The end of AMD Enthusiast-friendlyness?

[Bigjakkstaffa]

http://www.ocprices.com/?rev_id=179

--Jakk:t

[$1500-P4 gamer]

No. Its not. Buy a real mobo that auto unlocks multi's. Still can oclock just fine. Your NO enthusiest without one so to speak anyhow. Secondly MP chips are rated for duel use cause they pass certain signal erata tests at said mhz so they dont creat tons of eronous data when in MP. This happens all the time (eronous data) but more so when XP's are Mp moded. You want MP-pay for it. That simple. It will work better anyhow. And give amd the $ they deserve for making affordable servers to start with compaired to Xeons. Not to mention the advantage of hypertransport over Intels slow Mp system bus. But thats another story all together.:t

[MrBurns]

Most enthusiasts dont have multi CPU systems.

[$1500-P4 gamer]

"No. Its not. Buy a real mobo that auto unlocks multi's. Still can oclock just fine. Your NO enthusiest without one so to speak anyhow."

MrBurns, I was refering to the multi lock issue not the Mp one when speaking of "enthusiasts". Im sure most could care less about the MP issue too which isnt a problem anyhow. They are crying about the multi's not being able to be changed when used on older mobo's. Except kt400 which is newers but its Via-wadda expect! LOL. My point was "buy a real mobo then and dont ruin your cpu in the process". Maybe thats amd's point too. If your gonna oclock atleast keep up with the tech. Or dont do it. That simple.:D

[MrBurns]

I know, what you mean, and I know its about not being able to unlock the multiplier, but if you click on the link in the original post, and read this article, you canj read, that the new packaging that makes unlocking impossible is only used in a special dual CPU set.

[genesound]

LOL :D Welcome to Sysopt, MrBurns :t

[$1500-P4 gamer]

I did read the article thanks for the tip though.:rolleyes: Look again. ITs ALL THE L BRIDGES. Like the ones for the multi are there too-but under the same clear coating that the guy is complaining about. IF YOU read it you will see the comments at bottem of page about older mobo's needing the multi unlocked the traditional way and this not allowing it. Take your own advice before being so sarcastic please..Oh and I got all that info from NOT reading the article too LOL.

[genesound]

ROTFLMAO :D Don't ever change P4, Ill try not to either :D :D We have our premium mobos :D Let em fight over those VIAs :t

[morpheus kain]

It's not a huge deal since some mobo's auto-unlock, but I don't like the direction that this move is headed in. AMD don't be friggin fools cut it out and give the enthusiasts what they want cuz we're all that's backing you atm!

[genesound]

Naaahh, they're tryin to save a multi-billion dollar venture, we're (enthusiasts) a fly on the map. They need HUGE media and industry recognition to save their arses :eek: If they do , we'll get the boobie prize :t

[morpheus kain]

Boobie prize my ****! I'm in it for me not AMD and AMD isn't catering to enthusiasts with the new packaging, and they aren't fixing any problems either!

[genesound]

Well morpheus kain, I'm not disagreeing with you, I'm just sayin we're a pretty small target. They need to capture the BIG market, and hopefully not get us POd in the press, err um process :) I'm hopin they can figure it to do right by all of us :t

[mattyt69]

With this talk of mobos that will auto unlock a CPU will the ASUS A7N8X Deluxe auto unlock CPU's?

[Bigjakkstaffa]

As long as they are Tbreds or Barton's, yes

--Jakk:t

[Magua]

we're a pretty small target. They need to capture the BIG market

I don't see how changing the chip packaging would help do this. 95% of computer owners couldnt open up their case and point out the CPU. Right now Intel owns the retail market (do Dell/Gateway even commonly make AMD packages?) and AMD fills in gaps elsewhere.

What AMD needs to do is keep enthusiests happy (ie let them OC) while promoting there chips to the larger retail market. Changing the chip as the article states accomplishes neither.

