Slashdot Mirror
Tampered Athlons Hit Oz
"This is some very bad news -- thanks to Tim for alerting us. Rather than paraphrase, I will just quote his email:
'Well it looks like they [fake Athlons] have hit Australia I just recieved my K700 from [an undisclosed source] in perth and have opened it up to put the cold plate on it and to my horror the cpu is a 650 ... the resistor has been changed and serial numbers do not match ...'
We managed to get some pictures of the Athlon in question, and all the pictures are posted, including a summary with each picture, and help for those interested in finding out if their CPU has been tampered with.
More details are available on our frontpage.
We are working on finding an easy way (ie, for 'general' consumption) for those possibly affected to test their processor, and any updates will be posted, again, on our main page :)
Kind Regards,
Lucien Wells.
---
Lucien Wells
Editor/Assistant Reviewer & HTML Developer,
TechWatch"
The well-labeled pictures also serve as a primer to understanding the cryptic labels on the side of your processor. But as Lucien points out, checking this out will void your warranty.
CPU Identifier identifies the CPU you are running (on x86 systems only). This can be useful in three ways: 1) identifying remarked CPUs, 2) getting detailed information about your processor, and 3) when combined with the info given by /proc/cpuinfo.
The Athlon 800 uses 43 Watts under typical use and 48 Watts maximum. I'd love to know where you're getting the information that PIII-800s use only 17.2 Watts? I can't find a datasheet on the 800, but the 600 seems to use 37 Watts under typical use, according to Intel datasheets.
----------------------------
Actually it's more like 1.65x the wattage. not 2x.
----------------------------
Fry's Electronics does the same thing, but (if you're lucky) they will put a little blue sticker on the box. If you ever shop at Frys, *never* buy something that has a return sticker, or has an ameteur shrinkwrap job.
Actinide is a series on the periodic table (elements 89-103). The word I believe Sig11 is looking for is asinine, meaning stupid or silly.
You will be picking up your non-legit hardware over at Best Buy or CompUSA in the future (if you haven't already).
I don't know about CompUSA, but definatly at Beast Buy. Considering that defective returns end up back on the shelf, they can't be paying enough attention to counterfeits.
I am sorry, this is both offtopic and blaming AMD products, but this might affect my purchases.
I have been told not to base a server on Athlon because of power consumption/heat/stability problems. How is it, can it claim 100% uptime for a few years (an intel-level server lifespan)?
I had hoped to base the server on Athlon hardware but have been convinced otherwise
I think, therefore thoughts exist. Ego is just an impression.
Bogomips = Bogus MIPs
Because people don't look for fake $1 bills as often as they do fake $20s
If you get a remarked CPU, you can discover that with the appropriate software tools. That should be sufficient, although opening it up to further show the fraud might be more convincing. If that "Mom and Pop" place still refuses a return on the clear case of fraud, then they should be run totally out of business.
... e.g. get their own money back on the junk).
I don't have a problem with retailers that refuse to take back DEAD stuff because I understand how this stuff can become DEAD. OTOH, I've gotten DEAD stuff from 2 local dealers and they both took it back (motherboard in one case and RAM in another). I'm not even certain that I didn't zap them myself. I was glad they did swap out for me, but they could have just as easily told me I must have static zapped it.
OTOH, if I get a clear case of fraud, I will EXPECT them to take it back (and to even follow up with their distribution source
now we need to go OSS in diesel cars
The rationale for catridges was not security. Slot I catridges were introduced for the Pentium II, which had their L2 caches off die, because of the inability to keep the cache on die while keeping the costs low. Wrapping a catridge around the circuit board was simply for aesthetics and durability, so less people will break it by accident (similar to why console games have plastic casings wrapping their games). It was never designed to be a tamper-proof mechanism.
Go get your free Palm V (25 referrals needed only!)
