Source: Richard Swinburne
http://www.bit-tech.net/news/2008/05...-intel-s-p45/1
Source: Richard Swinburne
http://www.bit-tech.net/news/2008/05...-intel-s-p45/1
i have those on my x38 whats the point of anouncing that its complicated it isnt, and they didt mention pulls for ram
Bring it on. Namely Foxconn and DFI P45 boards.
This is nothing compared to the new bios options Gigabyte have in store for P45, you think things are complicated now?....lol.
gigabyte has no options on purpose thier market isnt the smartest its mainstream, get a dfi lt/ut then u will have the options and more now, or learn to hex edit on gigabyte biosesQuote:
Originally Posted by Richard Dower
thats the most uninformed post ive seen today, thanks zanzabar....
Yeah my X38's got that stuff too. You need to be a little more careful, but I don't know that it's reaching "complicated" levels yet.
Personally, I find those settings to be pretty forgiving at the lower levels. Once you start pushing limits, they become more important. But by the time you reach that point, I think you SHOULD have to learn thing or two about the equipment you're using. Just like any other hobby/sport/activity, right?
what do u mean, gigabyte isnt marketing to the overclockers, and they set most values that u dont see in the bios without u seeing themQuote:
Originally Posted by T_M
They are now, they've seen the light.
To make it clear: What Richard describes is a problem of all chipsets for Intel mainboards and not just P45.
For most people it is sufficient to let those options set to auto. If you want to squeeze everything out of the system, the mentioned options may give you another few MHz of FSB or memory speed.
Nothing new and certainly not the most complicated if you've ever played with Socket 939 DFI hardware.
I miss DFI S939 bios - mainly for the buttload of RAM timings.
I'm not interested in tweaking RAM on LGA775 boards as there's hardly any fun in tweaking like mere 2 or 3 secondary timings. Secondly, I hate the quirkyness of LGA775 in general. It's sheer draw of luck if my system boots up correctly, like every 2nd boot or so ends up in a black screen and you'll have to powerdown, switch PSU off and restart. Too often it resets a perfectly stable CPU & RAM OC to defaults...
I'm sooo irked by Intel chipsets...
My point in bringing this to the editors was that all Intel chipset boards which end users can run C2D and C2Q have this issue, FSB has serious limits without tweaking it to keep it in line, especially with quad cores.
Those who do not learn how to adjust skew and GTL are not getting the best from their hardware...;)
What i mean is that you obviously dont know much about Gigabyte in the last yearQuote:
Originally Posted by zanzabar
Gigabyte have made alot of advances in the complexity of their bios options, the "F1" hidden options is much better now for non knowledgeable overclockers then ever before. The hidden options make even more sense now and moreso when P45 based bios come out. Gigabyte are going to be adding tons of new options aimed at serious overclockers in P45 (screen grabs available)...so for the novice the F1 setting is a must. Personally i'm not into complicated bios settings for the sake of it, especially when Gigabyte and others don't explain what these bios options actually do.
So Tony is right on one level, on another level mobo makers must make more of an effort to explain what these bios options are for and do, or at least have documentation that explains it...these complicated bios options more then ever require mobo makers to help users understand more.
I lol'd.
If this is "complicated" for someone, they shouldn't be overclocking in the first place :shakes:.
As an elitist jerk, i say down with the incompetent :up:
I don't think playing with GTL's is a completely blind art. Once you know what %'s of VTT your chip likes, you can tweak accordingly. Title should read: 'Now everyone can experience the fun (or headache) of a DFI user on the P45'.
TonyQuote:
Originally Posted by Tony
i see what you're saying mate
you are right no doubt
actually my current bench setup is doing my head in
i am lucky to have a P965 chipset board which can shoot waaaay past 600MHz FSB on air cooling so i can easily test E8xxx CPUs and know what i should expect
having said that though this chip i have in the benching rig now is a 520-530MHz CPU on most intel boards and on this P965 i can validate as high as 628MHz FSB on air cooling
now on Gigabyte X48T-DQ6 i've come across some interesting findings and they specifically relate to CPU Skew of +200 >> i've been able to get the board to post past 530MHz FSB....in fact i've been able to get the board to post at 580Mhz FSB thanks to this setting HOWEVER.....the skew is throwing off the signal too hard and i cannot actually load windows at any decent FSB and i've even had reduced stability at 530....actually even 520 that's how much the signal is affected
however +200 setting is the only one giving me these results.........i've experiemented with GTL REF but GB have very simple options which may not be enough.....do you think that i need some more GTL tweaking to bring the signal in line and be able to stabilise reasonably high FSB (close to 580) or is it related to volts or something else?