[Someone Stupid]

What most don't seem to realize is that it only takes a few f'ups that are sent back to a reseller (which doesn't generally check chips) to have them complain to AMD. AMD can't **** off it's resellers, and the resellers don't want to lose those 10 out of 1000 chips in profit, as it still is profit that would have been made, but now it isn't as they had to send a new one. AMD is only saying, overclock with the right hardware, and we don't care. The enabling of dual processor usage also is a killer to them. If you really want multiprocessor usage, an MP is still cheaper than a Xeon. The laminate will most likely help them with sales rather than hurt them. Honestly, out of everyone you know that overclocks, or have read in forums from someone overclocking, how many times have you read of some fool trying to mess with something he or she has no understanding of? Not too often, but enough to where if you get the one's that admit it, you gotta remember 10 times that many won't most likely. I've seen people backtrack on what they've said countless times. This pisses resellers off having to deal with all the cost associated with inept overclockers. The idiot who posted the article, well I hope his reseller is reading that and knows him so he gets billed for two 2500's that he got as a replacement for cheating the system. If I blow something up overclocking, I don't RMA - if it's my fault for knowingly voiding my warranty I'm not going to commit fraud to save some cash and bury AMD deeper in the red. It's resellers AMD has to make happy, not overclockers. Overclockers are what are making the resellers unhappy as they are having to deal with overclocking ****ups and people who won't pay for their mistakes but hope it doesn't get caught on an RMA. Thumbs up to AMD in my book. You can't make money if nobody wants to sell your product.

[$1500-P4 gamer]

I see Someone Stupid and myself have similar opinions on this subject......;)

[genesound]

Yeah, good point ;)

[Magua]

Good point, Stupid.

[genesound]

:eek:

[morpheus kain]

Yeah I agree with ya stupid but I'm still kinda miffed.... Oh well... Can't win them all!

[Someone Stupid]

I was wondering if I'd get a positive or negative response from my post. :)

Long winded opinionated rant ahead warning:

I'm all for overclocking, hell, my comps overclocked and I'm practically waiting for my 2500 to die with all it's been through. The 8rda+ as well. Both have been tortured - the chip with 2.1 volts at one time, the 8rda+ with a vdd mod. If one goes out, well it goes out. To the trashcan it goes and I buy another one. I still come out ahead price wise from an Intel solution many times and nobody stuck a gun to my head and told me to overclock the hell out the poor first run barton. And if you can get a 2500, it really doesn't need overclocking at this moment anyhow, so you'd have time to save for a replacement beforehand - it's not like your buying an Intel which would take a good bit longer to save up for on a meager budget. The AMD, one game and an expansion pack less, the intel - well you get the idea. Intel locked it a long time ago, and they are priced considerably higher.

There comes a point to when even the small market segment we are does become a thorn in AMD's side do to some people in our community's inability to see that their actions due have an effect that isn't positive when RMAing stuff they broke as it adds up. I've even seen this even ENCOURAGED on some sites for chips, motherboards, and even video cards. Talk about shooting yourself in the foot. AMD doesn't eat the cost of the chip it sells to a reseller if the chip has been overclocked I'd imagine, why should it - they sold a working chip to the reseller. So naturally I'd imagine the reseller does have to eat that cost. AMD only has to reimburse them for chips which were defective to begin with. Though AMD has to spend extra money checking all these extra chips that shouldn't be coming in, I truly think it was more of a point to placate the resellers than to stop overclocking. If they wanted to do that, the nforce2 wouldn't be able to unlock multipliers to a limited degree, they wouldn't have let nvidia do that. That alone shows AMD isn't against overclockers - do you see any Intel solutions which do that? No, that's because Intel would have that chipset designer's head on a platter for doing such a thing.

Of course AMD could be cracking down, and we'll see no more automatic unlocking of chips. But if it happens, we have only ourselves to blame abusing the system. I mean this as a community for not discouraging RMA's from overclocked chips as of now it is generally not even frowned upon by many in the community sadly. A company that's running in the red can't afford to give handouts. That's like asking a beggar if he's got a dollar to spare. It just doesn't make sense.

--------

As for those who are upset, well we all should be. Not with AMD, but with ourselves, as if we were proactive with discouraging people to NOT RMA burnt chips and underperforming (to our expectation) motherboards and such, we wouldn't see such actions being taken. Looking back some of us - including myself when I was younger - have RMA'ed a burnt chip. None of us have been truly proactive in saying to someone highly vocally that burnt a chip up that we knew they were going to RMA that "hey, your only shooting yourself and the rest of us in the foot, just watch." I've known friends who've done this. I could have easily alerted the reseller of what really went down, but I didn't out of loyalty to my friends and out of ignorance. Well I'm seeing clearly now, and now the time has come to pay for our inaction on this as a community. We don't have any right to whine, we had a privledge and we abused it. Now we have to come up with new solutions if there are any to get complete unlockability or a dualie out of a single. I mean when you see overclocking and modding sites saying that it's okay to RMA a chip that one blows up or a motherboard that doesn't perform as well as one wants it, well this is the end result. AMD's hand was forced IMO.