Reminds me of a recent situation with Apple.. some of their G4 500MHz chips only had 450MHz stamped on them. Apple's tech note on it said that the chips did not need to be replaced. I've heard a number of computer dealers have been telling people who complain that messing with their motherboard/heatsink voids their warrenty, so stop looking. Hehehe
:P
The other thing is that living in Perth, I can attest to how downright dishonest some of our computer dealers are. I wouldn't put it past some of them.
Would have been interesting to see the chip before it got crushed
What you suggest would benefit people who bought systems that were overclocked with untouched processors, but the problem here is that the processor cartidge has been cracked so that the mobo thinks its getting a faster processor.
Post may contain irony: discontinue use if experiencing mood swings, nausea or elevated blood pressure.
But as Lucien points out, checking this out will void your warranty.
In otherwords, by opening this package (in order to determine its terms), you assent to being bound by whatever's contained therein. Hooray for shrinkwrap liscenses on hardware. At least there's no prohibition of reverse engineering....
"If one is really a superior person, the fact is likely to leak out without too much assistance" -- John Andrew Holmes
This has happened to Compaq products recently here in the UK - although this hasn't been Compaq's fault. What's worrying is that dodgy kit was supplied by reputable authorised Compaq dealers. The kit appears to function at first, but then servers starting having crashes with unusual stop errors.
It appears these Athlons work when installed, although there must be hundreds of Athlon users taking a second look at their processors right now.
insignificant sig
It strikes me ironic that he's complaining about getting a pre-overclocked processor.
Not really - if he was planning to overclock it to (say) 800, he will probably be annoyed to find he can't, as it has already been upped to reach 700, and without a cooling device (which probably means that in normal service it would have been more likely to fry than a "real" 700.....
--
-=DaveHowe=-
To save everyone the time :
Article Moved:
Do to the HUGE response to this article, we have had to move it to a different server. We are currently making arrangements and will notify ASAP.
Update: It is now available/mirrored @ http://ocworkbench.com/hardware/athlon/athlon.htm
Kind Regards,
Lucien Wells.
//Phizzy
"Most European technology just isn't worth our stealing," -- Former CIA chief James Woolsey, referring to Echelon
Well, on one hand that's true. But on the more practical side of things, I want to avoid retailers who are lacking either the scruples or the savy to get me the chip correctly marked. To a certain point, I don't care who did it, I just want to get what I pay for.
Trees can't go dancing
So do them a big favor
Pretend dancing stinks!
B.S.
While the vast majority of computer dealers are on the up and up, there are some very shady dealers out there. There are people that will buy hardware from anyone, without asking where the parts came from.
I used to work for a major microprocessor manufacturer in customer/technical support working very closely with the customers/victims of these shady charecters. I've met slimy hardware dealers that wouldn't care if a processor was legitimate or not. They get your money with a warranty thatlasts as long as you are in the store. They explicitly sell kit "as is" with no warranty from them. If the part fails, go talk to the manufacturer. What, they won't help you because the processor is counterfit? That's a shame. Too bad that you bought it "as is" If you wanted a warranty, I could have sold you this boxed processor for 40% more. I can assure you that I had no idea that the part was remarked. I can provide you with all the informaiton that I have on where I purchased the chip from, if you want to press legal matters further. I assure you that I had no intent to defraud you or any other customer.
To say that the vendor is ignorant in this matter is nieve at best, intentionally decitful in the worst.
Do you own a small computer shop? If you do, I'm certain that you know how to purchase legetimite parts from authorized channels. I'm also equally certain that you know of at least one "grey market" vendor that can get you parts for a discount.
I'll give you that clock / bus locking with substrate resistors is a stopgap measure at best. The only real wat to lock bus / clock speeds is to make microfuses in silicon, and blow the correct fuses in silicon during the test/burnin cycle of manufacturing. until this process becomes perfected (It's only a matter of time) fraud will be a fact of life in the computer industry.
-Goldmeer
(Boy, do I have some nasty stories of remaking that i wish I could tell...)
All you need is a a some plastic molding equipment (to make new cases for the processor with the new speeds molded in just the right way), some soldering irons (to make the modifications to the PCB), a quality printer and stock (To make lookalike boxes), and a shrinkwrap machine.
