dos

5/22/2006 2:51:10 PM

^^ They make reference chipsets. And the NForce chipsets have far fewer issues than AMD's reference designs did back in the Athlon days (except arguably the 760mp)

5/22/2006 3:32:39 PM

I owned the old AMD chipset back before the KT266A era (may have been the 760).


I dont recall an amd produced chipset since then, which is said because it was a good board.

5/22/2006 3:40:06 PM

overclocking is lame

you can get equal performance boosts by doing the OPPOSITE - under-clocking - and keeping the system cooler.
this will cause fewer errors, increase stability....

OR YOU CAN JUST WAIT THE EXTRA 1.8 MILLISECONDS AND PUT UP WITH THE ONLY 37 FPSes

5/22/2006 3:48:29 PM

Quote :

"like the first P4's which were slower than their P3 counterparts"

Uhhh. Noobs.
Okay first of all, let's all not forget the first ruel of teh computar architecutre:

CLOCK SPEED MEANS NOTHING.

Someone find some benchmarks and post em up.
Real vs. Virtual registers,
Caching algorithms, cache SIZES, levels;
instruction-level parallelisms, branch prediction algorithms, on-chip multithreading....OOE, various accelerations for MPGs and common instructions....
it is TOTALLY STUPID TO COMPARE CLOCKSPEEDS NOOBS LOL

anyway. even saying this, intel still pwns up AMD in any industry-accepted benchmark.
The only valid argument for AMD is the cost-per-performance, which is largely an economic discussion, not a technical one.

5/22/2006 4:04:42 PM

http://amd.com/duel


oooo AMD beat Intel, in a test designed, operated, and judged BY AMD...

Sigh.

5/22/2006 4:07:38 PM

Wait wait wait. You are trying to make the "clock speed means nothing" argument in favor of Intel? The company that perpetuated the "clock speed = performance" myth for years after AMD moved to performance rated part numbers?

Who in the hell is even comparing clock speeds directly here? We are talking about PERCENTAGE overclocks, not raw numbers, which you should be able to recognize.

Quote :

"overclocking is lame you can get equal performance boosts by doing the OPPOSITE - under-clocking - and keeping the system cooler. this will cause fewer errors, increase stability.... OR YOU CAN JUST WAIT THE EXTRA 1.8 MILLISECONDS AND PUT UP WITH THE ONLY 37 FPSes "

You do realize that processor lines are all basically overclocked parts right? Do you even know what the fuck binning is? Do you have any idea how the whole industry operates? While I absolutely agree that it's retarded for the casual user, it definitely is not going to decrease stability or cause more errors if done properly and within reason. And gaining 15-20% speed increases is hardly something to scoff at for people on tight budgets.

Quote :

"anyway. even saying this, intel still pwns up AMD in any industry-accepted benchmark. The only valid argument for AMD is the cost-per-performance, which is largely an economic discussion, not a technical one."

Currently, AMD and Intel chips are dead heat even on nearly every benchmark out there, Intel skewed or AMD. You are a fucking idiot if you think ANY benchmark isn't optimized better for one chip or the other.

Quote :

"I owned the old AMD chipset back before the KT266A era (may have been the 760). I dont recall an amd produced chipset since then, which is said because it was a good board."

Yep they have always made good reference boards, but they couldn't slam all the features on em that people wanted and it diluted their business by trying to go into the motherboard and chipset markets, so they (relatively) quickly got out of it.

5/22/2006 4:27:08 PM

Quote :

"Do you even know what the fuck binning is? Do you have any idea how the whole industry operates? While I absolutely agree that it's retarded for the casual user, it definitely is not going to decrease stability or cause more errors if done properly and within reason."

Umm.. that's kind of a contradictory statement. Do YOU even know how processors are sorted? There's a reason why your processor is rated at its listed frequency. So don't talk about how binning is "overclocking" all processors and then make a false leap of logic to then claim that "proper, within reason" end-user overclocking is the same thing. Its not the same thing at all and its misleading and/or ignorant for you to conflate the two.

[Edited on May 22, 2006 at 4:58 PM. Reason : s]

5/22/2006 4:33:07 PM

Yes I am fully aware of the binning process.

I am also aware that as production runs continue, the masks are altered to be more efficient and increase yields while decreasing production errors. The result is that over the life of a processor (or gpu for that matter), the original speed binning changes, but the market price structure doesn't.

The result of which is that later production iterations that are binned for low performance are quite often done because there is not enough demand for their actual rating. So you as a consumer can quite possibly get a low speed chip that was actually rated fine at much higher speeds. This gets continually better (or worse depending on how you look at it) over time.

