MSI P6N SLI-F

[crapito]

update on my P6N FI w/E4300 combo...

1400/350FSB CPUv +0.100; NB 1.40v; SB 1.6v; FSB VTTv +12%; PCI-E 105 & manual RAM timings (400DDR) = 3.15, runs WoW fine

my personal, geek-speak shorthand.

i've run much higher (like 400 FSB), but feel the heat and possible risk unnecessary as 3.15GHz runs anything i want perfectly fine (for now ).

btw: i use a Tuniq Tower 120 CPU cooler and Cooler Master RR-NCW-L4E1-GP Blue Ice
Pro Northbridge Chipset Cooler (just in case)

 

[jonks]

I'm not an expert at OCing, but is there a reason you're feeding so much juice to the southbridge? I thought it was the least instrumental part of the chipset in terms of performance.

I still haven't been able to get over 3ghz regardless of voltages but I'll keep trying with successive bios releases.

 

[crapito]

Originally posted by: sirjonk
I'm not an expert at OCing, but is there a reason you're feeding so much juice to the southbridge? I thought it was the least instrumental part of the chipset in terms of performance.

I still haven't been able to get over 3ghz regardless of voltages but I'll keep trying with successive bios releases.

if you are asking me, 1.6v is only +0.1v more than default and the lowest option over default, thus i don't geel like i'm "feeding so much juice to the southbridge." i don't know if i need any extra voltage on the Southbridge, but +0.1v is a small enough amount that i don't want to waste my time testing a lower, possibly unstable amount. +0.1v isn't going to fry or overheat anything, so why not? besides, i read on some overclocking site/review that +0.1v was the preferred amount (for whatever reason). it works for me, so i don't question it...

good luck on getting over 3GHz. for whatever reason(s), i couldn't run any FSB over 290 and still play World of Warcraft for the first week or so after getting my rig setup. now, 350+ is fine. why? who knows... it works. same BIOS and hardware, though newer video drivers (ATi Vista 32bit) btw.

peace.

 

[Assimilator1]

Spanki
It isn't contradictory at all ,when I said 'typical' I meant a typical overclock (hence 'can hit') not typical owner
Nm ,thanks for the info anyway

I'm going to be running a C2D at the maximium the mbrd/CPU will overclock to ,though I will only increase vcore very modestly.Typically I don't exceed a ~5% boost in vcore ,depending upon temperatures & the amount of extra speed I gain.
I'm planning on building this rig in about 1mths time ,& getting a E4400 or maybe the E6320 (when they come out) hence wanting a semi-decent FSB,mid 400s is good.
I'm new to C2D but my 1st (& not last) overclock was a Pentium 200MMX at 225MHz ,so I have a little experience their.

Cooling is something I haven't researched yet so I'm undecided atm ,but I will be buying a 'decent' aftermarket cooler ,any recommendations welcomed

 

[Spanki]

Someone recently called the thread over at hardforum The Mother of all MSI P6N threads (or something like that - it's up to 20 pages and counting).

If you check the first post in that thread, I've been trying to accumulate various P6N-related links and info, including end-user reported memory usage, as well as after-market cpu HSF installations (more recently added). If you wade through that thread, there's a lot of good info in it from both SLI-FI and Platinum owners, related to overclocking and all kinds of stuff.

 

[Assimilator1]

Cool ,thanks ......but 20 pages!:Q ,I've got my work cut out!

 

[Edge1]

Originally posted by: Assimilator1
...wanting a semi-decent FSB,mid 400s is good.

To be frank, I'm not sure this is the mobo for you then. Not to PO all my P6N compatriots, but PCStats got theirs to 430FSB, max. New BIOS revisions are supposedly going to up the OC ante, but we're waiting on those. Make no mistake - I love this board and an E6300 modestly OCed to 2.66gHz (380 FSB) is extremely quick, but did take slight CPU and NB voltage increases. What may separate this board fro other 650i or 680i options is its stability and compatibility, with seemingly many fewer RAM and other issues experienced by Nvidia-based board owners.

The P965 boards like the DS3 and P5B Deluxe apparently hit 500 FSB w/some regularity. More mature chipset with many BIOS revisions.

Just want to be shooting straight from the hip here.