These things are not really bulky or have high power requirements, unlinke years past, where you had to have a laser to etch the ceramic of older processors. These types of operations can easily be done in a mobile home or on a boat. That truely gives a new defination for an offshore enterprise, no?
That makes this type of operation very difficult to shut down completely.
Clueless moderator: I fail to see how the above post is insightful.
:)
In answer to the question, however, remarking an Athlon is far easier than modifying the equivilent coppermine. The Athlon multiplier settings (and how to change them) are actually well documented (see Tom's Hardware), and AMD even designed an overclocking system into the packaging (Gold Fingers).
The intel p3, on the other hand, has no available documentation for changing the clock multiplier. In your dollar bill analogy, the effort required to make the correct changed would be stated as, "Why make counterfeit one dollar bills, if you could make four dollar bills?"
Answer: It's not nearly as cost effective, and MUCH easier to spot.
What part of "shall not be infringed" is so hard to understand?
Anyone else notice that the picture here looks like the sticker has been peeled off and replaced? I also cringe at the poor quality of resoldering the resistors in this one. EEEk!
It's too bad that you can't find out for sure whether the processor is fake without voiding your warranty. Maybe AMD can help us out here and let us ask them about the serial numbers. I guess we should all expect tampering more often.
Thanks to TechWatch the people who did this now know that they need to use a better plastic, a sharper defined mold and better quality stickers. They'll spend a few bucks 'fixing' these issues and be back to ripping people off.
I second the vote for adding a CPUID field for what MHz the chip was designed to run at. Overclockers wouldn't care but it would prevent this from becoming yet another reason to get a good 'ol Intel chip instead.
There doesn't seem to be much in the world that people won't try to get away with. Too bad the people who did this are probably in some unidentifiable third world country.
Think, these people had to spend a minimum of $100000 for a limited production run on the replacement casings and stickers. They had some financial backing, probably. Can you imagine their business plan: "We need money so we can pay for staff, tools and supplies so we can rip off optimistic people looking for bargain processors"...
What a pain in the ass. Yet another sign that the world is coming to an end.
Actually, black for heat dissapation is not a new idea, when rebuilding my VW bus engine, the rebuild manual said I should leave the black paint on for that reason (They're air cooled and need all the help they can get staying cool.)
Supposedly they make special coatings for this purpose, not your average Kryon-bomb can of black. If the chip manufacturers are using some kind of thermal dissapation material for their casings, it would probably make the chips less reliable than before to simply slap a cheap plastic jobber on there.
(Of course the previous owner had already removed it, so I buffed it out real nice with a buffing tool. I also felt pretty stupid a few months later when it dropped a valve seat. Sigh. )
Anybody who runs intel can go download a program from their site which gives clock speed, and with the PIIIs the expected speed as well. Don't even have to run windows, there's a version that boots off a disk.
If not now, when?
(1)If you have an ASUS board, you can run the PC Probe utility that came with the board. It will give you info like: (under the info tab, DMI Explorer...)
Version: Athlon
External Clock: 100Mhz
Max Speed: 800Mhz
Current Speed: 700Mhz
(2) If you have a program like Intel LAN Desk, you can use the DMI Explorer to get the same information.
(3) Machines such as Compaq's and HP's come with diagnostic utilities that also explore this information...
This should match what you see on the case of the processor. If they don't match, or your clock frequency is running over 100Mhz, be suspicious!!
"History doesn't repeat itself, but it does rhyme." Mark Twain
Um, no. I don't think that's what he was suggesting at all. What I think he was suggesting--and what seems to make sense to me--is that sometimes education is pointless. This case would be a prime example. For reasons mentioned above, I don't think education about overclocking would lead to any less exploitation. So while education may be a good defense against exploitation sometimes, it doesn't always work and isn't always going to accomplish something.