Binning is effectively "underclocking" for less capable units. But as production improves faster than demand, you get excess.

So yes, "overclocking" a chip that was artificially binned lower than it's actually production rating is safe and has absolutely no ill effect. Figuring out if you have such a processor is not easy, and is many times just a crapshoot, but it is possible and has been done for years.

5/22/2006 5:53:21 PM

Quote :

"The result of which is that later production iterations that are binned for low performance are quite often done because there is not enough demand for their actual rating"

bullshit. this may be true for embedded processors, but for a home PC, there will always be demand for faster processors - and if AMD/Intel can offer faster processors cheaper or at the same price then they will. Or else, why do they keep designing newer and faster PC processors?

Furthermore, "Figuring out if you have such a processor is not easy, and is many times all the time just a crapshoot" does not jive with, "if done properly and within reason"

The simple truth is this: even though you think you got away with overclocking your processor, and even though you think your processor was artificially labeled at a lower clock speed (WTF WHY AGAIN?), just because it is running stably at a higher than rated clock speed DOES NOT mean that you've gotten away with it. Without a doubt, you're significantly reducing your processor's lifespan.

[Edited on May 22, 2006 at 7:04 PM. Reason : s]

5/22/2006 7:00:08 PM

They do you douche. Did you not notice during the life cycle of the P3, P4 or Athlons, that as refinements were made in the masks, they scale the offerings, dropping the lowest end offering for new high end offerings.

It doesn't matter how good the yield is on any cpu line, they aren'ts going to drop prices on the high end units just because they can. Increased yield SAVES the company more money than it would MAKE them by restructuring their lineups. There have been times, on occasion, the Athlon XP 1800+ in particular, where because of such high yields and production, they were actually cheaper than the slightly lower product models.

You can say it's not true all you want, but all the numbers point to you being dead damn wrong about it. Every single major consumer processor line that has hit the market has gone through core revisions, and as they have, enthusiasts can push them farther and farther.

Quote :

"The simple truth is this: even though you think you got away with overclocking your processor, and even though you think your processor was artificially labeled at a lower clock speed (WTF WHY AGAIN?), just because it is running stably at a higher than rated clock speed DOES NOT mean that you've gotten away with it. Without a doubt, you're significantly reducing your processor's lifespan."

The simple truth is, the over lifespan for a desktop PC is under 5 years ANYWAY. A moderately overclocked processor will last WELL beyond this. How many stories do you hear about processors "burning out" from regular use? About 0. I agree after probably 10-15 years it will take its toll, but who the fuck even still uses a computer 5 years old, much less 10?

You are arguing retarded fucking semantics. Millions of people overclock their cpu's within reason (aka no core voltage changes, no super cooling, not pushing the envelope) and amazingly in nearly 20 years of the practice, it still works just fine.

5/22/2006 7:35:22 PM

don't come crying to me when your "safely overclocked" processor burns out

Quote :

"How many stories do you hear about processors "burning out" from regular use? "

overclocking is not "regular use".

Quote :

"How many stories do you hear about processors "burning out" from overclocking? "

A whole hell of a lot.

[Edited on May 22, 2006 at 7:42 PM. Reason : s]

5/22/2006 7:40:37 PM

Quote :

"Figuring out if you have such a processor is not easy, and is many times just a crapshoot, but it is possible and has been done for years."

Heh, it's actually not even that hard. These days, AMD even tells you exactly what week into the production cycle the core you purchased was fabbed on, right on the chip itself. Besides that, people have been posting which steppings overclock better for yeras now. Whether you actually get a good chip or not is a crapshoot, but once you have it, it's fairly easy to figure out if it's worth trying to overclock or not.

Bottom line, basically nothing Noen has said is incorrect. Those of you are arguing with him are essentially trying to dispute fact, and it just makes you look like ignorant fanboys

5/22/2006 8:31:22 PM

wtf am I a fanboy of?

5/22/2006 8:32:47 PM

Quote :

"A whole hell of a lot."

Please provide me with even a few testaments to this phenomenon. Not liquid cooling, not vapor, not liquid nitrogen shit.

And perhaps you could explain why there is a multi-billion dollar market surrounding these crazy unstable and prone to failure chips you speak of.

