Forward: official GIGA 965P-S3/DS3/DS4/DQ6 beta BIOS

I have followed this thread for several months and thought I would post my finding. F5&F6 S3 bios tried, F5-F8(final) DS3 tried all cannot post beyond 450 FSB under any circumstances (see My Sig for rig details).

Why I believe this is a board limitation (D9 ram or is this really a ram issue at all!)


1. Ram Elimination…..Set CPU to 6X multiplier(to ensure CPU max is not reached), increased ram “multiplier” then increased FSB until ram hit 490 (980 effective). The FSB never exceeded 370. (up to 2.4Volt) *** Confirmed ram goes to 980 effective.

2. Then CPU elimination
Set multiplier to x6 (from x7) FSB until it would not post a between 445-455 (still). (ram at only 900 below what I tested it for)*** High probability this is not the CPU as it only was hitting 2.7gig

3 For the record played with all Voltages Ram timings, with no joy spent days ( as I enjoy it, this is not a complaint)


It has become VERY clear the 450 FSB+ issue will not be solved with a bios, and here is why.


1. People with the C1 chipset usually cannot go over 450FSB while those with C2 usually can (why update the chipset if C1 was ok)

2 They have been trying for months to fix this, if they can’t by now they never will.


I can reach 3.0Gig with this e6300 100% stable on this board which is not bad at all for the "price" of the board. I would like more but I think people are over reacting here, Overclocking has never been nor will EVER be guaranteed and many other boards are doing a lot worse!

i've got my patriot 800 up just over 1000 using divider on the f8i bios. my limit is cpu at 3.6. no water, just air so can boot but unstable at 3.6. if only i had some H2O!!!! patriot is micron D9 i do believe so your theory may hold...interesting that with winXP my mem timings ran solid all day long, now just upgraded to vista and had blue screen! had to knock back timings to 5-5-5-12 and now stable all day. vista x64 must handle memory different. guess that makes sense......more interesting vista stuff, my memory "score" in vista was 5.4 with the 4-4-4-10 timings, after relaxing to the 5's is now at 5.8. Not sure what the "score" is or relative to, but very curious to go up with looser timings....i think i'm out of interesting-ness.....

Quote:

---

Originally Posted by Battleneter

Why I believe this is a board limitation (D9 ram or is this really a ram issue at all!)

---

The only other possibility is that the Gigabyte engineers just dont know what they are doing with the BIOS. Asus P5B C1s can do 500+, as Im sure are some of the other 965s. I have a DS3 C2 that does up to 516 Mhz orthos stable however my beef was that the board on a cold boot would not come up overclocked and a series of reboots to get it to save and back into windows xp would occur (even though I was prime stable for over 16 hours for two runs).

Quote:

---

Originally Posted by theDUD3

i have a S3 mobo and gigabyte hasn't released a new bios since october, does anyone know of any beta bios around? or could i just flash my S3 with the DS3 bios?

thanks!

---

yes

Quote:

---

Originally Posted by dnottis

The only other possibility is that the Gigabyte engineers just dont know what they are doing with the BIOS. Asus P5B C1s can do 500+, as Im sure are some of the other 965s. I have a DS3 C2 that does up to 516 Mhz orthos stable however my beef was that the board on a cold boot would not come up overclocked and a series of reboots to get it to save and back into windows xp would occur (even though I was prime stable for over 16 hours for two runs).

---

I have seen your post, I am reporting only on the 450FSB issue.

There is more to a board than the Bios and Chipset. I am now 100% convinced the 450FSB c1 issue cannot be solved via a bios.

what's the c1 and c2? is that intel's mobo chips?

Quote:

---

Originally Posted by burn what?

what's the c1 and c2? is that intel's mobo chips?

---

Yes the latest boards are coming with C2, that are 450+FSB capable, it is a pity not everyone is quoting what they have when they post, this is extremely important.

Quote:

---

Originally Posted by Battleneter

Yes the latest boards are coming with C2, that are 450+FSB capable, it is a pity not everyone is quoting what they have when they post, this is extremely important.

---

what do you mean? everyone seems to have a pretty decent sig. i'm a nube to forums, but they seem fine to me..??

Quote:

---

Originally Posted by burn what?

what do you mean? everyone seems to have a pretty decent sig. i'm a nube to forums, but they seem fine to me..??

---

Lol this is not an insult, it is usually missed read back yourself.

Quote:

---

Originally Posted by burn what?

what do you mean? everyone seems to have a pretty decent sig. i'm a nube to forums, but they seem fine to me..??

---

lol you asked yourself

Quote:

---

what's the c1 and c2? is that intel's mobo chips?