--------

As for the Laminate - what is it exactly will determine whether it's "breakable." If it's some type of clearcoat, then some type of thinner should be able to be used locally to the bridges with some care - though if the bridges are bound to the laminate, be prepared to repair the bridges you need connected, if it isn't, wipe away and pop like usual. Or you can try sticking needles through the laminate to pop the bridges - just one problem - more current may be needed or AMD could have rewired the bridges when doing the laminate process all in one swoop to stop bridge play in it's tracks. If you want to find out, pay to play.

[genesound]

Hmmm... Automatic unlocked chips actually prolly contribute to the fraudulent rma problem. There's no way to just look at the chip and know what's happened.

I have a palomino I did bridgework on and nobody in their right mind would try to rma it, it's obviously been fiddled with.

Yeah, SS, this era of easy multiplier manipulation may be short lived. :rolleyes:

[MrBurns]

Originally posted by Someone Stupid
As for those who are upset, well we all should be. Not with AMD, but with ourselves, as if we were proactive with discouraging people to NOT RMA burnt chips and underperforming (to our expectation) motherboards and such, we wouldn't see such actions being taken.

There is nothing wrong about RMAing underperforming motherboards, because they can resell it to people, who are satisfied w/ the performance or dont overclock.


Originally posted by genesound
Hmmm... Automatic unlocked chips actually prolly contribute to the fraudulent rma problem. There's no way to just look at the chip and know what's happened.

You are right, that you cant find out by just looking at it, but if the CPU burnt down because it was overheated, and/or because of a sudden death syndrom, which are the two reasons, why overclocked CPUs burn down, you can find it out by technical tests, which are always done when you RMA something because of the warrantee. If your CPU is overheated, there is no warrantee, because its your fault, that you overclocked and/or didnt use the right cooling. Also Athlon CPUs dont have a overheating protection like Pentium4s have, so you cant blame the overheating on a not working protection.

[genesound]

Originally posted by MrBurns
You are right, that you cant find out by just looking at it, but ... That's just it, it means they're spending considerably more resources just to qualify rmas, it might be cheaper just to replace em blindly if there's no outward indication of malpractice.

[causticVapor]

This arguably can be met with a mixed reception, for good reason. There are very good and rather not-so-good sides to this.

First off, AMD is trying to erase the image of being a "cheap" CPU manufacturer. They unwantedly got that reputation through the first K6 generation that had an underperforming FPU, the Slot A Athlon that was nearly impossible to cool, and the Socket A CPUs that had no overheat protection.

It is also delivers an image of unprofessionality if a CPU can be changed to another model by simply closing or blowing a bridge, easily accessible on the package. People who realize that they have purchased remarked CPUs unawares are normally the loudest complainers, and CPU manufacters listen to them. They say, "Please, do something to stop this," and, ultimately, this leads to fewer options for the tinkerer, yet a better corporate image. In AMD's mind, thwarting remarkers ddes more for their bottom line than keeping a few enthusiasts happy, and I believe them.

Enthusiasts who did the L5 mod or wanted to use the bridges to perform a quick multiplier fix or voltage jump on a no-OC-option motherboard will be disappointed. With the exception of some very few odd cases, two XP CPUs of the same stepping and speed will work flawlessly together. All an MP designation gives you is the fact that you KNOW it's been certified for MP use and runs on the coolest stepping available... that's it.

If anything, this tells enthusiasts to get real MPs and real motherboards that can actually do multiplier unlocking by configuring registers instead of mechanically. However, there are those with older motherboards that can't do this, and would like to. In all honesty, anyone who has a KT333+Audigy won't see a performance difference and has no reason to upgrade to an nForce2 board. The Athlon doesn't gain much from a 400 FSB.

For all the people who need to mechanically set things, however, there is hope - the socket mod is completely reversible, and can do many of the things bridging L gaps do.