i'd appreciate your insight in this specific example
Dinos, I highly doubt you will ever see high (600+) fsb on X48(x38). That chipset is great for QX cpus ;)
Hahah. I OCed perfectly without touching GTLs...but I'm sure it wouldn't be the case if I had a quad otoh...or if I wanted better OC/vcore increase. *Shrug* Maybe for my next upgrade I'll look into it.Quote:
Originally Posted by WoZZeR999
For normal limits OC you don't really need to touch the GTL. With correct GTLs you can lower your other voltages, and lower overall system heat. With correct GTL's, you can also hit the upper limits of your chip with 100% stability (for those who like to run a prime marathon and create a space heater at the same time).
you know guys...the issue i see is since the 965 FSB with dual cores has reduced, tRD is able to run tighter on the newer chipsets but at the expense of having to buy CPU's with higher multipliers.
X48 is good around 420 to 480 with a quad with GTL tweaking and maybe Dram skew, after that forget it unless you get a cherry board or CPU..
You would get even higher if you set the board up bang on where it needs to be.Quote:
Hahah. I OCed perfectly without touching GTLs...but I'm sure it wouldn't be the case if I had a quad otoh...or if I wanted better OC/vcore increase. *Shrug* Maybe for my next upgrade I'll look into it.
FSB has hit its limit...hence Intel are dropping it soon IMO
Lol I came from an Ultra-D.Quote:
Originally Posted by CandymanCan
Problem is that I bought only half-decent RAM, so I reached the limits too quickly :(
This is the main reason my next uppgrade (more of a downgrade really ) will be AMD. VTT, GTL and what not has ruined it for me. Sure it's fun to overclock but it's not that fun to me. I wanna find my max OC in like a 1-2 weeks. Not a 1-2 months. I just don't have the time or patiens.
My first board was the Ultra-D, that was freakin' insane! I now have a Gigabyte DS3L and an E4500@3.2ghz. I'll take whoever caters to the noobish OCer thank you. ;) Or just copy settings from here whatever.....
+1, i'll spend a few days doing it and that's that, a few more MHZ isn't worth it i don't care about records.Quote:
Originally Posted by Gambit_2K
Everything can be set to AUTO, you don't HAVE to use the new settings.
I'm sorry, I fail to see how this is more difficult? I typically can leave these settings on auto and hit 480FSB with my quad with ease which is also much higher than necessary as I tend to boot at much lower FSBs and probe the limits. I suppose with a dual-core this could be an issue but how many users strive for 600FSB?
Quote:
Originally Posted by zanzabar
Gigabyte has changed their mentality on overclocking considerably. P45 should reflect this, hell they have a board that caters direct to us :cool:Quote:
Originally Posted by zanzabar
that article is very misleading... it makes it sound like ocing p45 is a nightmare and you need to spend hours and learn about electronics and ee before you can overclock it at all...
both of which is not true at all
I would put this article akin to something fudzilla or INQ would write. This article could go right after "We think Big Blue is stopping support for the overclockers, here is our *proof*" (in reference the to recent Nehalem article). GTL's are not blind, you don't just throw random numbers at it. It even says in the article where to start. If the 65nm proc needs somewhere around 67%, start at %66 and work your way to %68. Again, this only applies if you want a heavy o/c. For most overclocking, AUTO will work fine.
this review has the asus bios for the p45, it looks like the maximus with a few settings gone
http://sg.vr-zone.com/articles/Asus_...ew/5758-6.html
its still missing a bunch of ram options
Personally, I will not miss the FSB at all once it's gone and Nehalem is here.Quote:
Originally Posted by Tony
Tweaking GTL settings to get the most out of hardware is a very frustrating experience but can be rewarded with higher stable OC's.
I agree with someone else that said that many P45 boards are finally offering BIOS tweaks on the level that DFI has for years.
