys a MSI. Any clues about how I can proceed?
Thanks much for your reply. This is an extension of my first build and I'm still learning every day.
thanks for your reply loonym. so I'll stay that way for now. btw, do you have any advice in a good article/place explaining overclocking ? I mean I'm totally newbie in this, and I really need to learn from the beginning before trying.Quote:
Originally Posted by loonym
thanks for that ;)Quote:
Originally Posted by loonym
So far this thread is about the best advice I've seen. Some very good testing being done by these guys.Quote:
Originally Posted by gamin
yes, I know, but I have to admit I don't understand everything (told you, a real newbie in overclocking :confused: ). what I need is a step by step tutorial or something like that, where I can learn each step before trying to reach the max of my cpu... if only MSI would have documented the manual more completely...
EDIT: I took some time to read and understand theYipster's update#1 on opening message, and I guess I've understood a little now... :yepp: gonna give a try and post my results (if I don't burn or break everything of course... I'd be happy if I can reach and have a stable CPU at 3.6/3.8Ghz for now). don't expect results 'til next weekend, I unfortunately have a lot to do at my job this week :(
Ahoy hoy,
It's been several weeks since I've updated the first post, but I'll be putting up new information today (along with guidelines to OC the E8400.) For an early BIOS, 1.1B4 and P05 have shown to be quite good at supporting demanding OCs, but clearly they still have lingering issues (particularly with 45nm chips.) I concur with Loonym that an update is forthcoming. Keep in mind, however, that the reference boards have been on BIOS P03 for some time now, and that the non-reference boards often get BIOS development cues from the reference ones. After the 790i hoopla settles down (and hopefully it will quickly...) expect to see this and other 780i boards continue their march towards BOIS maturity.
As for my own endeavors, I'm sitting pretty at the moment at 3.6ghz. The P7N has gone into my main rig, which I recently reformatted with Vista SP1. The limited time I've had with my rig these past couple weeks has gone into getting my software, mp3, games, etc. back up and running, so I've paid little attention to my OC efforts. I'm looking forward to picking them up again once we see new a BIOS from MSI.
Andrew2004gto,
I'd love to have a pencil mod, but I don't think I'm going to be the one to find it. If anyone is up for the task, many of us would be eternally grateful! :D
Lutefisk,
That's a very interesting problem and it’s the first I've heard of it. I'm running two GTXs and I haven't experienced any frame hiccups or a loss of 3DMark score. (Hit around 18300 in XP, which is up about 500 points from my 680i.) Try updating to the latest BIOS, 1.1B4, and see if that resolves it. Two questions: Are you running SLI, and which slot is your graphics card in?
Ash,
From what I've seen the board is pretty easy going with memory, although I do recall people having issues with certain Corsair models on the MSI forum. Here are a few points to consider, and a few questions for you to answer. Please reply with your findings so that we hopefully help you further.
Point 1: MSI colors their DIMM slots differently than other manufactures, and in a way that many would classify as incorrect. In order for RAM to work in dual channel mode, the DIMMS can’t be set in adjacent slots. If you are running with two DIMMS, the best slots to use is the one furthest away from the processor and the one that is second closest to the processor. If installed correctly, one DIMM will be in a green slot and the other in an orange slot, with one and only one slot in between them empty. This can trip up many builders, as every other Core 2 board on the planet (except those from MSI) colors DIMMS by channel and not by physical pair.
Point 2: If you have four DIMMS installed, they should each be running at dual channel speeds. However, if each DIMM is a 2 gig stick, for a total of 8 gigs installed on the board, you might need to bump up the North Bridge voltage to achieve stability. 8 gigs is possible, but as on any desktop motherboard, it can be taxing on the NB. As such, you may need to manually adjust voltages before your system can successfully boot and remain stable.
Point 3: Standard PC-6400 DDR2 RAM is designed to run at 1.85v. Performance RAM, which can either be PC-6400 RAM with tighter timings, or RAM designed to run faster than PC-6400 / 800mhz, may require a higher default voltage in order to boot and maintain stability. Be sure to set your RAM voltage manually to your manufacturer's spec.
Question 1: What RAM vendors & models have you tried, and what DIMM positions have you tried?
Question 2: What are your BIOS settings for RAM? Please list voltages and timings.
Question 3: What tool are you using to determine whether or not your RAM is running in dual channel mode? If possible, please post a screenshot of the information and/or benchmark.
It may be that your board is defective, but before you look towards an RMA, it's important that you verify the basics. Please post your responses to these questions, and we can go from there. Also note that there is no such thing as DDR2 RAM running in quad channel. With four sticks installed or with two sticks installed in the correct positions, your RAM should be running at dual channel speeds. With an odd number of sticks installed, RAM will be forced to run single channel. This is a design limitation of DDR RAM.
Gamin,
Welcome to the community. I've found the P7N to be a very good overclocker, but unfortunately it's not as easy to manage as some other boards. This is due in part to the particularly cryptic voltage settings in MSI's BIOS, which we (and MSI's forum moderators) are still trying to figure out completely. In the mean time, our successes have come largely from trial and error, and while some find mastering a difficult BIOS to be a fun challenge (ask anyone with a DFI LanParty,) it can indeed turns others away. My only advice is to stay with it and to stay with us, and I'm sure you'll find success (barring any BIOS incompatibilities or hardware defects.) My own opinion is that a tricky BIOS is a small price to pay for a board that avoids the myriad issues found with the 680i and 780i reference models. I've been burnt by them twice, and I'm very wary of going back now.
