×

Welcome to the Slashdot Beta site -- learn more here. Use the link in the footer or click here to return to the Classic version of Slashdot.

Thank you!

Before you choose to head back to the Classic look of the site, we'd appreciate it if you share your thoughts on the Beta; your feedback is what drives our ongoing development.

Beta is different and we value you taking the time to try it out. Please take a look at the changes we've made in Beta and  learn more about it. Thanks for reading, and for making the site better!

AMD Readies "Lottery-Core" CPUs

CmdrTaco posted about 4 years ago | from the don't-roll-snake-eyes dept.

AMD 80

Barence writes "AMD has announced a radical shake-up of its CPU strategy, in an exclusive interview with PC Pro. The company has revealed that the next generation (codenamed Tyche) will be offered as a single 'lottery-core' SKU, with the number of functional cores in each part left for the customer to discover. 'We know gaming is very important to our customers,' explained regional marketing manager Ffwl Ebrill, 'and we're innovating to bring that win-or-lose experience out of the virtual world and into the marketplace.' Anyone discovering more than ten functional cores could consider themselves 'a lottery winner,' while unfortunates discovering their new CPU had no working cores at all would be encouraged to 'roll again.'"

cancel ×
This is a preview of your comment

No Comment Title Entered

Anonymous Coward 1 minute ago

No Comment Entered

80 comments

Incremental! (5, Funny)

TrisexualPuppy (976893) | about 4 years ago | (#31701818)

How about you get the number of cores proportional to the post number? FIRST!

Wait a sec...holding off on this one...

MOD PARENT UP! (0)

Anonymous Coward | about 4 years ago | (#31701850)

Coherently funny I say...or insightful lol!

Re:Incremental! (1)

pitchpipe (708843) | about 4 years ago | (#31702020)

How about negative cores:

Turn that spreadsheet into a random jumble of bits! It will even DECREASE the amount of entropy in the Universe!

Re:Incremental! (0)

Anonymous Coward | about 4 years ago | (#31702300)

Didn't Intel already do that with Pentium 4?

Re:Incremental! (0)

Anonymous Coward | about 4 years ago | (#31702550)

How about you get the number of cores proportional to the post number? FIRST!

Wait a sec...holding off on this one...

unless you order from newegg. then it doesn't work regardless of cores lol

Hmm....... (0, Troll)

ivanwyc (1649687) | about 4 years ago | (#31701820)

April 1st?

errr (-1, Troll)

Anonymous Coward | about 4 years ago | (#31701822)

A bit like the "will it or will it not crash policy Windnose guarantees you

Uh.... (1)

mdm-adph (1030332) | about 4 years ago | (#31701824)

Okay, I know it's a joke, but isn't it kinda true? Grab one of AMD's chips with disabled cores, and it truly is random how many cores (if any) you might be able to unlock using ACC. :P

Re:Uh.... (4, Funny)

LostCluster (625375) | about 4 years ago | (#31701876)

Let's not forget NewEgg's recent version of this promotion with their "Zero chips--- you're a loser!" packages.

Re:Uh.... (1)

kimvette (919543) | about 4 years ago | (#31702338)

I know it's April fools, but this is exactly the kind of marketing one would expect from a company which merged with ATI. So, I can't tell if this story is real or not. ;)

How to rig a lottery (1)

LostCluster (625375) | about 4 years ago | (#31701834)

The thing about lotteries is that they defy physics. Take a machine designed to give out specific units of force, and balls designed to weigh the same weight they are prescribed, and you've got a very predictable physical system.

So, how do they avoid drawing the same numbers every time? They let a computer mess up the given statuses... do they use Set 1, Set 2, or Set 3 of balls? Do they put the balls in numerical order? When does the machine start moving?

The thing is... you can't let humans decide these variables because eventually they'll spot the patterns and be able to rig the result. So, they do what the computer tells them to... yep, that's right, your state lottery is just as random as the PRNG at headquarters.

Re:How to rig a lottery (1)

Tubal-Cain (1289912) | about 4 years ago | (#31701866)

The thing about lotteries is that they defy physics. Take a machine designed to give out specific units of force, and balls designed to weigh the same weight they are prescribed, and you've got a very predictable physical system.

Predictable for how long?

Chaos theory (3, Insightful)

tepples (727027) | about 4 years ago | (#31701898)

Take a machine designed to give out specific units of force, and balls designed to weigh the same weight they are prescribed, and you've got a very predictable physical system.