 

[BigTuna11]

Wow..guess I need to learn how to post...lol. I hit reply and posted a blank. After lots of reading and playing I finally think I hit the wall at 3.4Ghz with my 6400. It was a frustrating journey at first. Took me a while to realize how important the NB adjustment was, gained lots of ground after bumping the NB and ram volts. I also had a strange issue with mixing IDE and SATA drives. For some reason I was maxing out at and unstabile at 3.1Ghz, I also was maxing on my NB at 1.4v (would not post higher). When I disconnected my IDE drives everything seemed to stabilize. I could jump easily to 3.5Ghz (not Orthos stable) and settled in at 3.4 (Orthos stable 14hrs). I thought it might be a power supply load issue, but two SATA drives run fine. I put the IDE drive back in and went unstable again.

Thanks everyone for posting great info! My first overclock experience and very happy with it so far.

 

[Edge1]

Originally posted by: BigTuna11
Wow..guess I need to learn how to post...lol. I hit reply and posted a blank. After lots of reading and playing I finally think I hit the wall at 3.4Ghz with my 6400. It was a frustrating journey at first. Took me a while to realize how important the NB adjustment was, gained lots of ground after bumping the NB and ram volts. I also had a strange issue with mixing IDE and SATA drives. For some reason I was maxing out at and unstabile at 3.1Ghz, I also was maxing on my NB at 1.4v (would not post higher). When I disconnected my IDE drives everything seemed to stabilize. I could jump easily to 3.5Ghz (not Orthos stable) and settled in at 3.4 (Orthos stable 14hrs). I thought it might be a power supply load issue, but two SATA drives run fine. I put the IDE drive back in and went unstable again.

Thanks everyone for posting great info! My first overclock experience and very happy with it so far.

That's a nice job on your first overclock! I'm on my first and can't get past 380FSB (x4=1520), although I haven't tried real hard yet. Your temps are darn good for bumping voltage the way you did. What are you using to measure temps?

I got to Windows but failed Orthos after 20 seconds after trying 400FSB, +0.15v bump to CPU (1.45v), +.05 bump to NB (1.35v), and changiong RAM voltage from auto to 2.1v, auto timings.

Idle temps also jumped a good 4C, from 33-34 to 37-38 using CoreTemp.

 

[Assimilator1]

Thanks Edge1 ,I was thinking this mbrd is on the borderline of the FSB I want ,everything else is alright though.
Unfortunatly the 965 mbrds are no use to me ,I have 3 IDE drives hence looking at the 650s.Though I could go for a 975 with a PCI IDE card added (it must have at least 1 native IDE channel for me).

I'm also trying to find out what the Abit 650 mbrd is like now that its out ,& it's only slightly more than the MSI mbrd.

Btw what's 'shooting from the hip' mean?

Also what's 'strap' mean in relation to FSB & RAM speeds?

 

[Edge1]

Originally posted by: Assimilator1
Unfortunatly the 965 mbrds are no use to me ,I have 3 IDE drives hence looking at the 650s.Though I could go for a 975 with a PCI IDE card added (it must have at least 1 native IDE channel for me).

Ah, I see. Then a 650i board might do nicely for you. I do love this board - its seems rock stable when dialed in. And OC performance is bound to improve - there's a circulating 1.22 beta bios (for Platinum) that is supposed to boost OCing. Well, if that's the case I hope you go with this one (SLI-Fi or Platinum). Gary from this site sent me a beta bios called "P04" which may be the same thing. I'm a bit too skiddish to flash given that I'm pretty happy right now.

Btw what's 'shooting from the hip' mean?

Well, I guess I meant "being completely honest", so perhaps that might be the wrong phrase...

Also what's 'strap' mean in relation to FSB & RAM speeds?

That's one I've heard tossed around but can't really answer.

 

[yacoub]

 

[yacoub]

This board is one I have my eye on for a nice affordable safe build.

 

[normalicy]

Just a note to everyone, I noticed that while I was overclocking higher voltages were not necessarily good. Especially if you're running the stock NB heatsink & Intel Heatsink. My system ended up being unstable because of heat. I couldn't run at stock speeds even with a .1 increase on the cpu. Dropped it back down & it went straight to 3ghz with no problems.

 

[BigTuna11]

Edge,
I found getting by the first set in Orthos was helped by leaving my memory timings in auto mode and bumped the mem voltage 2.25v. I only tightened the mem settings after I was stable. If still not passing then bump the NB to 1.4v or even 1.5v (my current setting), this had the most dramatic effect on my overclock stability. Using the included fan the NB stays nice and cool at 1.5v. I did notice my board (same as yours) did not like the .15v increase on the CPU, it was actually worse for stability than .1375v (my current setting). For me .1375v was the sweet spot. Everytime I increased .15v I had posting and booting glitches when I went higher on the FSB.

