Separate names with a comma.
Discussion in 'Motherboards' started by FrgMstr, Nov 28, 2007.
They aren't real quad cores, they are just two dual cores.
There's certainly not anything wrong with going with the 680i if you're willing to put up with its idiosyncrasies (especially with quad-core overclocking thrown into the mix) and distinct possibilities for RMA's, but I was just trying to say that there are other options to run a perfectly good SLI setup, if compromising on either CPU performance (AMD) or SLI performance (650i) is worth greater stability to you, even with no budgetary constraints. Not everybody needs the absolute top-of-the-line regardless of stability issues. You implied that there was no other option.
I know that you do, hence why I said that few others probably do. Triple SLI is a whole 'nuther ball of wax; I look forward to seeing how it performs compared to traditional SLI, and if the 700-series manages to fix any of the 600-series' problems, or really is just unchanged with a PCI Express 2.0 bridge thrown into the mix.
Don't give me that crap.
True, I didn't word that correctly. In any case, there are only two high end options, 680i SLI or 680i LT.
I am too. I have high hopes that the 780i SLI chipset delivers. Like you, I'll have to wait and see.
Hate to troll, but I will never buy an nvidia chipset mobo again, as soon as the rlses calm down and qx9650 is reasonably in stock and x48 come out im crapping this 680i to the shooting range (would sell it but that wouldn't be a proper farewell to this crapset) ...
As far as the G80 is concerned: I bought 2 of these for sli 1 year ago this month, and they are still the best cards out, but I will not be making any more video card purchases unless ATi's crossfire solution produces better results than G80 SLi or Intel comes out with their own gfx solution that kills G80 SLi, otherwise I will just run 1 G80 in my x48 OC'ed like crazy and use the second as backup.
Anyway the bottom line is: nVidia will no longer receive any money from me, having pulled my hair through a few reference 680i's and striker extreme I think the bad taste from 680i in my mouth may last a decade if not a lifetime. I mean these guys had a year to fix the issues with this chipset, and despite evga's (whose support has been excellent) effort, though valiant, proved impotent in the end.
Just for comparison, and to echo Dan in part.
I've had no issues with my board. As in, zero. The machine is stable and fast. I'm running a dual core, so I've not had to deal with any of the quad issues. Yes, I'm unhappy that the board won't support 45nm quads (perhaps with a mod, who knows) but it's really a non-issue for me as I'm not planning to upgrade any time soon. Yes, the northbridge fan is fairly loud -- that's the extent of my complaints. I'll probably switch over to full watercooling in a month or two and it'll be a non-issue.
The problem so far as I see things appears to be more with quality control and the occasional idiot user ("I ran my RAM at 2.4-2.5 volts for months and now it's dead! It's all your stupid board's fault! Other boards make a special kind of high voltage that doesn't cause chip degradation, yours just sucks!") than design problems, at least for dual core owners. With the quads I understand the voltage regulation may be questionable. Not sure about that.
It's a good board. Not the best, but a good one. Mine has worked as advertised, and time and again things I attribute to hardware prove to be software related. Beats the hell out of many VIA based boards I've had to deal with in the past, for example.
It's been a great board for me too, but it sure as hell doesn't "Work as advertised". Yorkfield support was advertised for it, and that's the whole point of Kyle's post I believe, that it doesn't support it. No intentions of bashing anyone, just pointing out that the board won't do what it was advertised to do.
We still don't know who's fault that is. I'd personally say Nvidia because they didn't wait to have a retail chip in hand to test before announcing they would work, but that's been fairly common practice in the past so it's difficult to say. No one was claiming 45nm support when I made my purchase, so if you want to get technical it's done everything I expected it to do.
OK, so what circuit change ? What exactly is the deficiency in functional terms that is the problem. ? We (previous workplace) had to mod boards all the time to accommodate engineering "changes" until a new PCB could be laid out. It usually is not all that difficult and while not typically visually appealing, is something the telecom industry has been doing since PCB #1. Lets see how much "customer support" is really there.
Ask Nvidia to provide details of what has to be done, the usual this will void your warranty etc could apply but give us the details so if someone wanted to make the 680i 45nm quad capable they could.