Even if it is predictable, it isn't necessarily tractable. The air inside one of those table tennis ball blenders, for instance, is a chaotic system [wikipedia.org]. The "specific units of force" aren't always constant given fluctuations in power supply.

your state lottery is just as random as the PRNG at headquarters.

More likely, the PRNG that dictates exactly how long the machine runs isn't entirely pseudo but instead tied to an entropy-gathering process such as hashing room noise received through the microphone.

Re:How to rig a lottery (4, Insightful)

CyprusBlue113 (1294000) | about 4 years ago | (#31701928)

Except they don't, because the vast majority of the systems are based on the concept of turbulent flow of a fluid (in this case air generally), which is for all practical purposes due to the number of variable points of deflection impossible to model for any time period signifigant enough to allow for predictions of these machines as they are designed to long pass this point before the balls would lock in position.

Heck trying to model turbulent flow on a fixed path is hard enough, trying to model turbulent flow through a mass of shifting floating deflectors is downright masochistic.

Re:How to rig a lottery (0)

Anonymous Coward | about 4 years ago | (#31702016)

It can't be *that* hard. Heck, I will give 50 million dollars to whoever can develop such a system and provide reliable predictions. With a bounty like that, there is no way the 'lottery ball problem' will go unsolved for long.

Re:How to rig a lottery (0)

Anonymous Coward | about 4 years ago | (#31702124)

Why even claim the prize? Just cash in everytime the lottery pool goes over 200 million. Why waste that opportunity for a mere 50 million? Maybe adding a figure or two to the prize would make it rewarding.

Re:How to rig a lottery (0)

Anonymous Coward | about 4 years ago | (#31703094)

Oh hey, it looks like you missed the joke!

Ah well, better luck next time!

Re:How to rig a lottery (3, Insightful)

strength_of_10_men (967050) | about 4 years ago | (#31702098)

and given that you have to lock in your picks before you even know the initial state, i wouldn't worry too much about it.

Already attempted by Intel. (5, Funny)

llvllatrix (839969) | about 4 years ago | (#31701838)

Didn't go over so well. Just ask newegg.

Re:Already attempted by Intel. (1)

fuzzyfuzzyfungus (1223518) | about 4 years ago | (#31704102)

We can only hope that the supplier they traced it to, once they realized that they were totally fucked no matter what they did at that point, responded: "You can't make a profit without breaking some neweggs..."

Give me an f-ing break (4, Informative)

fredrated (639554) | about 4 years ago | (#31701840)

Stop with the April fools posts already!

Re:Give me an f-ing break (1, Redundant)

toastar (573882) | about 4 years ago | (#31701854)

Stop with the April fools posts already!

This one isn't even funny

Re:Give me an f-ing break (0)

Anonymous Coward | about 4 years ago | (#31702206)

Which one was?

Re:Give me an f-ing break (4, Informative)

Beelzebud (1361137) | about 4 years ago | (#31701870)

I used to like April Fools day. Now it's just annoying. Every web page on the internet is basically useless for a day.

Re:Give me an f-ing break (5, Funny)

PolygamousRanchKid (1290638) | about 4 years ago | (#31701904)

Every web page on the internet is basically useless for a day.

Most pages on the internet are basically useless for every day of the year.

Re:Give me an f-ing break (2, Funny)

Anonymous Coward | about 4 years ago | (#31701966)

I like MySpace you insensitive clod!

Re:Give me an f-ing break (1)

steelfood (895457) | about 4 years ago | (#31702080)

Just 'cause it's April 1st doesn't automatically mean that everything posted on this day is funny. Guess what? It's not funny! Insightful! Informative! But not funny!

Re:Give me an f-ing break (0)

Anonymous Coward | about 4 years ago | (#31704662)

I push for a federal ban on April Fools day. Who's with me?

Re:Give me an f-ing break (0)

Anonymous Coward | about 4 years ago | (#31701872)

slashdot.org 127.0.0.1

hth.

Bringing together news from a wide array of source (0)

Anonymous Coward | about 4 years ago | (#31701900)

What part of "news aggregation" is unclear?

Re:Give me an f-ing break (1)

rdnetto (955205) | about 4 years ago | (#31714084)

The worst part is that they're all showing up as posted on the 2nd for me. Guess it's cause I'm in GMT+10 (Aus).

