Intel invents a way to prevent running CPU out of spec **updated**

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

zShowtimez

Senior member
Nov 20, 2001
544
0
76
I guess I should care.. But I dont. I have my new rig for the coming year or so. Ill just see whats happening then, and be pissed about it. :confused:
 

First

Lifer
Jun 3, 2002
10,518
271
136
BFG10K and Budman, remarking is an issue with pre-built systems (from SI's, OEM's, etc.), not an enthusiast issue. Your average Joe doesn't know what WCPUID is or how to find the clock speed of their processor by right clicking My Computer and choosing Properties.
 

ProviaFan

Lifer
Mar 17, 2001
14,993
1
0
The "CPUs are fast enough already" people just need to STFU. I overclock (it's an OCD), and wouldn't consider an Intel chip if they did this (if AMD did it too, then that levels the playing field, and I'd consider either based on which offered better price vs. performance). However, even though I don't like it or agree with it, I can still see their motivations for developing this technology.
 

sharkeeper

Lifer
Jan 13, 2001
10,886
2
0
LOL! If you think today's CPUs are "fast enough," you need to do some 3D modelling/rendering/animation. About 36 hours into a render session, you won't be thinking your CPU is "fast enough" any more And as for out-of-spec PCI, there are several popular platforms now which can do away with that issue entirely. nForce2 is a prime example of that.

I'm speaking of the average user. The end user that *needs* the cpu power wouldn't consider running *ANYTHING* out of spec EVER! What if the CPU was producing errant data? This is exactly what I'm talking about! I'll leave the ABMAX capabilities up to the people that know what they're doing and trust my outputs. Helps to get a good night's sleep which is rather rare in this department as of late.

As far as the nforce issue, it's possible, but I'm not going to explore the path with rogue chipsets. We prefer to stay on the AMD/AMD or Intel/Intel road and I'll leave it at that. :)

Cheers!
 

bluemax

Diamond Member
Apr 28, 2000
7,182
0
0
Lots of good points for and against... interesting read.

I wouldn't mind seeing them locked, provided the prices stay competitive and low, or improve. :D

You think Intel would charge less for a CPU knowing it WILL NOT come back burnt by overclockers? Less RMA costs could be passed on to the consumers.... if they care. ;)

Mind you... Overclockers are maybe 1% of Intel's market... of those, ~10-25% may burn out their chips. That's not a huge savings in RMA's.... but if you multiply by a hundred thousand CPU's in a year or two, it does add up.
100,000 CPU's * $200 = $20 Million.

I'd like to be the guy at Intel who saved the company ten or twenty million dollars. Can you say, "Bonus, baby!" :D
 

Adul

Elite Member
Oct 9, 1999
32,999
44
91
danny.tangtam.com
Originally posted by: shuttleteam
LOL! If you think today's CPUs are "fast enough," you need to do some 3D modelling/rendering/animation. About 36 hours into a render session, you won't be thinking your CPU is "fast enough" any more And as for out-of-spec PCI, there are several popular platforms now which can do away with that issue entirely. nForce2 is a prime example of that.

I'm speaking of the average user. The end user that *needs* the cpu power wouldn't consider running *ANYTHING* out of spec EVER! What if the CPU was producing errant data? This is exactly what I'm talking about! I'll leave the ABMAX capabilities up to the people that know what they're doing and trust my outputs. Helps to get a good night's sleep which is rather rare in this department as of late.

As far as the nforce issue, it's possible, but I'm not going to explore the path with rogue chipsets. We prefer to stay on the AMD/AMD or Intel/Intel road and I'll leave it at that. :)

Cheers!

its been done on the intel platform has well. And the nforce is the best chipset for the AMD platform IMHO.

BTW when i was working at inel, it was common to see higher speed processors clocked down a bin or two to meet market demand.

 

mechBgon

Super Moderator<br>Elite Member
Oct 31, 1999
30,699
1
0
Originally posted by: shuttleteam
As far as the nforce issue, it's possible, but I'm not going to explore the path with rogue chipsets. We prefer to stay on the AMD/AMD or Intel/Intel road and I'll leave it at that. :)

Cheers!
I see now. You're not going to OC your production systems at work (heck, I don't either), and therefore you don't want anyone else to have this option either.