The other impedance to a strait and easy OC with this board is that the enthusiast community of P7N owners is quite small (but growing,) especially compared to the reference boards, which seem to particularly attract people who are looking for high end hardware but aren't necessarily old hats at PC tweaking and tuning. As such, there is a wealth of information on hitting that easy OC with the reference boards (P0wner's guide is probably the best at the moment,) but little outside this thread for ours. You should indeed read P0wner's guide, as well as some other general Core 2 Overclocking guides. These won't apply directly to managing the P7N BIOS, but they do cover the basics very well and you can bring over much of the information they provide. Of course, as more and more people choose this board, I hope that we build a community (and information library) to match those of other nForce models. That's my goal with this thread.
To start you off, here are a few basic guidelines. I'm not going cover everything a proper guide will, but hopefully this can get you oriented. This may be a bit too general for your needs, but I thought I’d post it here just in case it helps you or others new to OCing.
1. As any guide will tell you early on, the "art" of overclocking is to run a component out of spec--particularly faster than its rated stock performance. Due to the way chips are manufactured, overclocking has been a cherished hobby of many for some time. Yet due to fantastic architectural and yield properties, the Core 2 has heralded in a golden age of OCing. 50-75% OCs can be had easily with good air and water, and insane 100-200%+ OCs have been achieved by some of the famous LN2 masters here at XS. The trick is to find the right motherboard that will support your OC ventures. The "experiment" with the P7N is to find a Core 2 SLI board that a) OCs well and b) is reliable and lasts. So far, the P7N is proving itself up to the task, but at the expense of a tricky and still (quite arguably) immature BIOS.
2. There are two ways to OC a Core 2 chip, but unless your chip retails for $1000+, you'll only have one. On Extreme chips, the user can adjust the processor's multiplier (more on that in point 3) for an easy one setting OC, barring any necessary voltage adjustments. For everyone else, the processor must be overclocked by making adjustments to FSB speed. These adjustments will also affect the speed at which your RAM runs, unless you set RAM to manual control and keep it at a constant setting. Since your sig' suggests you have an E8400, it's on to point 3.
3. Base FSB speed and the relationship between processor, FSB, and RAM: This is the key to overclocking non-Extreme Core 2 setups. The operational speeds of your processor, FSB, and RAM are all based off this base FSB speed--each by a different multiplier. (Note that while you can also overclock your PCI-E bus and graphic cards, these are handled independently.)
** Your E8400 ships with a default multiplier of 9.5, and its stock speed of 3.16ghz is obtained by multiplying 9.5 by it's target / stock base FSB speed of 333mhz. (Note that older 65nm chips have a base FSB speed of 266mhz, while some lower end Allendale and Pentium E chips have a stock base FSB speed of 200mhz.)
** Your rated FSB speed, like all Core 2 FSBs, is quad pumped. Hence, the FSB speeds you see advertised on the box (and the speed you set in the BIOS for FSB) is 4x the base FSB. As such, a 333mhz base FSB is 1333mhz rated / quad pumped. This is the default supported FSB speed of the P7N.
** Your rated RAM speed is 2x the base FSB speed when set to a 1:1 divider, such that RAM running 1:1 with a base FSB of 333mhz runs at 667mhz. Note that on an nForce system, you can set your RAM independently of your FSB, and even on Intel chipsets, you can change this divider. However, most recommend using this 1:1 divider when overclocking for two reasons. First, running 1:1 will have your RAM run only as fast as it needs to in order to maintain performance, and not more. Some point out that running RAM faster than FSB results in no noticeable performance gain on a Core 2 system, but whether this is true also depends on the chipset and motherboard used. Second, on nForce chipsets particularly, many find that they can best maximize their RAM's bandwidth and latency when running 1:1 with the FSB. This is why you'll see so many guides stress the importance of running RAM 1:1 when overclocking a 680i or 780i board.
4. To overclock, you simply raise the base FSB speed. This will in turn raise your processor's speed (and your RAM's speed if you're keeping a 1:1 relationship, but as of now you'll have to maintain this relationship manually in the BIOS, as explained in the front post.) The best way to overclock is to raise base FSB speed in small increments (like 20mhz.) Then reboot, test for stability, and repeat until you've hit a stability wall. Note that since the P7N BIOS accepts a quad-pumped FSB setting, you will want to times your increments by 4, so if you are raising your base FSB is 20mhz steps, you'll want to adjust the FSB setting in the BIOS by 80mhz.
5. Stability: Sooner or later you will find that you have OC'd past a stable setting. Assuming you haven't reached your OC limits (outlined in final point 6,) you can then go back to the BIOS and raise your processor's voltage to raise the stability ceiling. Raise vCore by one or two increments until your stable, then you can go back and raise FSB some more, repeating the process until you've reach one of the two OC limits described next. Note that after a certain threshold, you'll have to consider the voltages of other motherboard components (such as vNB or the GTL reference voltages) to maintain stability. This is where OCing on the P7N gets tricky due to the cryptic BIOS, but you likely won’t have to worry much about these settings until you’re past a 25-40% OC. When you’re at a point where secondary voltages are preventing stability, come back and review what settings others have used to get their chips stable.