So far, things seem to be heading in the right direction. The A64 CPUs deliver incredible performance numbers, have some very marketable extra features, and wipe away the previous image of overheating and fragility. Initial pricing for the 3200+ A64 in Tokyo is already below the $800 suggested price - $657.

If Intel can't keep its prescott heat problems under control and sporadic death issues result, it could be an eventual, ironic turnaround for the chipmaker.

[causticVapor]

Here's my opinion of RMAing, what is right versus what is wrong, taken from the overclockers.com forums:

"Almost every time I've broken something, it's because I pushed it too far. I generally don't kill stuff foolhardily, but I know the limits. Two motherboards burnt up at the ATX connector because the PSU couldn't supply ample power to the 5v lines. Did I fix it? Yes. Would I have gotten an RMA? Probably, probably not.

I think it's OK to RMA to a manufacturer if you tell them that you broke it, and it's your fault. Then they can decide whether or whether not to give you a second chance with a warning or to refuse. I know this because I've heard the techs from one particular company, Epox, talk - their stance is that if you tell them you broke the product, they will honor the RMA, but if you lie and they find out it was most likely your fault, they won't.

That's honesty, and there's nothing wrong with that.

RMAing because you got an early stepping, however, IMHO is not right. Returning it and paying a restocking fee is an abuse of the whole return-a-product system. It is there for those who wanted a vid card and accidentally ordered a CPU, for those who wanted a 1600+ and got a 1.4GHz Tbird. Most of the stuff that I've owned hasn't been very overclockable - the AGOGA 1800+ palomino, A1-cored 9700 Pro, 1300 SL5ZJ Tualatin, and so on. They were all early steppings. They all clocked like ****. Did I return them? NO! I got what was advertised."

9-6-03

[Someone Stupid]

Mr. Burns, serious holes in your arguement. First, AMD cpu's do have thermal protection on all relatively new boards as well as thermal throttling. As for resellers checking everything. BS. I know of a couple people who got by both googlegear and newegg with cpus with blow bridges. They don't have the time to check them all - they mainly look to see if everything is there - namely if it was a retail box, the heatsink. They do pull some for checking, but not all. You'd see a whole lot longer RMA process than you have now if they checked everything, as well as a few bucks more to the price of certain items. Secondly, if you RMA something that performs to specs but not above them, that now means that companies which offer free shipping as a courtesy now have to do that shipping TWICE for free to sell the same board which was what you ordered to begin with. Also they cannot by LAW sell the board as new, they have to sell it as used, that means they aren't making much of anything off of it other than enough to hopefully cover the cost of replacing and shipping the board, and maybe getting a few dollars if a restocking fee was involved. So yes, if you get a bad stepping, well you should have done your research from where you bought it from to see what they were selling at the time. It isn't hard, most sites have feedback that isn't editted to say the least if one bothers to check. So if it doen't meet it's specs, then it's RMA'able no doubt, if it meets them, no matter how little it can't go past it, regardless of reputation of going past it, you can't legally do it but generally can get away with it - they technically could partially bill you for another board had they found out that it met specs - the law can become involved as well if you lied to get it through the RMA process. You technically got what you ordered. You can pay the restocking fee for a refund, but that's an abuse of itself which cV covered in his pull from overclockers.com and I happen to agree with (but that one is a moral opinion - not a legal one) - but the option is still there and that one at least is legal. It's hard to kill a cpu if you don't unlock it to do some extreme overclocking if your competent in what your doing. The only way is to burn it, and with the nforce2s with some putting up to 2.2 volts out, it is possible, but that can be checked, but generally it will slide past a reseller for sure, and even AMD can't fully check it as well as they would like. Sure there are a few pin sets which can reveal what they know are weak links when it comes to voltage and some when it comes to clock speeds I'd imagine, but since to really tell what happened would involved a physical autopsy of the chip, they aren't going to do it. So automatic unlocking does have it's downsides - but most RMA's are from bridgeplay, else they would have told nvidia no to the unlocking deal, no matter how much nvidia through at them cash wise as that to them - the people who build the chips don't see as the problem. Bridgeplay was the main problem.