Here are some screenshots of my Orthos at more than 24hrs and the temps while it was running. There is also a screenshot of my Everest sensor readings and CPU-Z.

http://www.myhudsonplace.com/misc.html

 

[Edge1]

Originally posted by: BigTuna11
Edge,
I found getting by the first set in Orthos was helped by leaving my memory timings in auto mode and bumped the mem voltage 2.25v. I only tightened the mem settings after I was stable. If still not passing then bump the NB to 1.4v or even 1.5v (my current setting), this had the most dramatic effect on my overclock stability. Using the included fan the NB stays nice and cool at 1.5v. I did notice my board (same as yours) did not like the .15v increase on the CPU, it was actually worse for stability than .1375v (my current setting). For me .1375v was the sweet spot. Everytime I increased .15v I had posting and booting glitches when I went higher on the FSB.

Here are some screenshots of my Orthos at more than 24hrs and the temps while it was running. There is also a screenshot of my Everest sensor readings and CPU-Z.

http://www.myhudsonplace.com/misc.html

BigT - I'm with you. Thanks for those images - temps look great on a nice OC.

CPU definitely did not like the .15v bump. Temps too high (63C!) and unstable. Also agree about memory timings. I didn't push NB which looks like my next step, provided temps are acceptable.

Here's where I am:
CPU - 2.73gHz (390 FSB), 1.35v (+.05)
NB - 1.35v - actively cooled
FSB VTT - 4% (I honestly don't know what this is, but just chose this value over 0%)
RAM - auto voltage and timings = 2.1v, 5-5-5-15, 2T
Idle temp: 34C
Load temp: 57C avg I'd say, with lows of 50C and highs briefly touching 59-60C.

I'm almost at my goal of 400FSB. Your memory voltage is higher (2.25), and NB as well (1.5). My only concern at those voltages is temps.

Enjoying this board/build immensely:thumbsup:.

 

[BigTuna11]

Edge,
My NB was warmer to the touch without the fan at stock than with the fan at 1.5v, so I'm not really worried about the heat. My ram was rated for 4-4-4-12 @ 2.2v, so running timings looser and volts at 2.25 isn't a huge increase. The ram is warm, but not hot to the touch. I did try it higher initially, but wouldn't post at above 2.35v. Noticed in Orthos that failures or freeze ups on large files (first set) seemed to be mostly not enough ram volts or NB volts or ram timings to tight,. Failures on small files seemed to be CPU related, not enough CPU volts. My board looks to need a bump in CPU volts to go higher on the FSB, but gets wacked out with more voltage. I'm completely stable at 3411, if I go to 3450 I fail small ffts in Orthos. So I think I'm maxed out . I do think the board is limiting me from increasing the CPU to 1.5v or greater, but I'm not looking for bragging rights, just a nice stable overclock . I'm there, so happy with the board.

I'm just starting out with overclocking, so don't hold me to any of this as they are just my observations with my setup. Here is what I would try.

1. Since you have a fan on your NB, bump it to 1.45v or 1.5v
2. Bump your ram to 2.25v and leave the timings in auto. Later after everything is orthos stable, then tighten the timings up a little.
3. Bump your CPU to the .1375v increase value. This seems, at least in my case, to be the highest stable increase. Anything higher caused problems with stability.
4. Set fsb vtt to 2%. I saw no difference in anything by raising it higher. I read somewhere that raising this causes more heat to the FSB components and no significant gain in performance.
5. Bump you FSB above 1600 and keep your ram speed at 800 1:1. Run Orthos for an hour, I bet your stable.

I see no problems with my temps at these settings (YMMV) and you should have your best chance to increase your FSB and get stable. After your stable in Orthos, you can start backing them down one setting at a time and dial in your optimal settings. Or if things are looking good at 1600 FSB, increase the FSB and RAM 1:1 until Orthos fails. Then back them down until stable again. This should be your top end.

Have fun .