However, most people, some of whom have posted here, will simply ignore them and use the hammer that is voltage. It reminds me of an old saying... "If all you have is a hammer, then everything looks like a nail!" :yepp:
It's not saying that... it's saying that if you want to get the absolute max OC out of your system, you do need to pay attention to these electrical engineering aspects of your system.Quote:
Originally Posted by saaya
I would say from some of the responses here that some overclockers are only concerned with the max OC they can get with voltage tweaks. They are missing out. :shrug:
Time and patience is needed if you want to OC you're system. For the SOHO users, they wouldn't care anyways.
not reallyQuote:
Originally Posted by kiwi
Pt1t boots on air cooling at 600MHz FSB with his E8500 (Gigabyte X48T-DQ6)
i'm sure we can tweak FSB to get there
i've been benching 600+ FSB with E6850 and even E6700 on X38-DQ6
i've also benched another of my E8500s with Asus P5E3 DLX/Gigabyte X38T-DQ6 at 560MHz 32M Superpi on air as well
so under cold it will be even better ;)
never heard of that one before hahahaha, awesome :lol:Quote:
Originally Posted by virtualrain
its not giving the impression that ocing p45 is a nightmare and you need to studdy vtt and gtl and clock skew tweaks? then whats this?Quote:
Originally Posted by virtualrain
wrongQuote:
overclocking is going to become an increasingly more complicated art with the release of Intel's P45 chipset
wrongQuote:
if you want to buy and overclock on an Intel P45 board (or X48, X38 and P35 for that matter), Leach believes that you must learn how to use:
* GTL Reference Voltages
* CPU VTT and its relation to GTLs
* Clock Skews
* CPU PLL Voltages
misleading, if you want to reach the max, then you you need a LOT of time to tweak the system, but hows that new?Quote:
It’s a case of spending a lot of time increasing the CPU VTT
misleading, you dont HAVE TOQuote:
Next you HAVE to play with the clock skews
wrongQuote:
The kicker that this is a completely blind art
misleading, you dont HAVE TOQuote:
you’ve quite literally got to sit there for hours and tweak the nuts off the board trying combinations of GTL and Skew settings until you find something that works
misleading, makes it sound like when i upgrade to a recent intel chipset i need to invest days into tweaking my system to get it stable above stock speedsQuote:
So there you have it, prepare to invest in some serious time if you’re upgrading
all those tweaks are helping to reach higher ocs, yes, but they only give very small boosts and arent necessary to overclock your system. this all makes it sound like its some extremely complex science or art to oc a system which will only scare people from buying a new board or from ocing it.
this is a 1 page article! and yet there are 3 wrong and 4 misleading statements in it...
i agree with whoever said this article is what youd expect on fudzilla, but not bit-tech
No, no, and no.Quote:
Originally Posted by kiwi
Once its tweaked, it will fly with every cpu at high FSB including wolfdales.
There are LOTS of hidden optionns in BIOS for every board, once manufacturer would put them into BIOS, everyone would be upset again. Some of the people are complaining about tweaking board at xs, this is where xs belongs, if you don't want to put all your effort into overclock, you definitely hit wrong forum.
http://www.anandtech.com/mb/showdoc.aspx?i=3309&p=8
Anand appears to have no major issues. He mentions raising the MCH voltage to 1.5 V ... but also notes that the test version was not final reisions either.
Proof? Show me any "tweaked" resultQuote:
Originally Posted by Kasparz
it wasnt anand, it was kris aka freecableguyQuote:
Originally Posted by JumpingJack
kiwi, pm pt1t :D
Majority of times leaving most options at Auto is enough, i personally have only needed to adjust (G)MCH voltage for FSB stability, the time consuming part is fidning the right voltage level to attain stability, throwing +0.20V may not be needed..spending some time testing could allow you to use less voltage. Rule of thumb...you try leave voltages at normal as much as possible while getting the best overclock possible.
While his 65nm chip doing 630fsb on x38 is nice it is still 65nm. Do it with 45nm and I will be sold :D In fact you can probably count a number of 600+ 45nm validation results on your left or right hand fingers :)Quote:
Originally Posted by saaya
Pt1t has done 600+ with 45nm chipQuote:
Originally Posted by kiwi
OK, let me just try to explain my point - my news post was NOT designed for the readers of XtremeSystems - you guys have the time to play and you know a system inside out. It was designed to introduce bit-tech readers, the average overclocker, to more advanced BIOS settings - that is all. Can you imagine the effect if I just had written "yea, these are new and you'll see them everywhere, they do stuff" - no one would have bothered to notice. I don't write with you guys in mind, I expect you lot to be more educated and read between the lines with what I (and other sites) put who aren't aiming their reviews at the extreme end of enthusiasts.Quote:
Originally Posted by saaya
Your statements are just wrong though - I have sat down with Taiwan engineers from Asus, MSI and Gigabyte who have all told me the exact same thing I posted: to get the most out of a system you need to learn the other options. 1.25V-1.3V through the P45 northbridge is all you need.