Why not a Pentium III, since you could get more monedy from it. In otherwords, why make counterfit $1 bills, if you could make $20's? tcd004
well, insincere as this is, if the pirates perceived enough market demand for amd chips (as inexpensive as they are, compared to intel) then I guess AMD chips are being considered more and more for end-user systems.
so even though the fake cpus are a hassle to put up with, the good news is that AMD is becoming a more and more serious threat to intel and a real alternate choice for x86 style cpus.
(ob consp theory: maybe intel is doing the fakes, trying to taint the good name of AMD?) [vbg]
--
--
"It is now safe to switch off your computer."
Years back I wanted to upgrade from my Intel Pentium 120MHz to a K6 200MHz. I bought the CPU and ran it. It was so unstable and hot that I was forced to return it 2 days later and get a Pentium 200 (for 2x the price).
A week later the store I got it from told me that the K6 I had purchased was a remarked K6 166MHz. Which explains the high heat and instability.
I have seen everything from remarked P120s (marked to 133MHz) to K6's, and now apparently Althons.
Some people seem to think that they won't get the wool pulled over their eyes because they will looks for all the little signs. Let me say that some of the remarkers out there can do as good a job as AMD and Intel on the cases and covers. You would swear that the K6 I had was real, but I saw with my own eyes the test software reporting that it was a 166MHz CPU. The only way to foil them is to read the speed from the CPU core.
The real problem is that the decision how to clock the CPU is made by the marketing department, not the engineering department. The difference between the chips which are clocked differently is really only one of quality. The good ones don't heat up as much at higher speeds. Wouldn't it be a funny coincidence if a foundry just happened to produce the same quantity of each quality that the marketing guys think they need?
The real irony is that the chip makers suffer from the fact that their quality is too good, so they downgrade a lot of their production. The reason that it's so easy to fake chips is that they all come off the same assembly line. The measures to prevent overclocking are all cosmetic.
I bragged about my Karma at a job interview but I didn't get the job.
All the comments about karma whoring aside, really, Sig, that's an incredibly dumb idea. The public does not need to be educated about what overclocking is. The majority of the public does not need to know that overclocking takes place or how to do it. They just need to know that there are fake Athlon chips out there, and honestly, the 'public' doesn't need to know that either, it's the resellers who need to know. I doubt the shop that sold this chip to this guy realized they were selling a fake chip. The retailers need to run their chips through a test of some sort, some test that doesn't require that they actually boot up a computer with the chip, but maybe a simple electrical or physical test.
Education may be the only defense against exploitation, but you have to be careful who you educate. Educating the public in this case would be futile because half of them wouldn't understand what you were talking about and the other half wouldn't care (not to mention the incredible resources necessary for 'educating the public'). It's better to educate the knowledgable public (ie. the ones selling the stuff) so they can provide good service and abide by the law (which is plenty sufficient, BTW).
If everyone knew everything, that'd be great, but there's no chance that that will ever happen, so make sure you get the right people to know the right things.
I think the public needs to be educated about what overclocking is, how it works, and how to detect it. The industry needs to stop trying to prevent the dissemination of information if it wants to prevent widespread abuse. First, allow enthusiasts to change the clock settings. If they blow their chip up, fine - make it blow an EEPROM if you're worried about warranties should they wish to change the clock setting. Second, make it happen in software - like with the Asus "softmenu" boards.
Education is the only defense against exploitation. The law is insufficient.
I don't think these will help you tell much beyond what stepping number of processor you have. Since 650 and 700 are probably produced on the same process (and the better quality ones are marked 700), it probably won't help.