Discovering their new CPU had no working cores (0)

Anonymous Coward | about 4 years ago | (#31701864)

Users discovering no working cores are expected to turn up in person with their cpu and motherboard for a stoning. And not the good kind of stoning.

Re core lottery (1)

freddieb (537771) | about 4 years ago | (#31701888)

Actually it would probably work as long as the cpu had enough cores to be worth the cost. I would go for it!

This will not be good for less well off consumers (0)

Anonymous Coward | about 4 years ago | (#31701980)

This is what will happen:

Folks will be hording them and testing each one and when they find one with a lot of cores, they will then eBay them at a premium. The rest of us poor slobs will have to settle for the 1, 1 1/2, 1 5/16ths core CPUs while the well to do will get the 2,4,6,8,10 core CPUs.

It'll be great for AMD sales but I see it as hurting them in the long run.

Just stop it ! (those April 1st Jokes) (1)

burni2 (1643061) | about 4 years ago | (#31701942)

Those jokes aren't even funny nor interesting.

They are just bad.

Except the youtube 1/4 that thing was good !

TAFF! (0)

Anonymous Coward | about 4 years ago | (#31701984)

"explained regional marketing manager Ffwl Ebrill"

'Ello taffy boy :P

Wow, that's nice they... (0)

Anonymous Coward | about 4 years ago | (#31701992)

are naming this after that guy at Penny Arcade.

The luckiest one gets what (1)

roman_mir (125474) | about 4 years ago | (#31702006)

It is all fun and 1 of April games until someone gets a 0-core processor.

Re:The luckiest one gets what (1)

natehoy (1608657) | about 4 years ago | (#31702714)

I'm thinking of installing a small generator on my abacus powered by the motion of the beads and calling it a minus-1-core processor. I put processing in to it, and I get energy out.

Isn't it what AMD already does? (0)

Anonymous Coward | about 4 years ago | (#31702040)

I mean come on, twelves cores? What happened to 8 or 16?

Oh, I get it. It's probably 16 mebicores.

Triple Cores? (0)

Anonymous Coward | about 4 years ago | (#31702082)

In a way AMD already does this with the X3. Sometimes the fourth core can be unlocked giving a quad core for the price of a triple core.

Even before multi cores, overclocking provided an aftermarket solution to getting more then you paid for out of a chip. But was dependent on the quality of the chip itself. Binning reduced the random chance by some amount but still left some degree of randomness.

Re:Triple Cores? (1)

BitZtream (692029) | about 4 years ago | (#31702150)

This is in no way unique.

You can essentially 'overclock' everything in existence to some extent.

It generally doesn't happen outside of PCs because most people realize that you're trading off something else.

In the case of overclocking you're giving up reliability and data integrity in exchange for a marginal speed boost because you were too cheap to buy the better chip or you're just going nuts trying to see just how far you can push it.

Overall though, unless you're 15 and still inschool with no job, Overclocking is always a waste of money long term. The chip may be cheaper, but any gains you made are lost almost immediately with all the fucking around you have to do to get it to work. Your time has to be worth less than minimum wage for it to make sense to overclock for anything more than a 'I made my chip run at 6ghz' factor.

There is a reason the chips are slowed down and cores are disabled, they don't pass the required tests for better performance. I guess the other reason legitimate reason for doing it would be gather entropy for a PRNG, using the psuedo random errors introduced by the chip itself to add entropy, but its far easier to just wire up a Geiger counter or a microphone and use background radiation.

Re:Triple Cores? (1)

natehoy (1608657) | about 4 years ago | (#31702308)

But the triple cores truly are a lottery.

If you're going to fabricate a bunch of quad-core chips, you expect a certain number of them to end up with 1 bad core, a certain number to have 2 bad cores, and some will have 3 bad cores, and some will just be garbage. That's the nature of the beast. AMD is just smart enough to test them until they hit their desired yield of quad cores, then test until they have their desired yield of triples, and test the remaining ones and those either become double-core or get thrown out. I'm not sure if they have any single-core processors that started life as quads.

The benefit to the consumer is that, while the number of cores is lower, the cache is still high, so a quad core with two cores tied behind its back may still be faster than a purpose-built dual core. And, heck, it was a useless part as a quad due to one or more bad cores, so AMD is making some good money selling these "defective" parts as "partly good". They can charge a slight premium over purpose-built duals, and it's pretty much "found money" to them.