You have to understand my position as well - I've been thrown stuff from all angles harping on about all these things. I've sat through presentations explaining how to get the most of out these boards and they all revolve around the "new stuff" and not voltages, so it's hard not to think it.
MSI and Asus are specifically bringing down the X48 tweaks to P45 - everyone is committed to offering MORE, not less to play with. I will always be pushing for more too, but in a way that's elegant, intuitive and user friendly (for example, if the technical field was level: DFI are not, but Asus are).
It was just prior warning to my readers given the significant effort I had to put in trying to overclock the DFI X48 T2R recently. I will accept that the two things have influenced one another to a degree and the intended outcome of replies is not as I had hoped. I had expected people to be interested in a "wow, OK then" approach, not negative.
That's not to say these boards won't be overclockable from just drop in and play - they absolutely will - so those with Q6600s at 3Gs will still be able to do this easily enough. I meant it as just an extension to the simple terms of voltages that some already know. However if you spend the time playing with skews, vtts, plls and GTLs you won't need half as much voltage increase on CPU, NB or memory for the same effect - this gives your hardware a better lifespan (providing that is, you respect VTT and PLL voltages ;)). Whether BIOS engineers make automatic adjustments inside to aid all this, like some do (at the time of writing, having played with a few early P45 BIOS this didn't seem to be the case but it IS changing rapidly as launch approaches), has yet to be finalised :)
For those of you looking for OC or benchmark records having EVERYTHING at your disposal is a given. X48 will still be better for this, but despite Intel telling me A2+ revisions of X48 silicon will be 65nm at CeBIT at the 4-series "launch", it doesn't seem to be verified/the case at all.
The one thing I regret putting in is the Nehalem stuff, and only because Fudzilla posted something about it a day earlier so we ended up being associated with that. I was talking to the TWese and a few other industry people last weekend about it and they were echoing my suspicions about the two tier SKU approach Intel is taking (why else? You shouldn't need triple channel memory if it's more efficient?!). However, having just spoken to Asus this very afternoon - the PM of the RoG line has recently looked at the Intel stuff on paper and already "has ideas" so I think we'll be safe. As it is anyway, mainstream parts won't be around for another 9-12 months anyway.
I will keep pestering TW about Nehalem between now and December :) I'll resist posting anything about it until then, promise (can't let BT seem like a FUD site when it isnt :|) I'm gonna go back to work now so I regret to say I won't read the replies - just that ya'awl have a nice day and I'm sure I'll be reading your ocing stories soon ^_^:yepp:
EDIT: FYI: Asus P5Q BIOS looks :banana::banana::banana::banana:ing awesome. MSI have pulled its socks up too, but it still needs work :) I've got a meeting with DFI to see their new DK/LT too :D Feel free to drop me a line or forum post over there if there is anything specifically you want me to ask. I spose you lot don't really care too much about the greener-power saving features too much, do you?
I like a challenge, bring it on. :)
Brings me back to my DFI Ultra-D where I spent weeks finding my max clock on my TeamGroup Cronus Micron B5-F :D I say bring it :up:
As for overclocking on P35, I adjusted my GTL voltages to nearly every possible voltage combination and it didn't help me gain any fsb :down: Too bad the IP35 Pro does have these other options mentioned (DRAM skew, etc)
Anyone know any hex hacks? :D
For nOObs there's the CPU up level on the Asus boards, for real men there are bios settings... if we just know each settings is correctly explained in the booklet or what it affects that would be nice... overclockign indeed has evolved from upping some voltages and FSB to general tweaking to get best performance...
I go with El Donko : bring it on , she can't be that hard as 680i with a quad...
And really Gambit you gonna fight to breach the 3ghz barrier for lesser performance, beats me... forum threads will help ya well on the way, like usual... stay with the dark force...
Yea, he posted results as we discussed. Well, he is the 1st one and hopefully not the last one with more or less decent result for 45nm + X48Quote:
Originally Posted by dinos22