6. Limits: The biggest limiting factor to your overclock is heat. Raising your CPU's operating frequency increases its operating temperature, and raising the voltage will increase its temperature doubly-so. As such, it's important to have good cooling when overclocking, and it’s important to run stress and stability tests to ensure that a) your processor is stable and b) that it doesn't get too hot. On a Wolfdale, you should look to keep core temps below 75c, although temp monitors aren't known to be accurate with Wolfdale CPUs yet. The second limit you need to consider, especially with your Wolfdale, is a maximum safe voltage. Previous generation Conroe and Kentsfield CPUs can withstand a lot more voltage than a Wolfdale or Yorkfield, and you will have to keep this in mind. Wolfdales run cool, and if you are shooting for a big OC, you will likely find that you can raise voltages past a safe level before temperatures rise past theirs. Early on, the consensus for a safe 24/7 Wolfdale / Yorkfield voltage limit was 1.45v, but now some are suggesting less. If you can achieve 4ghz-4.5ghz below or at around 1.4v, you'll be in good shape.
Note that while I've covered some of the basics of Core 2 overclocking, I haven't discussed some of the necessary prerequisites, like turning off Speedstep or the like, nor any of the pitfalls, such as the famous nForce FSB hole. You can get this information easily from any Core 2 OC guide. I also haven't touched upon what software you'll want to use to stress test and monitor temps, but that information can be found in many corners around here. Hopefully though, this will get you on your way :)
Good luck and keep us posted on your results.
GEEEEZ !!! :shocked: that's what I call an Answer !!! (notice Answer, with A capitalized). I'm gonna be busy for the next 2 weeks with all the info you gave me, I thank you for that :) I'm gonna take time to understand everything before trying (already said : newbie here). this is not my first attempt in building my own computer (17th to be exact), but this is the first time I took so many time for choosing and buying my components, and this is also the first time I build such a tricky system. P7N diamond is my choice, definately what I wanted and needed, so there's no way that I return it and buy another one. to tell you the truth, this my first non Asus mobo. I was a little scared after reading all the posts at MSI forum with people getting problem from this component or this other one etc... all I can say : study everything, read a lot and choose the right components. I wanna keep this computer for the next 3 years, so it deserve right stuff, and no cheap components. my computer booted perfectly the first time and I didn't get any issues for now (running XP32 SP2). I'm gonna give a try to XP64 and vista, but I'm not in a hurry. I first have to learn "the beast" completely.
things I didn't say about my computer is the cooling system. I'm using a Lancool K7 tower (alu), and I replaced the fans with 3 Noctua NF-S12 1200 (yes, I want it silent). for the CPU, I choosed the NH-U12P. all I can say is this is definately silent, and performing good cooling. I'll post pictures if I can. so, about your advice on a good cooling system, I think it's all right.
the only thing I don't get for now is which bios am I supposed to use ? stay with 1.0, P05 or 1.14 beta ? I guess 1.0 if I don't overclock, but which one if I wanna try ?
many many thanks for your post and for all your explanations. I'll try to learn, and I promise to post my result in here. :up:
No problem, and best of luck with your efforts :) As for the BIOS, it's always best to see how far you can go with what you have before upgrading. If you choose to upgrade, you might consider trying both P05 and 1.1B4, although most have found 1.1B4 to provide for best results.
Btw, while this is not an OS thread and I by no means want to start a discussion on this topic, you will no doubt find Vista x64 to be a better option than XP-64 for a home / gaming rig. I would suggest forgoing XP-64 all together. Vista x64 has better application and driver support, largely in part because it shares a code-base with Vista 32. (XP-64 is built on Server 2003 and is an entirely different beast compared to XP-32. Support and compatibility is far below XP-32 levels or Vista for that matter.) For instance, if you use iTunes or own an iPod or iPhone, you will find native 64-bit application and driver support for Vista x64, and no support for XP-64.
are you reading in me or what ??? :p: yes, I'm using iTunes for my iPod... so, event if it's not what we are here to talk about, I'll follow your opinion and skip XP64 for Vista64. thanks again ;)
Hi all,
I updated the first post with new information, including E8400 settings, tips on NB cooling and software to use, as well as a blurb on two-card SLI.
Ash,
For RAM and NB, the BIOS allows you to set an actual voltage. Only with vCore do you set an increased amount. Your RAM should be set to factory specs and NB can be left on auto unless you are overclocking or if using eight gigs causes the system to be unstable.
I haven't heard of anyone else having this issue with dual channel mode, and if this is your second board and it does the same thing with all four sets of RAM, then my guess is that the BIOS may not be reporting your RAM settings correctly and that you are indeed running your RAM in dual channel mode.
Install two DIMMS in the position I detailed in my previous post, so that you are running with four gigs. In the BIOS, load optimal defaults and verify that your RAM is set to run at 800mhz. Then in Windows, download Everest (see first post for link) and run the Cache and Latency benchmark. Post a screenshot of the results. This will tell you whether or not you're really running in dual channel.
I'm sorry I can't test this but I think 8gb mem is going to require timings somewhat more relaxed than manufacturer ratings.