The other benefit is that, for some of those units, the triples (and maybe even the doubles) are not "tested bad", they are "untested". AMD needs a certain amount of doubles and triples in those lines to meet demand, so whatever is left once they meet demand for quads is a mix of tested-bad and untested chips.

So the "Lottery Core" CPU is not vaporware. It exists today.

Huh? (0)

Anonymous Coward | about 4 years ago | (#31702084)

For a moment there I thought I am reading The Onion.

April Fools (1)

Delusion_ (56114) | about 4 years ago | (#31702086)

Enough already. It's done to death and not funny.

At least the OMGPONIES joke was fun, since it didn't rely on an endless avalanche of stupid fake news stories.

Dontruinitwithtags ... it was already ruined. (3, Interesting)

BitZtream (692029) | about 4 years ago | (#31702120)

To whomever decided to add the 'dontruinitwithtags' tag ...

IT WAS RUINED ALREADY WITH THE OTHER 9 GOD DAMN APRIL FOOLS POSTS. ENOUGH ALREADY.

I get excited at a glimmer of hope of something to read while I sit here and wait for this batch to run ... only to find out another moronic April Fools post was made ... because 10 in one day just isn't enough.

Re:Dontruinitwithtags ... it was already ruined. (1)

MORB (793798) | about 4 years ago | (#31702158)

Who would be stupid enough to believe this anyway even in the absence of tags and of a billion other jokes?

Re:Dontruinitwithtags ... it was already ruined. (0)

Anonymous Coward | about 4 years ago | (#31702476)

Not everyone lives here ya know. Maybe you should get a job that requires more out of you than, well, reading 9 April Fools stories.

Whatever Happened To... (0)

Anonymous Coward | about 4 years ago | (#31702176)

AMD's plans to get multi core processors appear as one very fast core? I can't imagine how this could be done (presumably it couldn't and they gave up) but it would have been very useful for me since I voted "Below Average" in the current slashdot poll. Whenever I program a multi-threaded real-time system it always ends up deadlocking in some situation that I can never recreated and it takes me all eternity to work out why.

One very fast core would make things so much easier so I was looking forward to this technology but I never heard anything about it beyond the initial announcement. Did they every publicly say they'd given up?

It's not a joke, quite. (1)

Areyoukiddingme (1289470) | about 4 years ago | (#31702198)

Isn't there a tech site that sells a daily "Bag of Crap"? People buy it, every day. Sometimes it's full of nice stuff. Often it's crap. People buy it. Because the price is right.

So, same here. I'd go for it. I'd pay $10 for a "Box of Crap" from AMD. It can have anywhere from 0 to 12 operational cores. As long as neither I nor AMD know ahead of time which it is, I'd buy it. I'm willing to gamble my $10 on the yield of AMD's production line.

'course, given their yield, the price would probably have to be higher than $10. They make quite a few usable chips. Lottery tickets are $1 because the yield is so low.

Re:It's not a joke, quite. (0)

Anonymous Coward | about 4 years ago | (#31702772)

The local Dollar Store does that. Small paper bags filled with unknown items for $1.

I've watched in utter amazement as people rip open 5-6 of the bags, consolidate all the good stuff into a single bag, grab a few things off the shelf and add them, leave all the other bags and their contents scattered about the shelves, then take the torn bag that's obviously overfilled with stuff to the cash register to pay their $1.

I don't go to dollar stores often. When I do, I find myself looking for a sharp enough piece of plastic to slit my wrists.

Believable if this had been Intel (0, Troll)

kbg (241421) | about 4 years ago | (#31702222)

Well if this had been Intel then this would have been believable since they already have a lottery-virtualization core, that is you don't know if your CPU supports virtualization or not even in CPUs of the same generation and family. However this is AMD so this must be an april fools.

Copycats.. (0)

Anonymous Coward | about 4 years ago | (#31702234)

This has been Intel's strategy for decades.

Save money on QA (2, Insightful)

DarthVain (724186) | about 4 years ago | (#31702236)

A boon for Overclockers really..

No longer are things priced by binning, now every one has an equal chance.

I would seriously buy one if the price was right. I mean I already own a C2D 1.8 I bought on the assumption that I could overclock the thing to 3.2 or whatever, and then was sorely dissapointed when it couldn't take an OC at all. I mean the ONLY reason I bought it was to OC it.