---

there it is

i dont have it listed in my sig but i figure by saying rev2 it's pretty much implied currently

Quote:

---

Originally Posted by ziddey

lol you asked yourself there it is

i dont have it listed in my sig but i figure by saying rev2 it's pretty much implied currently

---

Forgod sakes dude, I said "Usually" and it still is!!! what are you mentally retarded?

And I wasn't directing that comment at any one person.

you guys lost me, but i realized i had the wrong bios rev in my sig so i'm now up to date...

Quote:

---

Originally Posted by Battleneter

Forgod sakes dude, I said "Usually" and it still is!!! what are you mentally retarded?

And I wasn't directing that comment at any one person.

---

lol you're calling me retarded? I wasn't even talking to you!?!? are you retarded?!?!

but in all seriousness, i've had a little too much tonight but still think I'm fairly comprehensive?? If I'm not reading this situation improperly, what you're saying is for people to state in their sigs c1 or c2. And burn what? is the one asking what the hell you're talking about, and I was trying to clarify it to him. And then I just interjected that I didn't explicitly state my revision since I figured rev2 ds3 is assumed to be c2. I guess I'll read back a few more posts to make sure I'm understanding you correctly. I hate it how I'll get a ton of email notifications but when I click it sometimes jumps me 3-4 posts ahead of the last one I saw. So I pretty much just assumed based on the 3 posts abov e mine what you were talking about.

If you'd like though, I'm 100% retarded vegetable!!!!!!

edit: and for the record in that other post I was in no way attacking you at all!!! why you so on the defense??

edit2: 2 beer bongs later, I checked back and still think that I got the situation properly but you misinterpreted my last post as an attack on you? I was trying to state that he himself asked between the c1 and c2 being intel chipset revisions (or actually rather asking what c1 and c2 were referring to, which you answered in your following post). Then he said he still didn't follow, to which I told him to go back and look at the posts (two of them) above that. I think you somehow thought I was quoting that to attack you.

Or are you trying to say that rev2 isn't sufficient enough to state c1 vs c2 (which to my understanding, there are tons of rev1's with c2 since it was introduced somewhere around the end of summer / early fall.)

But for what it's worth, I completely agree with you that it's a board issue rather than ram. Possibly even more than just c1 vs c2 differences. I remember reading something about an additional resistor or something being added to make it more "compatible" with d9's and higher mhz scaling.

For what it's worth, I've got a very recent ds3 (rev2). Right off the bat, I brought the fsb up to 400 at 1:1 and set timings and voltage as speced (ram is 800 5-5-5-15 2.1v). Installed my os and began to experiment. Due to temp issues with the cpu that I will address soon with watercooling, I just bumped the mem divider up from 1:1 to 4:5 and got 1000. After seeing this be a success, I dropped timings to 4-4-4-12 and added an additional 0.1v (for 2.2v), I began testing. It was absolutely stable and I haven't bothered to drop the vdimm any to see if it works any lower, since I figure 2.2v is fine for everyday use, and I could just eventually scale higher to see what else it was stable with. Also, I've noticed that the vdimm set in bios is actually about 0.1v lower than what is actually set. This is seen both with speedfan as well as with dmm measurements, at least on my board.

Hope this helps you out some?? And sorry if you took things the wrong way

Quote:

---

Originally Posted by Battleneter

1. People with the C1 chipset usually cannot go over 450FSB while those with C2 usually can (why update the chipset if C1 was ok)

---

I have to disagree with you there. I have the S3 and my chipset is rev C1 and I'm running 24/7 at 490x7. Ram at 980. I'm only running 2x512MB kit of ram though which if it's a D9 issue and it's related to 1GB sticks (2x1GB kits) then the point still stands about the issues. It's not the chipset revision. Also, I have been able to post and boot into windows at 515 and lower cpu multiplier.

http://www.thepengellys.com/stuff/cp...etrevision.jpg

If you look at the bottom left of a DS3, there's another revision number on it. I got a C2 board but it read revision 1 on the bottom left and I'm not sure if Gigabyte's referring to C2's as being the revision 2 or those with revision 2 printed on the actual board.

Quote:

---

Originally Posted by HeadOn

If you look at the bottom left of a DS3, there's another revision number on it. I got a C2 board but it read revision 1 on the bottom left and I'm not sure if Gigabyte's referring to C2's as being the revision 2 or those with revision 2 printed on the actual board.

---

C1/C2 is the chipset revision #, look at my post above yours, the cpuz screenshot shows the revision #.

Quote:

---

Originally Posted by HeadOn

If you look at the bottom left of a DS3, there's another revision number on it. I got a C2 board but it read revision 1 on the bottom left and I'm not sure if Gigabyte's referring to C2's as being the revision 2 or those with revision 2 printed on the actual board.