{cp [n][y][n][y] is correct. Try four gigs first at stock settings (set RAM to 2.2v as required) and you won't need to up the NB voltage. Later, once you've verified dual channel mode is working via Everest, you can try 8. Unfortunately I don't know what settings you'll need to get it to work. It might work perfectly w/o adjusting voltages, or as Loonym suggested, you may need to reduce timings in addition to bumping up the NB. For starters though, lets just verify that dual channel mode is working with the 2 DIMMs.
got an issue with the p05 bios...it crashes my vid drivers...tried all nvidia drivers and even the 169.25 dont work...any ideas???in the mean time im gonna remove my sata drivers:hitself:
Hi
Been reading this thread since I got my new P7N and it helped me to get a fairly easy overclock of my Q6600 to 3.6GHz using the suggested settings in post 1. It also seems to need less voltage to the core than the same chip did on an Intel X38 board.
My only issue is the statement that using the second card for SLI in the white slot works ok with no performance loss. I tried this to reduce temps on my main card (which was a success) however I noticed a drop in my 3dMark06 score from 17600 to 13400 and a drop in my Crysis benchmark from 33.5fps to 24fps (1900x1200, High, no AA or AF).
Just worth noting that not every board seems able to run SLI on different slots without losing performance.
Xraymtb,
It is very strange that you experienced a performance loss when using the white slot for the second card. What bridge cable are you using between the two cards and have you made any changes to the BIOS regarding the PCI-E bus?
That's interesting xraymtb. I haven't tested the difference but I am running my second card in the white slot. I'll try a couple things and post back, thanks for the post.
edit: Could pci-e 2.0 really make that much diff? My cards won't use that but I have some 9600gt coming today.
Hi All,
First, Thank you to x-built-Stan & theYipster for informations about o/c with the P7N diamoand. :up:
I've got the same problem like pacific vortex with the P05 bios. (BSOD with Vista 32 bits, when loading the video driver and the same with linux when loading proprietary drivers from nvidia) I'm using the 1B4 bios for now. :confused:
I could push my Q6600 to 3,5 Ghz, but it's not 100 % stable. (3dmarks06 ok -17348 with 2 GTS 8800-512 SLI, but occt freezes - 2 minutes after the beginning of test, prime95 freezes too after 15 minutes)
My VID is 1.225 (coretemp used) I've tried much much combinations to go to 3.6 Ghz without success. I've never exceeded 75° C in full load @ 3,5 Ghz. I think it will be better to use 450 Mhz FSB with a multiplier of 8, but can't do that with actual bios.
RAM: 800 Mhz (my Rig wont start with ram@1066 Mhz) Timings to AUTO.
Vcore: 1,4 (with from 0,4 to 0,6 vdroop when full load)
Vdimm: 2,10 V
Nb: 1,425
Sb: auto
CPU GTL REF: 65
FSB Term: 25
All AUTO
Any ideas to help me please ? (Sorry for my poor English):shrug:
PCI-E 2.0 shouldn't make a difference. I have to double check the manual, but it may be that the white PCI-E slot is fed from the south-bridge, such that data between the two cards has to transfer over the hyper-transport link.
Still, I don't think this matters, as on my system I score 18,300 on 3DMark and performance in games is equal or better to what they were on the 680i using the same components.
Phard,
On a whim, try a GTL REF around 30-35 and see if that makes a difference. We're still not quite sure what the setting for GTL REF actually relates to, but if it's microvolts and not percentage to another voltage as we had thought, then 30-35 would be in line with what people are setting on the references boards and on the Striker II Formula.
Bonne chance!
I too found an increase although not quite as dramatic as that posted by xraymtb for obvious reasons but a full 5% with all other things being equal except the slot change. It's not a huge difference but it could be the difference to being #1 or 2 on orb for 7800gt or being #6 or 7.Quote:
Originally Posted by theYipster
This is cards at stock in slot 1 and 3
http://images6.theimagehosting.com/pci_e_1_3.th.jpg
And in 1 and 2 still stock no cpu changes made
http://images6.theimagehosting.com/pci_e_1_2.th.jpg
A full 5% increase (or more?) just by slot change matters to me and I think also to others.
Hmm, interesting. It's probably a result of both cards being on the NB vs one on the NB and one on the SB. I'm not sure if I'm going to go through the hassle of switching to test for myself, since compared to the 680i I saw no loss in performance (plus I have that X-Fi installed.) Btw, Loonym and xray, what BIOS are you using?
I'm still on .P05... I moved the xfi to the 16x slot 3 and my pci-e raid card to #4 :D
Loonym,
If you would indulge us, could you perhaps swap back to the stock BIOS or 1.1B4 and see how that affects your 3DMark 06 score with one card in the white slot? I've seen that a lot of people are having issues with video drivers when using the P05 BIOS, and it may be that this BIOS incorrectly configures or manages the PCI-E or HT buses.
hi guys, i decided to sell my Maximus Formula and buy a 780i to make a SLI with mu 8800gt
Do you suggest me P7n over XFX reference (but 6 phase power) and over Asus asus p5n-t ?
Maybe p7n has more quality, even in Oc, i can find @ 232€
Does msi p7n any particular problem ? will be q6600 and corsair twinx pc6400 ver 6.2 cl4 ( i have 4 modules of 1gb) good for this board?
thank you for reply:)
I'm confused... When you are running with two or four sticks, you can't post? Or is that you believe you are running single channel when it is supposed to run in dual channel?