Guess what? I don't drive my car above the speed limit on public roadways. Perhaps I should come over to your house and install a governor on your car that limits it to precisely the speed limit...? :D Ah so, grasshopper... :)


edit: as Adul points out, some of Intel's own chipsets can lock the PCI and AGP bus too. And with a fair number of nForce systems in my fleet, I can vouch for them being a fantastic platform, very stable and trouble-free lil' guys :)
 

sharkeeper

Lifer
Jan 13, 2001
10,886
2
0
I see now. You're not going to OC your production systems at work (heck, I don't either), and therefore you don't want anyone else to have this option either.

Guess what? I don't drive my car above the speed limit on public roadways. Perhaps I should come over to your house and install a governor on your car that limits it to precisely the speed limit...? Ah so, grasshopper...

Bad analogy!

Don't laugh, but that may be coming as well! It's only a matter of time...

Not that I would agree to THAT...

Cheers!
 

FishTankX

Platinum Member
Oct 6, 2001
2,738
0
0
Originally posted by: bluemax
Lots of good points for and against... interesting read.

I wouldn't mind seeing them locked, provided the prices stay competitive and low, or improve. :D

You think Intel would charge less for a CPU knowing it WILL NOT come back burnt by overclockers? Less RMA costs could be passed on to the consumers.... if they care. ;)

Mind you... Overclockers are maybe 1% of Intel's market... of those, ~10-25% may burn out their chips. That's not a huge savings in RMA's.... but if you multiply by a hundred thousand CPU's in a year or two, it does add up.
100,000 CPU's * $200 = $20 Million.

I'd like to be the guy at Intel who saved the company ten or twenty million dollars. Can you say, "Bonus, baby!" :D

The P4 costs about 30$ to produce, thus if demand isn't at max I doubt you'd be saving that much.
 

ProviaFan

Lifer
Mar 17, 2001
14,993
1
0
Originally posted by: shuttleteam
Bad analogy!

Don't laugh, but that may be coming as well! It's only a matter of time...

Not that I would agree to THAT...

Cheers!
Wrong. Good analogy. You might happen to have a car hobby, while Mech has a computer hobby. Pretend that Mech has a job working on cars, or working in law enforcement; he would probably love to see the speed limit enforcer (note that you, the enthusiast would be greatly distraught to see such a thing mandated).

If I could remember it, I would insert the "they came for the [some group], but I wasn't one, so I didn't care; then they came for the [another group], but I wasn't one, so I didn't care..." (greatly paraphrased, of course) quote here. The double standards and hypocrisy that some people show continues to amaze me. :disgust:
 

mechBgon

Super Moderator<br>Elite Member
Oct 31, 1999
30,699
1
0
I also doubt $20 million is of much consequence to a company like Intel.

Really, OC'ing is generally about one of two things: taking a gamble on getting a cutting-edge CPU at a low price, at the expense of one's warranty, or else getting a CPU that no one else has yet (3.5GHz Pentium4, 2500MHz AthlonXP, 3GHz Celeron-478, etc). It's a modern-day sport, and I see no harm in enthusiasts having their bit of fun this way, any more than they do by putting different cams in their engine, dualling off their exhaust system, etc. :)

edit: thanks for the word of support Josh :) I actually have a bicycle hobby, and yes, some days I would like to put limiters on certain folks' cars :)
 

Cerb

Elite Member
Aug 26, 2000
17,484
33
86
IMO, they should have the chip know it's speed. If it is more than X% out of it, it should have a warning in big red letters when the BIOS comes up telling what it should be and what it is.
We wouldn't care.
Joe "I use it for email" Smith should.
 

Wingznut

Elite Member
Dec 28, 1999
16,968
2
0
Originally posted by: Adul
Intel invents a way to prevent running CPU out of spec
What makes you guys think that they couldn't make it impossible to overclock RIGHT NOW? If Intel or AMD wanted to make their cpu's un-overclockable, they could already do it.
 

sharkeeper

Lifer
Jan 13, 2001
10,886
2
0
Well with any discussion you have your partisans and antagonists. :) I guess I should've known better to even speak out in such a debatable topic.

Next week, they'll be an article about broadband ISP's programming Packeteer to choke P2P throughput to oblivion.

I'll avoid that one like the plague for sure.