---

CPUZ will show the chipset revision - not PCB or board revision. The DS3P is the version 2 of the board. Here is my DS3 with chipset C2. My Asus P5B-e is C1.

http://img223.imageshack.us/img223/3433/3500140vmy2.jpg

Quote:

---

Originally Posted by HeadOn

If you look at the bottom left of a DS3, there's another revision number on it. I got a C2 board but it read revision 1 on the bottom left and I'm not sure if Gigabyte's referring to C2's as being the revision 2 or those with revision 2 printed on the actual board.

---

Alright!! Now we're getting somewhere!!!! Recently, gigabyte announced a new revision 2 board, like t he one I have. I'm not sure intimately of what the differences are, but I can say that the rev2 has a different nic than the rev1 ( in fact rev2 is same nic as in s3 boards). At least this is what I was informed of (8056 vs 8053 I believe). I know that gigabyte used c2 revision/stepping or whatever you call it p965 chips in later rev1 models of the ds3. So it goes with reason that rev2 will have only c2 chips.

Personally, I haven't had any "d9 issues" or maybe I have but just don't know it. But like I mentioned, I'm using 2x1gb sticks of d9gmh at 4:5 multipler at 400fsb for 1000ddr2 at 4-4-4-12 and 2.2v set in bios (2.3v actual)

So it does look like it is very well cured in newer versions of ds3, as was said about a massive rma to take place soon.

I will also note that I am on f7 bios (as shipped from factory) and am hesitant to flash to f8 after seeing all the testimony (although these results sound like they're from people with earlier ds3's)

The only gripe I have about this board is the failed o/c problem and the coldboot cycles that follow with it turning on and off repeatedly every 3 seconds. For me, this is usually an infinite cycle and I'd have to cut power to the psu and then it'd fix itself.

I also run pci-e locked at 110mhz if that means anything to you guys (not saying it doesn't work at 100. just saying I stress everything more in general)

Quote:

---

Originally Posted by dnottis

CPUZ will show the chipset revision - not PCB or board revision. The DS3P is the version 2 of the board. Here is my DS3 with chipset C2. My Asus P5B-e is C1.

http://img223.imageshack.us/img223/3433/3500140vmy2.jpg

---

actually dnottis, I have great great great respect for you (I've spent the better part of today just reading through posts here to learn more about different products for watercooling), but I think the ds3p is a nameplay being done by gigabyte. The board itself, if you look at photos or specs, actually more closely resembles another gigabyte board (I can't remember off hand but if memory serves correct, the dq) but without the heatpipe system. I'm currently involved in a [H] thread about it

Quote:

---

Originally Posted by MadMikeSS

I have to disagree with you there. I have the S3 and my chipset is rev C1 and I'm running 24/7 at 490x7. Ram at 980. I'm only running 2x512MB kit of ram though which if it's a D9 issue and it's related to 1GB sticks (2x1GB kits) then the point still stands about the issues. It's not the chipset revision. Also, I have been able to post and boot into windows at 515 and lower cpu multiplier.

http://www.thepengellys.com/stuff/cp...etrevision.jpg

---

I will have to back peddle a we bit as you are right (half right at least).

I have noticed people like yourself running 512MB sticks over 450FSB and with C1. I got a little carried away as I have tried 512MB sticks and had the same problem (maybe just my bad luck). As you stated the D9 issue is 1 gig stick related.

But I will say however C2 seems to "nearly" always go over 450FSB with 1 gig sticks so I stand by that. I personally don't believe the 450FSB D9 exists on C2 (of course cold boot is another matter).

BTW the F8 final bios introduced the cold boot issue on my S3 board as did the all the other F8's where the latest S3 bios is fine (F6).

Quote:

---

Originally Posted by burn what?

i've got my patriot 800 up just over 1000 using divider on the f8i bios. my limit is cpu at 3.6. no water, just air so can boot but unstable at 3.6. if only i had some H2O!!!! patriot is micron D9 i do believe so your theory may hold...interesting that with winXP my mem timings ran solid all day long, now just upgraded to vista and had blue screen! had to knock back timings to 5-5-5-12 and now stable all day. vista x64 must handle memory different. guess that makes sense......more interesting vista stuff, my memory "score" in vista was 5.4 with the 4-4-4-10 timings, after relaxing to the 5's is now at 5.8. Not sure what the "score" is or relative to, but very curious to go up with looser timings....i think i'm out of interesting-ness.....

---

please tell me exact model of your patriots its SL or EP?

Still no new BIOS for DS4 ...

How is the f8 bios? Should i update?

do you have problems right now? if not, then no. what are you running right no. so far it seems to be more trouble than good

This thread has decended into chaos!...please keep things ON T, update - DQ6 2.0 ordered and German dudes confirm money transfer complete. Shipping today, but most likely won't get it until January 3rd.

In other news, check out the awesome new Transformers teaser:

http://uk.promotions.yahoo.com/transformers/

ps. that's Peter Cullen (original Optimus Prime voice) heard at the end.