Video card RAM can't be set into any mode, and if everything is running correctly then it should behave completely independently. However, it may be that something is screwy with your card, causing errors elsewhere in the system. If you can, try your card in another machine and see if what happens.
Ash,
I had a similar issue when I first installed my video card. I seated the video card and when I mounted it to the case, it poped out slightly. It was not even noticable. Reseating the card and adjuecting the retaining bracket solved the problem. If willing please try the following:
1) Unplug the unit. Very important.
2) Remove all but 2 mem sticks {CPU, Y, N, Y, N}
3) Remove all but 1 video card - blue slot closest to CPU.
4) Pop out the CMOS battery to reset the CMOS. Move the small tab on the right side on the battery to the right. It will pop out easily.
5) Replace the battery after 10 seconds, Push down firmly until it seats. The unit should now be set to the factory settings. Returning to the factory settings will eliminate any possible configuration errors in the bios.
6) Plug in and power on the unit. Watch the hard disk access light. When the light is flashing the unit is booting up. If you get no video signal, power off and re-seat the video card. Try again. It may take a couple of attempts.
7) If the HDD access light is on steady, then the unit is not posting indicating a hardware problem.
Finally, Check the beep codes when the system boots. I posted the beep codes a while back. You should get a single long beep indicating a good post. Note that the single long beep may be followed by a barely audible "tink" (not beep) for each USB socket (not device attached) in use on the unit.
Hope you find this helpful.
Yipster,
I started a project last week to build a surface stability map of the v1.0 bios.
Basically, I've been setting the CPU GTL (Y-axis) and FSB terminator (X-axis) and running OCCT. The z-axis is a stability precentage based on a 30 minute OCCT run and computed as follows:
Stability = (Test Duration - Run time to CRASH) *100 / Run Time to CRASH
Where Test Duration = 1800s (30 Minutes) and
Run Time to Crash = the time reamining in OCCT when a crash occurs (in seconds)
A test that runs for 30 minutes achieves a 100% score. A BSOD or failure to post receives a 0% score.
I'm about half way through the first surface map. It should be interesting to see how these to settings interact with respect to stability. I'll post the results sometime around the end of the week.
My MSI P7N Platinum didn't boot to Vista with 4GB ram. After I installed this kb929777:http://support.microsoft.com/kb/929777 I added another 2GB and now it's working fine with 4x1GB.Quote:
Originally Posted by ash5555
That sounds great X-Built-Stan. I look forward to your results. Btw, which version of OCCT are you using? The new version (2.0.0.a) has a stronger CPU stability test based upon Prime95's small FFT algorithm.Quote:
Originally Posted by x-built-Stan
Also, Aaron Yuri (the board mod at MSI) just received his P7N and will be tinkering as well. Hopefully we'll get final confirmation on what the integer-settings refer (and whether or not they all use the same metric.) My bet is that FSB VTT and GTL REF are either percentages added or subtracted to a mother voltage (you're theory,) or that they're microvolts (like on the Striker II Formula.)
Hmm, the fact that the same issue has happened on two boards, and that you have tested four sets of RAM each suggests to me that your video card may be at fault. I see that it's a non-standard 1 gig model. Unfortunately I don't know much about it, or if it can present compatibility issues with certain setups. I would consider what X-Built says -- that may be the issue. Otherwise I would search for people with issues concerning your video card and see if anyone has a similar problem.
up.Quote:
Originally Posted by edo
so as i read, there could be problems with 4x1gb ?
Edo,
Please do not bump posts on this thread. It isn't a for-sale board.
The P7N in my opinion is the best 780i board on the market. It can OC easily and has far better components than the reference boards which should result in significantly improved reliability. Remember that the P7N Power Phase design is dual-phase. It's actually more robust than the 6 phase design used on the reference board.
thank you for reply first of all.
Yes that's not a for-sale thread, but, i asked for tech info, and there's no better thread than the owners one;)
do you mean dual-4phase , 8 phase total, like asus right ?
what about 4 sticks of 1 gb ?
Edo,
We're all enthusiasts here and none of us work for MSI. We try to answer all technical questions, but please understand that no one is bound or required to answer them in any set amount of time. Bumping posts, especially when they are only 3 hours apart, is not what I would consider appropriate for this thread.
Anyway, it's a dual four phase design which is designed to match or exceed the robustness of ASUS's 8-phase system. Both MSI and ASUS's power phase design are considerably better than that used on the reference boards.
I haven't tested four 1 gig sticks, although I can't imagine them being a problem. Search this thread as well as the threads at forum.msi.com.tw for people using 4x1 gig setups and any issues they may have. Forum signatures are great place to look.
Thanks.
theYipster,
I know you all are enthusiasts, we are on XS forum and not MSI :)
Sorry for bumping ( i think you mean "quoting" ) , anyway my post would be never noticed and like in others thread i think is not out of regulament, sorry anyway.
I got all information i need
Thanks.
I am using the standard bridge cable that came with the board. When I set up with the white slot I had to use the longer cable as the short one was a few mm too short.Quote:
Originally Posted by theYipster
In my case, I have no need for the PCI slot and I have set Rivatuner to keep my gfx temps to a reasonable level so I will be keeping the cards in the recommended slots anyway. I am also using the original shipping bios if that makes a difference.