I don't blame AMD for the laminate. They have to look after their bottom line and they have always been offering some of the best bang for the buck. I have a first stepping barton. Isn't the best overclocker unless you feed it enough voltage to make one cringe, but hey, I got what I ordered - I didn't RMA it a while later saying it died since it would likely slide past in hopes of getting the newer stepping as I missed it only by a couple weeks max as I built a system for a friend with a 2500 a few weeks later, and it was a better stepping and went up with very little core increase comparitively. Also have an R1.1 Epox that I had to vdd mod to go to 400 FSB stable - I didn't RMA it in hopes of a better one. I made do. I've had it happen to me in many other instances, wound up with a first run 4200ti a decent while after release - don't know why it was still in stock, but it was, tough luck on my part. It took a whole lot to get it to 4600 ti core speeds, but I didn't complain - I took it in stride.

cV: I share your sentiments - if they are the same as what you quoted :) on the restocking fee deal as well as one another reason as to why AMD had to go to the laminate - shoddy resellers playing with bridges and more than a handful of RMA's from bridgeplay by overclockers. I also know of several people who got by googlegear and newegg with blown bridges on their cpu. Resellers really can't check everything, so you have a fairly good shot of sliding on through long as everything is there and it's packaged up well.

[morpheus kain]

*nods* Yup Stupid I definitely agree. People should just take what they get and f'in live with it!

[MrBurns]

Originally posted by Someone Stupid
Mr. Burns, serious holes in your arguement. First, AMD cpu's do have thermal protection on all relatively new boards as well as thermal throttling.

Yes, but the protection is from the board, not from the CPU, so the warrantee of AMD is not for overheating. If your board has overheating protection, you have it enabled and then it fails and your CPU burns out, then you can RMA the board to mobo manufacturer and complain, that your CPU also doesnt work because of the failure in the overheating protection. If you can proof to them, that the CPU burned out because of overheating, and that you had the overheating protection enabled, the board manufacturer has to pay you the price of your CPU.

About RMAing CPUs because of bad stepping: I never buy anything online, so in my case, there are no shipping costs, so I never think of shipping costs. I think, that it isnt wrong to RMA something because it doesnt deliver the right performance if you buy it in a normal shop, because there are no shipping costs, and they still can sell it to someone else.

About checking parts: I once RMAed a floppy drive, because it didnt work. It was worth 13$, but the man in the shop checked it, and he said he did it, because they are not allowed to accept a RMA because of warrantee, when they dont check the parts. After half an hour the check was finished, and he said, that the little motor on the back of the drive, which should move the read/write heaqd was damaged. The reason for this is, that the resellers arent responsible for the warrantee, only the original vendors (like AMD) are responsible. So if googlegear doesnt check all RMAed parts, they probably break the warrantee contract w/ AMD. Also all checks have to be paid by AMD, because they are responsible for the warrantee.

[$1500-P4 gamer]

"Yes, but the protection is from the board, not from the CPU, so the warrantee of AMD is not for overheating. If your board has overheating protection, you have it enabled and then it fails and your CPU burns out, then you can RMA the board to mobo manufacturer and complain, that your CPU also doesnt work because of the failure in the overheating protection. If you can proof to them, that the CPU burned out because of overheating, and that you had the overheating protection enabled, the board manufacturer has to pay you the price of your CPU."

ROFLMAO- thats not how it is at ALL. Ever been to AMD site? Check their approved list. ALL mobo's on that list when used with XP CPU are guaranteed BY AMD to not overheat. The A7N8X is one-the 8rda+ another. Theres more. The diode is in the CPU. AMD lab tests approved mobo's at their lab and certifies they will shut off before damage under all conditions testable. READ the amd site once before making such accusations! So yep-board AND cpu manufacturer are guaranteeing it. They also tell you NO warranty on non approved mobos at all. See since the Toms H scare amd FIXED the issues but some just cant see the light and continue to spread Rumors. XP's from day 1 have had thermal diode implemented yet mobo co's didnt pay the $1 per board to enable it. (tons cheaper to make then intels Thermal unit) Thats the mobo co's slacking NOT amd at all. They have put their foot down now and its being enforced with the warrenty deal! Also no AMD experiences Sudden Death syndrome. Thats not even the name. Its Sudden NOTRTHWOOD (as in P4) death syndrome cause of too high of vcore they sometimes die early. This isn't happening with AMD XP's. Also P4 can burn-you'll see. Check the search engine EVER? LOL-temp isn't the big issue with them its volts. More than one way to fry a chip. And temp isn't the be all end all guaranteeing a sound O'Clock. It helps-but thermal protection wont stop a P4 from burning from a super high clocking. Guess you don't know about the gate issues and such powering these things.:t

[NDC]

As for dual configuration. AMD will still hold a good market hold for non-corporate SMP systems since Intel Xeons are much more pricey vs. AMD MP processors.