It is realized this would be beyond the capability or skill of most users but I believe there would be a fair portion of people able to carry out the mod. Frankly I would be surprised if only "a", as in one, circuit was the issue.
People at XtremeSystems are already modding the power filtering and regulation side of things extensively in search of higher OCs on phase and peltier cooling, and they report much better results post-mod with the current quads. I doubt it's that straight-forward though, as you say.
LOLOL PCB#1 of the telecom industry didn't have 12-16 layer PCBs. They also didn't use signaling up around 1ghz and greater.
No! Yorkfield support was NEVER advertised as a feature of the 680i SLI chipset based boards. NVIDIA only claimed 1333MHz FSB CPU support. Guess what? The 680i SLI chipset based boards do in fact support 1333MHz FSB CPUs. There are plenty of Conroe and Kentsfield 1333MHz rated CPUs. People like yourself have drawn the conclusion that 1333MHz FSB CPU support means guaranteed 45nm, Yorkfield, Penryn and Wolfdale CPU support.
This discussion used to come up all the time when the 1333MHz FSB CPUs were coming out and some manufacturers added support for them to their P965 Express chipset based boards. I said it before, and I'll say it again, 1333MHz FSB CPU support doesn't equal 45nm, Penryn, Yorkfield, or Wolfdale support.
The point of Kyle's post was to clear up the issue concerning the rumors that 680i SLI boards didn't or did support these new CPUs. Kyle has now provided an official answer to this long debated topic.
I think what I've heard claim is that at one time, eVGA DID claim 45nm support with only a bios update. But once qualification testing was complete and they found out they could not support it on 680i boards, they removed that claim. And replaced it with the forum post that clarifies that Wolfsdale will be supported, but not Yorksfield. I didn't see the original claim first hand, just rumor, so take that with a grain of salt.
However; a quick google search, pulls up an Anand article from back in October where they did some leg work talking to board manufacturers.
Shows that while DFI, ECS, and Foxconn are still doing qualification. eVGA has finished, and states that only a bios update is required for T1/A1 revisions:
You don't get much clearer than that.
Similarly for ASUS, things have changed:
With no mention of specific board revisions. But as seen from the ASUS link tossed around. They have also backed down on their claim of supporting Yorksfield.
I understand that, but I don't think that 45nm support was promised in any official documentation. Unless you want to count forum posts as official. (BTW I don't.)
I'd love to actually try a Yorkfield processor on an EVGA 680i SLI, or a Striker Extreme and see what happens. I've got an extra of each, so anyone wishing to find out can PM me and send me a Yorkfield. I'll give it back I promise.
I didn't link forum posts. I linked an Anandtech article, where they claimed they talked to all these board manufacturers and are reporting for them.
Anand typically has a pretty decent reputation. When you, or Anand, or even tomshardware, or a half dozen other of the popular tech sites claim you've talked with companies and are reporting back directly. The community tends to believe you.
So do you think Anand reported falsly? Or do you think eVGA report to Anand falsly?
Honestly it doesn't really effect me at all, SLI doesn't interest me. So I have stuck intel/intel. It's an issue of integrity that has caught my attention.
I didn't read the posts. I wasn't questioning any other sites integrity, or accuracy. I thought we were talking about eVGA forum posts. I have seen plenty of information on the lack of 45nm CPU support for the 680i SLI, but nothing that seemed to be more than conjecture or second hand reports. So in this case my response to your post was based on an erroneous opinion. You mentioned forum post, and I had conjectured that the original information was in that form as well. My bad.
You know, now that I think about it, I had read this on the Anandtech article linked above; "EVGA is one of NVIDIA's premier launch partners and utilizes the NVIDIA reference platforms. This is an advantage the majority of time when it comes to quick BIOS releases and general support. The only open question currently is the GeForce 7050 motherboard. The 650i/680i boards will have optimized BIOS releases shortly that will fully update Wolfdale/Yorkfield performance on these products." and I think I didn't worry about it after that. It wasn't until recently that I started to wonder once the rumors in various forum threads were stating that Yorkfield and Wolfdale wouldn't be supported on the 680i SLI.