Edo,
I run 4x1GB dual channel in my rig. See signature. No problems here.
Xray,
Are you using two bridge straps? When I put my cards (2) in SLI, I originally only had 1 bridge strap installed. After I got lousy 3DMark scores, I checked the MSI manual and diagram showed 2 straps. Adding the second strap got my scores up.
FYI: both cards were installed in the Blue PCi-E slots.
X-built,
Interesting. By how much did your scores go up?
really thank you :toast:Quote:
Originally Posted by x-built-Stan
I don't recall exactly. I do remember that it struck me as though I was running in single card mode. That is why I checked the manual.Quote:
Originally Posted by theYipster
Xray,
I do recall that I also had trouble getting the SLI mode enabled in the nVidia control panel after I had made some hardware changes. I had to set the SLI mode, reboot, and then re-install the video drivers. You might want to try that also.
It's weird that some would have this issue and others don't. X-Built, what is your 3DMark score with SLI correctly enabled and your proc at 3.6ghz? If you don't know of hand, can you run the benchmark and post results?
Interesting evening installing 9600gt parts. What I discovered is that using bios .P05, XP64 will not load with any available nvidia driver. It stalls at welcome screen every time. My only recourse was to flash .114 in order to use new parts.
An interesting point for the GTX/Ultra owners but the GT/GTS only has one SLI connector.Quote:
Originally Posted by x-built-Stan
When I ran the benchmarks I had the SLI indicators visible and they looked normal suggesting SLI was enabled correctly. The difference I have seen does appear to be a lot larger than others have which suggests something else was having an effect.Quote:
Originally Posted by x-built-Stan
hi there.
I've tried some overclocking using theYipster's unvaluable tips and tricks, but I have to admit I'm a little scared when my computer freezes or give a bsod, so I'm really gonna try to learn the beast completely before trying again.
I have some questions regarding the board and the cpu :
- as said in my sig, I'm using E8500 C0@3.16Ghz with 4x1Gb Crucial Ballistix Tracer PC6400@4-4-4-12 2.2v (had to set it manually in the bios otherwise they were only 5-5-5-18) with bios 1.14b. when I look in the bios for my cpu temp, it says 33°, which seems to be quiet good. but when I look at the temp from windows (coretemp, everest, hwmonitor), it tells me core#1 53° and core#2 43° on idle. is it possible to have such a difference between both cores or is there an issue with the reporting softwares ? if yes, what have I done wrong ? it seems to be so hot. I'm currently using a Noctua NH-U12P with fan at 1200rpm and Noctua NT-H1 thermal paste. shall I try Arctic Silver 5 or Arctic Ceramique instead ? (got both at home, so I can try)
- once, I noticed that the bios (using 1.14b) automatically changed my cpu multiplier to 9 instead of 9.5 (so cpu @3Ghz) and ram fsb to 533 instead of 800Mhz. to be honest, I can't remember after which operation this happened. I've had to manually put right settings in bios for them to come back and I left them that way.
- for those having troubles with usb keyboard, I noticed that each time I had to clear cmos and to restart the computer, my usb keyboard wasn't responding before getting into windows. when you load default optimized settings, everything is back to normal and I can use my usb keyboard to enter the bios (have to do it with a PS2 keyboard plugged of course). can anyone tell me which is the f!!@#ing setting in the bios that control that ?
told you, still learning from the beast... ;)
thanks to anyone who can help
- When overclocking, computer freezes and blue screens are to be expected. It just means that your current settings are unstable. Overclocking is all about trial and error and no matter how good you are, chances are there will be quite a few errors.
- 45nm CPUs have well know issues with temp reporting software. It's a common bug and I'm not sure if new software resolves it, or if a BIOS update is needed. Note that you haven't done anything wrong in reading core temperatures. If you haven't already, please make sure you are using the latest versions of CoreTemp, Everest, etc. You may also want to browse the E8400/E8500 thread, where many others have discussed the same issue. I don't know if a common solution has yet been found, but that's where I would look for one.
- As for the multipliers, note that there might be lingering compatibility issues with the current BIOS in supporting half multipliers. If you see that your BIOS isn't maintaining a 9.5 multiplier (even though you haven't changed settings,) you should inform MSI.
as usual, thanks for the quick reply :) so temp should be ok (no freeze at all, stable, fast, etc...)
I forgot to mention 2 things :
- my DRAM voltage is marked RED in the bios after I changed it. normal ?
- cpuz is reporting a 5:6 ratio when I manually set everything to what it should be when normal (cpu 3.16Ghz, RAM 4-4-4-12@2.2v). am I not supposed to get 1:1 ? if so, any clue on how I could get my 1:1 back ? when I leave everything default, it says 1:1, but RAM is only 5-5-5-18 then ?
You should set RAM voltage to the manufacturers specification, regardless of what the BIOS colors it.
With your system at default settings (proc at 3.16ghz, RAM at 800mhz,) you won't be running on a 1:1 ratio. Remember that a 1:1 ratio means that yoru RAM is synced to the base FSBs. Since the base FSB of your proc at stock is 333mhz, RAM needs to be set to 667mhz to run 1:1 (remember that the RAM mhz you set in the BIOS is 2 x base FSB.)