 

[Edge1]

You've done your homework - doesn't sound like your first overclock. I will indeed give this a try, hopefully tonight. I was thinking of a bump in NB to 1.4 for starters, adjusting FSB VTT back to 2%, and trying things from there w/o a CPU voltage bump. I can bump RAM up too, perhaps incrementallyto start, say 2.15 or something. My RAM is advertised as 1.9-2.1v, 4-4-4-15, so I don't know if that has any bearing on whether I can go above 2.1. I'm no RAM expert, so I'll leave timings on auto.

If that doesn't fly, and temps are stable, I'll bump NB a tad more (1.45) and retest.

What kind of case are you using?

PS. I haven't done anything with the SB at this point.

 

[BigTuna11]

Seems that your ram is a little better rated than mine, it needs less voltage to have the same timings. If you try it at 2.15v and you pass the large file tests in Orthos you should be good with your ram. If they are failing you might try to bump the NB more first as you are already cooling it and that was the most significant adjustment to get me past those tests. For my purposes, the real world difference between 4-4-4-12 and 5-5-5-12 is something I probably won't notice. So, I left my ram there to make the system more stable. While fully auto my timing were read in CPUZ as 5-6-6-20. After getting stable I manually locked the first 4 timing lines at 5-5-5-12 and left the rest auto. I did try 4-4-4-12 and it gave me errors after a couple hours of Orthos.

The I bumped the SB up one increment, not sure if it was meaningful or not. I didn't play with it much. I just figured a little extra might help the cause. My SB is still very cool, so the extra .1v doesn't heat it up at all.

I'm using the case in the link below.

http://www.newegg.com/Product/Product.aspx?Item=N82E16811163043

I have a home theater and needed a desktop style case for the shelf. I did have some trouble with the height of my CPU cooler, it wouldn't fit under the center case support bar. Had to notch out a channel in the CPU heatsink fins to fit it under the bar..PITA. It is a little crowded, system temps are not to bad at 29C. The fans are not the quietest, but there are 7 running inside the box. I'd rather a little extra noise and have good air flow than risk higher temps by slowing them down.

I'll be interested to see how you make out with your adjustments, good luck!

 

[Edge1]

Thanks BigT. I had an interruption which required the full use of my pc, so I had to cut orthos short for now. Stable for 3+hours with the following adjustments:

CPU bumped one more notch to 1.3625v
NB bumped to 1.4v
Ram manually set to 2.15, timings on auto showed up as 5-5-5-15
FSB VTT dropped a tad to 2%

Only problems were CoreTemp results, showing high 50s into the 60s and even up to 62C at certain points. A little warm for my liking, but this PC certainly won't be running 24/7 with loads like that. Also, TAT never hit 60C which is curious - topped out at 58.

Got my work done, so now I'll run Orthos overnight I suppose. Thanks again - you've been a big help. I agree that the NB is probably the most critical part of this equation.

PS. That's a nice looking case for a shelf. Mine's got fans galore as well...sounds like an idling jet just before takeoff, but what the hey.

 

[BigTuna11]

Glad to hear there was some progress. Sounds like you have more room to play .

I use TAT for my temps, since its made by the guys who did the CPU . Everest seems to agree pretty much with TAT.

I did have dreams of a silent PC, but after this experience I don't see how that would be possible without a lot of specialty stuff and $$$.

I just signed up over at the other forum with the 20+ page thread. Was hoping to find some info on the 1.22 bios, wondering if it's worth installing.

 

[Edge1]

That's a great thread. Spanki's the man:thumbsup:. He's helped a lot of folks out by consolidating so much info.

Funny - Everest acts like SpeedFan to me - way low. I agree about TAT. If those are the real temps, I'll take it. Makes sense given that, as you said, they made the chip.

 

[yacoub]

Spanki - the Hard forum is queer about GMail addresses so I can't register there (and I don't remember my account from years ago) so I'll ask here: Have you no reports from folks with Ultra 120 heatsinks? I'd love to know if the Ultra 120 and the Ultra 120 Extreme will fit before I go buy one. Thanks.

 

[Spanki]

I haven't personally seen any reports of the Utra 120 being used yet, so I can't say for sure - sorry. The only real issue mentioned so far has been that there's some circuitry near the cpu on the underside of the motherboard (at least on the Platinum, but I assume the SLI-FI is the same)... some HSF backplates would contact with this, and so might require some dremmeling/routing out to make room.

There doesn't seem to be any real issues on the topside of the board clearance-wise.

 

[Assimilator1]

Does no-one know what 'strap' means in relation to FSB & RAM speeds?