I'm pretty sure the box my striker came in had a big old "45nm support" sticker on it.
Def says 45nm something on newegg's photo's...http://www.newegg.com/Product/ShowI...r+Extreme+ATX+The+Ultimate+Gaming+Motherboard
Wow. I was just looking at my Striker Extreme's package and it says nothing of the kind. The 45nm sticker is absent. It says on the box that 1333MHz FSB supported when CPUs supporting 1333MHz FSB become available, but nothing else.
Early adopters such as you and I will not have that sticker. It was only put on a few months after the initial wave.
I would like to add that I'm not going to buy a 780i, I'm happy with my A8n32, I just played crysis at high settings with an overclocked 3800+ To 2.5 ghz.
I have 2 8800 GTXs in SLI and a Xeon X3220 with a 680i board.
My Xeon setup could not crack good FPS at even medium in vista, high was laughable.
here I am on my 939 RIG PLAYING CRYSIS IN HIGH ON XP.. WHAT GIVES?!
I'd say that your Xeon rig isn't equal to your Athlon machine.
ASUS has been claiming 45nm support from bios xxxx on upward, I guess they started slapping it onto the box after. I should check the striker boxes I have around, all of them might not say it, but the one I'm currently using does.
I has 3 huge stickers, 1333fsb support, 45nm support, quadcore support
eh? do you mean his xeon is not as good as his athlon =p? cuz his xeon 3220 would be a q6600 in xeon clothing ^^
He didn't mention the specifications of the complete machine. His Xeon 3220 machine may not have the same video cards, memory, etc. as the Athlon machine does. If that's the case it should be blatantly obvious why his Athlon runs Crysis better.
If the machines are equal, then something is wrong because the Xeon system should be slightly faster, all other things being equal.
Mine has the Quad-Core ready, and 1333MHz FSB support stickers. Mine is one of the earliest retail boxed Striker's though. I'm curious, what PCB revision is yours? Mine is version 1.00G.
Thank you Menelmarar!! Thank you for saving me from digging through my garage, it's a hell-hole of storage, and it would have taken forever to find that box with the 45nm support sticker on it.
So Dan after seeing this maybe your answer would have been,"Yes! Yorkfield support WAS advertised as a feature of the 680i SLI chipset based boards"
It seems that every time I find proof that this was said somebody reads into it and figures out a way to manipulate it. The Yorkfield is a 45nm chip, and it says 45nm is supported. I didn't draw any conclusions as you stated, I simply read the advertising, perhaps you drew a conclusion about me.
I guess the next thing someone will say is,"the dual-core is 45nm, so they're not lying".
LOL, what's next!? EVGA has already officially stated that once the 780 boards are available they will make a determination on whether or not to provide a step-up program.
That's EVGA's official response for the moment, so only time will tell if their customer service lives up to the hype. (and yes, I know it's not their fault, but it's their responsibility)
It would seem that you are the one reading into things and drawing conclusions of your own about what the advertisement implied. The fact is that 45nm CPU support does NOT mean that the Yorkfield chips are supported. This conclusion of mine is bolstered by two facts. One, Yorkfield isn't the only 45nm CPU Intel has. Second is that since dual core 45nm CPUs are being supported, the 45nm support sticker is obviously not a lie. You read "45nm CPU support" and drew the conclusion that Yorkfield was included.
It doesn't really matter if I say that or not, but I did. What will matter is that the lawyers would say that if someone tried to make a case out of it. Not to mention that there is always the fine print on the package that says "specifications subject to change without notice". Such text is placed there by the legal departments of various companies to absolve them of guilt or responsibility.
The problem is that the end users typically interpret something like a 45nm support sticker into a guarantee for Yorkfield compatibility which is obviously a stretch.
Who knows? They already stepped up to the plate and gave us the A1/T1 versions to fix the quad core overclocking issue when legally they were not bound to do so. They took care of their customers. This equals good PR, but the fact is that like any other company they are in this business to make money. I don't expect that EVGA or NVIDIA will be willing to basically replace nearly every 680i SLI board they've ever sold with a brand spanking new 780i SLI board. I just don't see it happening.