5/22/2006 11:33:05 PM

ive been running my 2.4C shit at 3.2 forever it seems with ram and processor voltages jacked up to a close max.

it is indestructible. anyway, all P4 have thermal diode which protects the processor from overheating.

every damn processor i had was overclocked and nothing ever happened. one is still running after 7 years of abuse. (celeron 667 mhz)

[Edited on May 23, 2006 at 12:44 AM. Reason : asdf]

5/23/2006 12:43:37 AM

why bother

Anything either you or I post, is just going to boil down to being anecdotal. We're just gonna have to agree to disagree on this. Luckily, for me, though, I have science on my side.

[Edited on May 23, 2006 at 7:54 AM. Reason : s]

5/23/2006 7:53:53 AM

No you see, had you actually just read my post,

I will accept even a few anecdotal cases as evidence here. Lets put the number at say 10. Find us ten independent cases where a person was modestly overclocking and that caused the CPU to be destroyed. Just ten. Since there are a "whole lot", this shouldn't be a problem.

Although, if you had science on your side, you wouldn't need to prove your point through anecdotal evidence, because you would actually be able to back it up with, say, science. But the reality is, the fault tolerances on processors are retardedly large and overclocking doesn't make anny pronounced difference in the longevity of a processor's actual lifecycle.

5/23/2006 9:10:12 AM

so you would claim that the failure rate on CPUs would not increase if they were all overclocked the way you describe?

5/23/2006 10:06:42 AM

Isn't physical failure not as much of an issue as increased instability?

5/23/2006 10:30:00 AM

If by "failure rate" you mean the physical, permanent inoperability of the chip, then no, the failure rate would not increase in any significant way by overclocking as I described.

Both AMD and Intel have working thermal throttling that prevent the CPU from being able to fry themselves out. If you aren't boosting voltages, there's very little the average consumer can do to cause permanent damage to a desktop processor.

Now if you are talking about failure rate in terms of instability during operation, absolutely it would increase if every chip was overclocked.

Here's another interesting tidbit to think about. What about the hundreds of thousands of P3's and Athlons that were remarked on grey markets and sold as higher speed chips? Very few consumers ever knew about it, because there was no big difference to the consumer. It mean't slightly cheaper prices.

However for Intel and AMD, they lost quite a pretty penny because these markets were buying budget chips and remarking them as higher speed parts, giving themselves the profit margin difference. So Intel and AMD started the age of multiplier locks. It had nothing to do with "failure rates" and everything to do with lost revenue.

5/23/2006 12:05:45 PM

Quote :

" Noen: I described...because Sun uses AMD Opterons...now... also give the...days of Sun...a big lumbering giant...is absolutely dominant...on all three fronts...but it's...in gobs. Plain and simply that is not...The Athlon XP. Dell, Gateway, IBM and so on used Intel processors...only...they are finally bringing the Core. ...Thank god. "

I believe by merging your previous posts into one meta-quote, I have summed up your stance succinctly?

I respond by reiterating that overclocking is lame and so is AMD.

5/23/2006 4:52:40 PM

[Edited on May 23, 2006 at 5:05 PM. Reason : ]

5/23/2006 5:05:00 PM

^^ how can anyone take you seriously when you make comments like

Quote :

"overclocking is lame and so is AMD"

That is quite opinionated and unsubstantiated.

[Edited on May 23, 2006 at 5:05 PM. Reason : -]

5/23/2006 5:05:33 PM

Quote :

"overclocking is lame and so is AMD"

what's lame is talking out of your ass

5/23/2006 5:25:37 PM

Quote :

"I believe by merging your previous posts into one meta-quote, I have summed up your stance succinctly? I respond by reiterating that overclocking is lame and so is AMD."

No, not at all actually. Sums up your idiocy and apparent excess of free time quite well though.

I respond that overclocking is what it is, calling a technical process lame is ignorant.

I fail to see how AMD is "lame" at all. You might PREFER Intel technology, but that in no way establishes a correlation of AMD being lame.

5/23/2006 5:38:39 PM

^ lol, a proponent of overclocking casts aspersions on someone else's "excess of free time"

[Edited on May 23, 2006 at 8:24 PM. Reason : s]

5/23/2006 8:24:18 PM

Did I say I personally do it? No. My time is more valuable than any small performance gain I might see. I did go through my tinkering phase back in 01-02, but I, like many others, moved on.

If it's a hobby, you can't argue with the legitimacy of it at all.

5/23/2006 8:43:53 PM

Hey Mr. Strawman, please come back with another one of your baseless retorts.

5/25/2006 9:21:44 AM

from what i was told, overclocking is just for nerdy people as a way to feel better about themselves

5/25/2006 11:31:48 AM

lol

5/25/2006 11:38:02 AM