This would fill that niche. You don't have to sort or test them, heck, you don't even need to market them... what are you gonna say... erm we don't know how fast it will run, nor do we really know how many cores it contains... buy one and find out! Sure I will, just price 'em cheap and I would be all over that! (so long as your secretly not binning all the good ones and just selling rejects!)

I dunno, sounds find to me if the price is right. Overclockers everywhere would have a field day posting their finds! Not sure how you would do the package with a possible variable number of cores. I guess your MB would have to be uber compatible with everything...

Yes yes I know it is an April Fools joke. I still want one for fun though! :)

Re:Save money on QA (0)

Anonymous Coward | about 4 years ago | (#31703312)

AMD already kind of does this. They have random SKUs every once in a while. For a current one, do a search for the Athlon X2 5000+ (make sure it's the 45 nm one clocked at 2.2 GHz stock. It's basically an AM2+ Phenom II X4 in disguise.) The only issue is the low 11x multiplier, but for ~$40... yeah. (I just snapped one up after seeing it on Slickdeals.)

Re:Save money on QA (1)

twosat (1414337) | about 4 years ago | (#31704840)

Sir Clive Sinclair and his Anamartic company were involved in something like this with "Wafer-Scale Integration" many years ago which used the "Catt Spiral" developed by Ivor Catt to make solid-state memory disks. Mr Catt is proposing something very similar to the "Lottery Core" cpus with his plans for the "Kernel Machine" http://www.ivorcatt.com/3ew.htm [ivorcatt.com]

Dangerous Precedence (1)

CherniyVolk (513591) | about 4 years ago | (#31702294)

First, I am a fan of AMD and I'm glad they are being creative with how they might approach the market in the future.

But, this "lottery core" concept, if legal, is a very dangerous precedent pawned off to the consumer. Like atm transaction fees, or any financial transaction fees whatsoever, the only reason they persist is because no one has decided to sue for the practice; all those who might afford the lawsuit in some way benefit from the practice or the short changing effect has negligible effect on them to care.

Anyways, bottom line, is this seems to be an attempt to circumvent all consumer merchandise laws, lemon laws and all other business codes of transaction of goods and services with the justification that it's a "lottery system".

AMD get's away with selling defective products, with refusal of refund, what's to change anyone else from doing so? How would you like for to tell you that you either continue chasing the repair costs or buy another one of their cars? Who is going to regulate AMD to make sure they don't pump defective cores to encourage increased sales volumes (thus profit)? Real lotteries, licensed gambling establishments... even financial services like the stock exchange and investment ventures are all HEAVILY regulated, by organizations that are state/government sanctioned and have real teeth and have in the past used their teeth. AMD is simply a business... who is going to regulate them? Is the Nevada Gaming Commission going to govern all the gaming machines that determine which chips are lotto winners? Are they going to come in and close production lines indefinitely until the algorithms are fixed? Are they going to make sure there are no back-doors? Are they going to prosecute whoever if fraud is found? What about all the merchantability laws that protect consumers purchasing goods?

This concept is not new... real lotteries exist. What is new more or less is to expand this as a means justify pawning defective goods off on the consumer as a 'lottery system' based purchase.

I strongly condemn this proposal and I hope to god this is an April Fools joke.

Re:Dangerous Precedence (0)

Anonymous Coward | about 4 years ago | (#31703112)

This is clearly a joke.

But thanks for the thesis on the what the ramifications of this story would be if it were true.

You must be a blast at parties.

In other unlikely news NVIDIA ships Fermi (0)

Anonymous Coward | about 4 years ago | (#31702548)

Fermi finally ships today, April 1st.

Intel beat them to this (1)

G00F (241765) | about 4 years ago | (#31703148)

Intel was already getting ready to roll out their own lottery-core [slashdot.org] However there was a bit of public backlash because units leaked out early.

Check for New Comments
Slashdot Account

Need an Account?

Forgot your password?

Don't worry, we never post anything without your permission.

Submission Text Formatting Tips

We support a small subset of HTML, namely these tags:

  • b
  • i
  • p
  • br
  • a
  • ol
  • ul
  • li
  • dl
  • dt
  • dd
  • em
  • strong
  • tt
  • blockquote
  • div
  • quote
  • ecode

"ecode" can be used for code snippets, for example:

<ecode>    while(1) { do_something(); } </ecode>
Sign up for Slashdot Newsletters
Create a Slashdot Account

Loading...