It is not their fault, and I fail to see how it is their responsibility. A forum post probably wouldn't be binding enough in the eyes of the law. That whole subject to change without notice being printed on the packages would surely come up if anyone tried to make a case out of it.
For the record, I think that the end users need to be aware that there are no guarantees for future CPU compatibility when those future CPUs are still in the design and validation phases of development. Furthermore, history has shown that such promises of forward compatibility are often met with dissappointment. I also believe that manufacturers making such statements should avoid making them. Obviously people will take offense to it if it turns out that such claims are invalid. That means legal action, and a damaged reputation. ASUS really shouldn't have bothered sticking 45nm CPU support stickers on their board packages. Obviously they did this due to sell more boards to people who might have otherwise avoided those boards and purchased something different, or waited for the release of 45nm CPUs before investing in the 680i SLI platform. That sort of thing is fine if they can validate these claims in advance. Obviously, such validations were probably not made in this case prior to them tagging the boxes.
Current one in PC, that came with the 45nm sticker, is 1.00G and shipped with the 901 BIOS, which is a BIOS that is not listed to support 45nm
So really I don't know what the hell ASUS was thinking.
At best it's misleading since most are going to assume they meant quadcore+45nm+1333fsb and we know that's not going to happen. At worse they pulled a fast one and only meant dualcore, and then shipped the sucker with a BIOS that won't work with it.
Well any board purchased before a specific CPU was released is going to ship with a BIOS that may not work well enough to POST the system, and then update the BIOS with the new CPU installed. This is understandably unavoidable.
They were thinking about making money. ASUS has always had a good relationship with Intel, so if any 680i SLI board would support 45nm Yorkfield and Wolfdale CPUs, it should be the Striker Extreme and the P5N32-E SLI.
I agree that it is misleading as many will draw that same conclusion. They should have realized this. EVGA isn't guilty of that. They only advertised 1333MHz FSB CPU support on their packages. There are plenty of 1333MHz Conroe and Kentsfield CPUs, so they've done thier job at least meeting that criteria. I can't speak to anything else though. Anandtech's article mentions validation, and as for forum threads; while some might find posts from various EVGA or NVIDIA employees official statements you really have to maintain some skepticism when reading those. All those posts aren't likely from engineers or are they cleared by the legal department.
A quick personal experience I thought I would share with you guys.
After reading Dan_D's response to my post, I thought he was full of it, so I gave my lawyer friend a call. We went to high school together and I knew he wouldn't lie to me. (Is that possible for a lawyer?). So I sent him the link and he read the post and looked at the newegg picture. "Supports 45nm CPU", as best as we could read the fuzzy picture.
After about 3-4 min's he said,"How much does this 45nm CPU weigh?" I was like, wtf are you talking about??? He said,"If this ""45nm CPU""can physically sit comfortably on your motherboard without causing damage, then it supports it, anymore questions?"
Well, needless to say, after that it went down hill. I ranted for a while and he proceeded to explain a multitude of cases where things just as ridiculous as this one held up in court.
So I guess Dan and the people that share in his manner of thinking, are in fact, not full of it at all, I am.
I've re-thought my expectations, and I guess now I can hope that perhaps EVGA will offer a discount if I upgrade to a motherboard that will:
Support the Yorkfield 45nm quadcore processor from intel, electronically functional and overclockable just like past central processing units have worked in other motherboards. (Try putting that on a sticker, hah!)
That is AWESOME!
I think in very literal terms. So even though I'm not a lawyer and never would want to be, I have to admit that sometimes we speak the same language. Personally I think it is bullshit that manufacturers put cryptic shit like that on their boxes, but there is nothing I can do about it.
Regardless, you and I, and for that matter most here, know how to update a BIOS. Most of the public, even most DIY builders have not done this. To ship something claiming it works, when it flat out will not, is bad business when some poor sucker tries to slap it and nothing happens.
To make matters worse strikers are known for failing flash updates left and right. So when said user tries to use ASUS update prepare for the worse.