To keep a 1:1 relationship, set RAM manually in the BIOS to 1/2 your FSB. Note that this is really best practice for overclocking. When running at stock, there is no issue having your RAM run at 5:6 ratio (so that it maintains 800mhz.)
For your RAM to run 1:1 at it's stock setting of 800mhz, you will need a 1600mhz FSB. This will give you an overclock of 3.8ghz, considering your chip's multiplier of 9.5
Yipster,
Project M3cury 3DMark Score: 18736
Here is the link:
http://service.futuremark.com/result...eResultType=14
Everyone - Here's a leasson learned.
Gamin's post reminded me of a situation that everyone could learn from. This happened last year when I was overclocking my first system. As is expected, I experienced several BSODs and freezes. What I didn't realize at the time was when that happens you run the risk of slowly corrupting your hard disk. Mostly due to Windows delayed disk writes and caching. Anyway, after numerous occurances I was no longer able to boot Windows - to corrupted on the disk. The result was a complete re-install of the OS and all software - it took a couple of days - what a pain.
After that happend, I got a backup utility (Ghost). Now I always make a backup before I start an OC series - just in case. I really recommend using a backup utility before you start OCing.
FYI: If you buy one, make sure it support the size and type of HDDs that you use.
Great point,
I myself keep my main system environment on a central RAID array (properly backed up of course,) but then also keep a hard-drive called "Bench Drive," which contains a bare-bones XP-32 install plus all necessary stress test, temp monitoring, and benchmark software. I conduct my overclock endeavors on this bench drive, so that my main data and system information is free from any ill-effects caused by bluescreens or corrupt writes.
Yipster,
I also use a barebones harddisk for OCing my builds. But a re-build is certainly more painful than a restore. Additionaly, when the failure on occured on that first build, I thought I was protected because the drive was mirroring in RAID. Well it turns out that they were mirroring and the corruption was mirrored onto the second drive - both were useless. Mirroring is really designed to protect against a physical drive failure, not corruption of a drive's data - Also a good lesson. :doh:
Great tips X-built and Yister about making sure you have a full back up of data before doing any overclocking. I also learned the hard way years ago. Before I do anything BIOS related I make sure I have whatever I need to do a quick re-install of the OS.
Also to Gamin, I have that problem with the half-multipliers not showing up with my Q9300, at least if you change yours you can get it back without swapping CPU's.
I have a Q9450 OEM on the way which has a 8 ratio but is still listed as "under testing" on the MSI site. I'm still going to try and overclock it though. :) With my luck I'll only get half-multipliers for it though!
What's up with a new BIOS release though, MSI is normally quicker than this aren't they?
Ah yes the great RAID-1 debacle :) As a software architect, I've come across my fair share of developers, sys-admins, etc. who are lured into the false sense of security RAID-1 provides. Moral of the story is that no matter your hard drive setup, always backup, backup, backup! :)Quote:
Originally Posted by x-built-Stan
Yipster,
I ordered a 250mm (850 RPM, 20 dBA, 120 CFM) side fan for my rig. That should help get the NB temps down. I plan on installing it this week. I'll take NB temp measurements before and after and post them with pics.
ok, here I am with another problem. the X-Fi sound card... does anyone else is experiencing poor quality, scratches when moving the mouse, etc ? XP32 and Vista 64, same problem : awful sounds, sounds crackeling all the time, ...
I've read many things, and it seems to be a common problem with nforce chipsets since nforce 4 if I'm right.
for now, I've tried the latest vista drivers available from creative website, but it did nothing. maybe I should move my X-Fi to another PCI-e slot (in the yellow one actually). I'm gonna try theYipster's tip with SLI and put it in one of the PCI-e 1x with one of my 8800GT in another slot.
I've read somewhere that a guy solved the problem by putting all the settings in the bios to the right one for his hardware manually, but it was in 2006 with an older mobo... maybe another way to solve the problem... I'll keep ou in touch.
My best x-fi solution has been Youp_Pax drivers by Robert McLelland. The site is here: http://www.youp-pax.org/.
There's also a big thread about this at DriverHeaven forum here
edit: If you do decide to experiment with drivers, be it audio or graphics, do yourself a favor and grab driver cleaner pro from guru3d site or somewhere. A traditional uninstall still leaves unwanted files and registry entries.
well, with 2 NX8800GT Zilent, there's no need to move one to the white PCI-e : there's enough space for the X-Fi to go in the free PCI-e 1x slot remaining. anyway, this fixed the problem, but not completely. this is not worth as it was, but there are still some glitches, scratches, etc...
thanks for the link loonym, I'll give it a try tomorrow after my job day :( have to take some sleep after this 2 days testing session...
also thanks for the tip with driver cleaner pro (will go in my survey utilities folder), but I already made a drivers free XP image, so I just need to redump it :p:
I'm already learning to hate this board with a quad, just like every other Nvidia chipset board I've been through (EVGA 680i SE, XFX 680i LT, Abit IN9-32X MAX).
Tried Mark's suggested settings and plenty of vcore with my X3220 -- it'll get in to Vista occasionally, but I've yet to get more than a couple of minutes in with Prime95 on 4 cores. It doesn't just stop the threads, it locks hard.