Cheer up folks, it's just a computer! :p

Cheers!
 

flxnimprtmscl

Diamond Member
Jan 30, 2003
7,962
2
0
Originally posted by: Jhhnn

Overclockers cost them money in other ways, too- after smoking their chip with too high a voltage, they return it under warranty, demanding a new chip. Overclockers are an extremely small segment of computer sales, and undoubtedly more trouble than they're worth.

That was the exact thing I thought until I considered it for half a second. Look at it this way. Probably somewhere around 40% or so of all cpu's sold to enthusiasts would have to be rma'd for intel to break even (retail value wise) on enthusiast sales. You there's anywhere close to that number of rma's? I don't for a second. And even if there was and intel was only breaking even on enthusiast sales (a very very small percentage of their sales) would they even really care? So whatever might cause this to be implemented I doubt it would be the enthusiast market.

 

spanky

Lifer
Jun 19, 2001
25,716
4
81
Originally posted by: Evan Lieb
BFG10K and Budman, remarking is an issue with pre-built systems (from SI's, OEM's, etc.), not an enthusiast issue. Your average Joe doesn't know what WCPUID is or how to find the clock speed of their processor by right clicking My Computer and choosing Properties.

thats what i was gonna say. also... if intel implements this... there is always the chance that amd will follow suit and license it from intel.
 

Junkman

Member
Nov 27, 2002
44
0
0
Sorry to ask a few dumb questions... I've done some mild overclocking in the past just as a learning exercise mainly and I am far from being an expert at it but here goes..... The anti overclocking they are talking about does it apply to the processors or the chipset or both? Also does Intel examine processors that are RMA'D to determine the cause of failure...its my understanding that if you burn up a processor by overclocking that its at your own risk and have violated your waranty...or is there no way for them to determine this?
 

Goose77

Senior member
Aug 25, 2000
446
0
0
I believe that even if INTEL does go the route to lock it down, there will be some doofy techie that will spend all of his free time trying to figure out how to unlock it. In the end, he will get by it, word will spread through the net like wildfire, and we all are off and OverClocking again..

 

AdmiralTiger

Member
Feb 17, 2003
119
0
0
Originally posted by: McCarthy
Will that prevent underclocking too? :(

One can only hope! No Underclocking, no overclocking.. talk about a pretty straightforward CPU... sounds like the Sims working in an environment with no recreation nearby, just straight to work, then home, sleep.... then wake up, go to work. Boring.

Obtaining a patent is one thing.....

Announcing they'll use the patent - that's when we ought to whine about it. For now, we have nothing to worry about (for me anyway, I'm an AMD fan.) :D
 

AdmiralTiger

Member
Feb 17, 2003
119
0
0
Originally posted by: Junkman
Sorry to ask a few dumb questions... I've done some mild overclocking in the past just as a learning exercise mainly and I am far from being an expert at it but here goes..... The anti overclocking they are talking about does it apply to the processors or the chipset or both? Also does Intel examine processors that are RMA'D to determine the cause of failure...its my understanding that if you burn up a processor by overclocking that its at your own risk and have violated your waranty...or is there no way for them to determine this?

You're right... overclocking voids your warranty and you have to buy a new CPU yourself, no RMA's if you ever kill your CPU. They can determine what went wrong with the CPU (obviously by the burn marks in the die itself or something like that.. but I'm just speculating.)
 

techfuzz

Diamond Member
Feb 11, 2001
3,107
0
76
Just because a company applies and receives a patent doesn't mean that it will be ever used commercially. Someone patented a dimple-making device in the 1800's here in the US, but I doubt anyone ever went about mass producing it!

techfuzz
 

Robor

Elite Member
Oct 9, 1999
16,979
0
76
Originally posted by: shuttleteam
About time they do this.

Hopefully AMD will follow suit!

Cheers!
Why celebrate an idea like this? Intel charges way to much for their premium processors. If I want to buy a cheaper 2.4 CPU and run it at the same speed as an expensive 3.0 what's the harm? Yeah, I know the answer. CPU damage, corrupt data, etc. I don't care. I know how to monitor the voltage and temp of my CPU and it's my data so if I corrupt it pushing the system to far it's me who fixes it. I do see the arguement about preventing fraud but there's got to be a better way than this. I've been Intel/Intel for a while but this idea would push me to AMD.