I'm sure some people will slap their new CPU in, get frustrated and call ASUS, get told to use ASUS update, and brick the board.
Fair enough. The striker is also not a reference board. I've seen screen shots of people with the retail chips working... but they don't work at their rated speed and can't be OC'd.
Maybe there was some method to ASUS madness behind the striker design and it will show it's teeth on 45nm, that would redeem it in my eyes.
Agreed and I don't. I'm speaking as to what was plastered on the box.
Looking at their site though they make it quite clear that 45nm quadcore is not on 680i (though I don't know if they removed that check recently).
ROFL and from that site it seems ASUS is stating that 780i is quadcore 45nm no go as well.
That'd be funny if it was true :/
What are the specs of both machines? Obviously if the machines are comparable, the Xeon machine should assrape the Athlon machine, at least in benchmarks. As for actual performance, you shouldn't notice a big difference especially not at stock speeds. Though the Xeon 3220 should absolutely kick ass overclocked. In regard to Crysis, this should make some noticable difference.
Oh I believe you, I just believe that the 6 680i's I've gone through have made me shudder. Almost all of the mem controllers on them give out and go into bitch mode (6400 only).
Right now I've got an X2 3800+ overclocked to 2.5 ghz per core on air (Zalman 8700 NT cooled)
A8N32-SLI Deluxe board.
2 8800 GTX's
1 X-Fi fatality
Enermax Galaxy 850w Power supply (Great PSU, too damn big )
Basically the only thing that changed was the board and CPU (Got my xeon x3220 and 680i sitting somewhere, I am due for a new 680i but my store told me eVGA stopped spinning 680i silicon and thus there's a halt of supply in the market and they'll be due in next month so I've got to wait until then)
I'm just so surpised that I could play crysis in the Very High hack mode of XP at very playable framerates. I'm running 2560x1600.
On vista, dx9 or 10 it didn't matter. It would make my ram passout and fry my board prematurely, or any intensive game that uses lots of ram (WoW was another HUGE board killer for me).
Mushkin has been very lame handling my RAM situation. They are very steadfast on the belief that the Mushkin XP2 8500 996535 Ram (2x1GB) works just dandy on quad cores with their memory and 680is without issue. I've asked them so nicely if they could give me a 2x2GB XP2-6400 set of ram as a replacement since they're equal cost and they flatly told me no that if I was to need a replacement it would be the same set.
I know this is industry standard but dealing with Corsair (Who understood my 8500C5D woes that I had before the mushkin) they realized that when I bought the TWIN2X-8500C5's last year for nearly 500 dollars that this was unacceptable since my second set did the same and gave me a FULL refund of what I PAID (Not market value!) and offered to give that as credit as well in case I wanted different ram.
Nonetheless, this whole situation has those 2 items of mine tucked away (and most likely sold) because I feel it's just not going to get fixed soon. I also want skulltrail and I don't know the ETA on that and this 939 is literally holding it's own so far so I'm pretty much stuck. I am told skulltrail is due late Q1 I'm hoping for a Jan/Feb release date since it's going to be close between it and the 780i (for me atleast).
Any help is appreciated, Thanks!
I've been following this thread with interest and I have to say that your post was VERY refreshing. It is so nice to hear someone show intelligence and investigate the other guys opinion instead of bashing it. Thanks for your insight and thanks for checking on this. I shared Dan's opinion, but didn't have any facts to back it and its nice to get corroboration from someone in the legal field.
That's weird that Mushkin is treating you that way. I have bought exclusively their memory since 1999 and had very good experiences with tech support and the RMA process. At one point I had asked about the feasibility of changing from what I bought to another type and paying the difference, and they were willing to do it. (This was in the last year.)
Have you tried posting in the support forums at all? Some of the moderators their are helpful at putting you in contact with Mushkin personnel. Maybe someone there can get you a better answer on the memory trade.
Thats very interesting,as this is third or fourth time I have heard that the 780i series wont support 45nm quads.And here it is again in print,on Asus's own website.
They've stated 100 times they're not going to exchange it for a different set no matter how much I plead.