I run both, Dual Intel Xeon 2.4Ghz and Dual AMD MP2000+ and the AMD definetly gives you the bang for your dollars. However, I prefer my Xeon duallie over the AMD...

[MrBurns]

Originally posted by $1500-P4 gamer
Also no AMD experiences Sudden Death syndrome. Thats not even the name. Its Sudden NOTRTHWOOD (as in P4) death syndrome cause of too high of vcore they sometimes die early. This isn't happening with AMD XP's. Also P4 can burn-you'll see. Check the search engine EVER? LOL-temp isn't the big issue with them its volts. More than one way to fry a chip. And temp isn't the be all end all guaranteeing a sound O'Clock. It helps-but thermal protection wont stop a P4 from burning from a super high clocking. Guess you don't know about the gate issues and such powering these things.:t

No, there the sudden death syndrom also occured on AMD CPUs. It only occured on 0.13µ models, and because Northwood was the first 0.13µ core, it was originally named sudden Northwood death syndrome. If you dont believe me, go to www.overclockers.com, and search for the story "death of a thoroughbred" in the tipps and technics section. I cannot make you a link there, because the section doesnt work now, but maybe when you read this post, it works, and otherwise you can still look later.

Edit1: The tipps and technics section of www.overclockers.com works now. Here is the link to the story "death of a thoroughbred": http://www.overclockers.com/articles736/

Edit2: I never said, that P4 cant burn out, I just said, that they cant burn out only because of overheating.

[Someone Stupid]

:rolleyes: You cite ONE person's experience with it as a Syndrome, you know how many Northwood's die from tinkering with the voltage, a good bit for it to become famous enough that it gets the term syndrome. Voltage playing on it is much more brutal than with an AMD by far. Your citing one overclocker compared to the experiences of many, myself included with the 3.02B core. Honestly, your research should have more than one person crying as there are bad chips in any bunch.

EDIT: Read P4 gamer's post - that is all fact. So you really showed some real ineptitude on your part there with the lack of thermal protection on XP's.

[causticVapor]

I remember reading that article long ago.

There are always a few CPUs that will fail, and that guy probably got a CPU that would have died rather quickly, even at stock speeds. The truth of the matter is that most have been OC'ing for YEARS without problems. Again, there will always be an odd component here or there that is likely to fail.

Tbreds are harder to cool, because they dissipate more watts/mm^2 than Northwoods do. However, most models actually consume less power than the northwoods, because they have far fewer transistors. And earlier northwoods dissipate heat in an extremely uneven fashion, one that is prone to early death. Elevating the voltage only a little too much causes mad ripple from the overworked regulators. This creates surges that can eventually fry the CPU.

The northwood 3.06B is the first CPU to actually incorporate bidirectional CPU throttling - the CPU throttles itself down if the voltage regulators are overworked. This way, no damaging power surges are delivered to the CPU.

I bet you'll very rarely, if at all, hear of SNDS from a 3.06 or any of the 800MHZ FSB northwoods. As a matter of fact, the D stepping reduces the thermal gradient significantly, allowing for it to clock higher.

The Athlon doesn't have such problems, because the core is more rugged and distributes heat evenly. It sends a massive heat spike through the heatsink upon powerup, and then levels off (consumes 2-3 times more power than nominal to get going). Its 'hot spots' are far less pronounced than those on the northwood, partially because the ALU is running at the CPU clock speed, not double that. Thus, the core has more voltage and temperature resilience.

So what can we conclude? Early northwoods (<3.06) and Athlons on older motherboards are actually at equal risk. The latter will combust when cooling is taken away, the former under high load and overcurrent conditions. In reality, both the 3.06B/800FSB northwoods and Athlon XPs on thermally-protected motherboards are equally safe to use. :)