Tried upping a few things, but the northbridge is already running hot from feel and I'll be the first to admit I don't know what half those settings are referring to, so I'm a bit leery of setting them sky high. Mark's default settings don't normally get in to Vista. With a bump, sometimes they do, sometimes they bluescreen before getting a desktop up, other times they black screen.
I've seen plenty of "DRIVER IRQ NOT LESS THAN OR EQUAL" and similar bluescreens, and multiple references to 'ecache.sys' if I'm remembering correctly. Nothing very helpful; just general instability I figure.
It's fine at stock, of course... but I didn't pay this much to run stock. I'm pretty pissed off.
Hi Silent,
Standard questions apply here:
what bios are you using?
what are you trying to OC to?
what are you setting for vCore and vDimm?
how far have you pushed other settings?
Lastly, what is the VID of your X3220?
Just for :banana::banana::banana::banana:s and giggles, try a GTL REF between 30-35 instead of 65-67 and see if that improves things any.
Bios is 1.1 B4, taken from the MSI forums. I haven't tried the Performance BIOS as this one was apparently better for quads in most cases.Quote:
Originally Posted by theYipster
I'm trying to hit 3.6Ghz (400Mhz FSB x9) stable with the X3220 and move up from there.
vcore is +0.2 (probably more than it needs) and vdimm is 2.1v which is suggested by Patriot (at 5-5-5-22 2T, not pushing the sticks at all, just trying to get stable first)
I just bumped things a notch or two above your recommended. It's very hit and miss, mostly miss right now. So far this isn't doing any better than the 680i LT did, and it cost a hell of a lot more, so I was expecting /some/ improvement.
I don't believe this board is reading my VID properly; the IN9-32X MAX and the 680i LT said it was 1.18, this one seems to think it's 1.24.
Cooling is a Black Ice Pro II with a pair of Panaflo low speeds at ~8 volts on a rheobus and a Swiftech Apogee GT with AS5. I've cranked the fans and had no better results. Temps are higher on this board, presumably due to its setting the vcore higher, but they don't top ~45C load at stock and I haven't seen higher than high 50s C when I've managed to get it to do anything at 3.6Ghz.
Silent,
I have never seen a G0 Quad with a VID below 1.2. In cases when they've been reported, the user usually had speed-step turned on or some other issue which can cause VID to be reported incorrectly. I don't recall the differences between your Xenon and the G0 off hand, but if they are more or less the same chip, I wouldn't be too sure if 1.18 is accurate.
For your next step, I would raise vCore to +.3 and work down from there. It may be much more than you think you need, but stable Quad settings with this BIOS do seem to rely heavily on lots of vCore. This may be in relation to GTL Ref and what it takes to get that stable. In my experience, there appears to be a correlation between vCore and GTL ref. GTL Ref with the latest BIOS does seem to be a bit more finicky than vCore, such that GTL Ref may require a certain range for stability -- thus requiring a higher complimentary vCore.
With 1.1B4, I need a vCore of +2.5v to maintain stability. This is higher than I think is ultimately required by my chip, given its ability to bench at 4.05ghz at 1.55v. MSI is aware of this issue, and I'm hoping I'll be able to reduce it either with a future BIOS update and/or by playing more with GTL Ref when I have an opportunity to sit down and do extensive benching.
I managed to kill Vista at some point, cause it won't boot at stock now. Time to play the reinstalling game.
Everyone,
For those wondering, here is a thorough and popular explanation of what GTL Ref does.
http://edgeofstability.com/articles/.../gtl/gtl1.html
I've seen a lot of people set GTL Ref on this board when using dual cores, and I wonder whether or not that's necessary (as GTL Ref is intended for quads.) For those with overclocked dual cores, was setting GTL ref necessary in order to achieve stability? What happens if you put it on auto?
I think you are trying to overclock an already unstable os. Those look like trademark errors of running and or installing vista w/4gb (+) mem. Did you apply the hotfix?Quote:
Originally Posted by silent-circuit
I'm also wondering if anyone here has or is using bios A7510NMS.116? I received it to test with my 9600gt/XP64 issue. It still has the multi issue with kentfield and I haven't done any oc except my standard conservative 3.2 so I don't know much about that aspect but I'm curious if it could solve the multi problems people are reporting using 45nm quads.
Please put that up (.116) somewhere, as it seems to be more recent than anything I currently have access to.Quote:
Originally Posted by loonym
Also, the OS was unstable at that point, but I reinstalled, patched up to near current via SP1 (which so far as I can tell includes the hotfix, and the hotfix returns "not applicable to your system" after installing SP1) and tried to OC again only to get the exact same error within 10 minutes. Not happy with this board at all.
I'm also running in to a number of USB related bugs which hamper use even at stock. One of my keyboards (Razer Tarantula) wouldn't even be recognized as a keyboard by the BIOS until after a firmware update. Now if I plug anything in to the hub on it it still isn't. My other keyboard (Apple Pro Keyboard) won't be recognized if I plug /anything/ else in to the other USB ports at the same time as it. The only way I've managed to flash my BIOS is by connecting my USB stick to one of the ports on the hub on the Apple Pro. These issues are persistant with every BIOS I've tried (1.0, 1.1B4, 1.1P5). What the hell, MSI? This is :banana::banana::banana::banana: I haven't had to deal with since USB was new! These keyboards work with literally every other BIOS I have ever tried them with, so what are you doing that's so different? (and clearly... well... wrong.)