--
I'm the inventory controller and technician and a local "Mom and pop" computer service center. It's my job to check in ALL the inventory that comes our way. It's also my responsibility to check for *good* hardware. About 5-6 months ago i recieved a K6-III 450 that was on order. There was no etching on the cpu cover plate. The specs we're merely penciled in!?!$ Well...after calling the vendor (who gave me an RMA number), i got a little pissed. I mean who would want to give their customers a bogus part! I wound up talking to Corporate Investigations at AMD. They were OUTSTANDING. What they did was: 1) replace my processor after hearing about the bogus one. (they said it was a "test-fab unit" that should have never been released to the public..) 2) Contacted the vendor we purchased it from. 3) Offered the vendor to replace ALL their bogus cpus
with the real thing at no charge..IF they told AMD who was supplying to them. very nice if you ask me. Felt good too...ensuring that my customers and evyone elses got good solid product.
you never lose in ure razorblade shoes......Beck-Hotwax
The site notes that the Athlon core was the .25 die size and 650 Mhz. Now AMD was selling these 650 cores underclocked at 500MHz, because it was cheaper than going back and making 500 Mhz cores. Pricewatch lists Athlon 500 at ~130, 650's at ~150-160, and 700's at ~190. I think that whoever did this bought Athlon 500's (with the 650 core), took 10 minutes of soldering to undo the underclocking by AMD and overclock it a little to 700, and instantly got over $60 profit per chip (probably more if they got the 500's in bulk). I don't think anyone would take the time to solder a 650 into a 700. Also, I would be very pissed because the guy wanted a 700 because of the .18 die size, which is more overclockable than the .25 of the 650 core.
Am I reading this correctly?
In order to determine if your Athlon is counterfeit, you have to void your warranty.
Someone want to call up AMD and ask them how they suggest we then find out?
Now the problem with fakes shows up. Firstly, it have been said on top of this comment that "now everybody knows how to do it.". Security by obscurantism is no security.
Now to the practical side: To what temperature does a non-overclocked Athlon's plastic case get ? Is it imperative that the case be black for thermal dissipation or could it be transparent so that people could peek at the core inside without voiding the warranty ? Not a translucent case like iMac but a truly glassy plastic.
Not only it would be good to avoid frauds, would look nice too. :)
damit I think they riped me off, when I opened my Althon box all I found was a fried egg and 2 strips of bacon. It actucally worked out well because I was hungry and the egg was sunny side up, Mmmm my favorite. Oddly even though, the bacon did clock in at 700Mhz without over heating, but now my dog is sniffing at my computer case...
I should of notice when grease was driping from the bottom of the box, and another thing, how the hell can you have a good breakfast without hot grits? What are these people thinking.
Atleast Intel gives you free coffee refills
"`Ford, you're turning into a penguin. Stop it.'" -THHGTTG
My experience comes from messing with recycling plastic on a hobbyist level - remolding plastic soda bottles and the like. It can be done very cheaply - see description for this book.
The description of the case - brittle, shinier than the original - sounds like plastic that has been remolded. I propose that the counterfeiters might be just remolding the original cases. Any plastics engineers want to comment?
I wish they actually disclosed the name of the retailer. Such retailers hurt customers AND AMD, and their names should be widely known and disclosed.
Go get your free Palm V (25 referrals needed only!)
AMD has several useful utilities to determine information about your Athlon. You can find them at the following URL:
http://www.amd.com/products/cpg/bin/
Among the programs listed here are things to measure the clock speed, find the CPUID, and other information. Binary and sources are available.
I have no clue how useful this would be in these cases, but it's certainly worth looking at.
What I really wish would happen would be for BIOS manufactures and CPU manufactures to get together and make the BIOS display both the real CPU speed and what it is currently running at in large letters during POST. If it was overclocked it could even display a message stating that in a very clear way (text that fills the whole screen saying "Warning: Processor overclocked"
People who overclock wouldn't be hurt because the banner would be there only for 5 or 6 seconds and would remove Intel's excuse so maybe they would stop locking the clock (yeah right.)
Normal people would freak out if they put thier new 750Mhz Duron in thier motherboard and see such a message.
I am sure that some remakers would get around this by hacking the BIOS, but it would stop people from getting remarked CPU upgrades. Also if someone happend to flash thier BIOS in the future after buying a new PC, they would be informed of the trickery then.
As far as software, could H. Oda's CPUID program be used to verify what exactly you have in your system. It tells me I have a Celeron 300a running at 450MHz. Maybe it is different with AMD CPUs.