My Xbox360 Drama

[Sky]

I think we all know how I feel about pc gaming on my 61" hdtv + 5.1 surround. ;)

It really shows off the ugly of last gen consoles like the xbox and gc (I don't own a ps2). I'd buy a 360 if I wasn't saving for a Conroe or Kentsfield pc next year, but I wouldn't play shooters on it. Fingers > thumbs.

[Morfiend]

So, it looks like my new Xbox lasted a week. Today it started freezing on me every 10 minutes, and I got an error at start up one time. I think it was error E64 or some thing like that, which when I looked it up was hardware failure.

Looks like its time to take it back and get another new one.

[schild]

Wow.

[Strazos]

I hope this place is close to your house.

[Oban]

I was about to buy the XBOX360 premium today, but remembered this thread about the reliability issues. 

So instead of spending money on an unreliable console, I added my name to the PS3 preorder list at my local EBGames.

(Fixed for Schild.)

[schild]

Oban, please please please tell me your post is a giant lie.

Cuz I'm getting a divide by zero error every time I read your post and I can't help but think blood is spurting out of my ears.

[Morfiend]

So I picked up my 4th console today. So far so good.

As to PS3, really all I can say is 

Even with the problems I am having, I would still much rather have a xbox 360 than a ps3. Wait till you see what kind of problems that things has at release.

I would suggest getting a Xbox 360, but get it some where that has a very good retun policy, like CostCo, or buy an extended warranty on it.

All you have to do is look at the games being released in the next 2 months, and you should be sold on a 360.

And im not a salivating fanboy like schild.

[schild]

Sorry for liking good games. You bastard.

[Sky]

I wouldn't buy a PS3 until they can reliably make 8 cores work. Some shady stuff going on there right now.

[schild]

Go go team service plan.

[Sky]

But failing cores is in the realm of 'as intended' last I heard.

[Rasix]

Go go team service plan.

For service plans, when you buy one does it cover you for repeated failures like Morphiend has had?  Does the counter reset for each new console or is it just for a set period of time?

I've decided against buying a next gen console until they seem relatively stable. PS2's releases should keep me happy for a while.

[Morfiend]

I didnt buy a service plan. I got my Xbox at Target, and they dont do that. It comes with a 90 day warranty, and this does reset every time I get a new console. Luckly my system went bad before my 90 days where up, this way I got a brand new console. If it had been after 90 days, I would have had to pay MS $130 or so to have them repair it(see ship me a refurbished console).

As far as I know most extended warranties that you buy do NOT refresh if you get a new system under the warranty.

So far my new system has been working great. It runs a lot cooler than my last 3, also it makes a lot less noise while running. Im keeping my fingers crossed.

[Trippy]

I wouldn't buy a PS3 until they can reliably make 8 cores work. Some shady stuff going on there right now.

It doesn't matter. Even if they can improve yields so that they don't have to disable one of the SPEs they won't stop doing it. All software is being written for 6 SPEs (the 7th is dedicated to OS security -- yes that's how fucking paranoid Sony is about people hacking the system). Suddenly having 7 available would complicating coding to the point that programmers would simply ignore it.

[Sky]

But the problem is 2 or more have been failing with frequency. That's getting into bad territory. Besides that, I find the whole idea of buying a flawed processor a bit weird. I buy a processor, I kinda want it to work 100%.

[Trippy]

But the problem is 2 or more have been failing with frequency. That's getting into bad territory.

I'm not sure what you mean by "failing with frequency". If you are referring to wafer yields then yes even with one SPE disabled the yields haven't been good, which is to be expected given the complexity of the process.

Besides that, I find the whole idea of buying a flawed processor a bit weird. I buy a processor, I kinda want it to work 100%.

There is no such thing as a "perfect" processor. Even though we think of CPUs as digital things they are still created by a very analog process and defects and imperfections are always present. This is why, for example, you have CPUs that can run at different frequencies. All the different speed CPUs come from the same wafers but because of the above variations some can operate faster than others. "Flawed" CPUs are also where you get things like the Celeron and Sempron CPUs. Those also come from the same wafers as their "parent" CPUs but because memory caches are particularly susceptible to defects, dies with many defects in their caches (but not too many) get binned as the "lesser" CPUs. In fact memory logic is so sensitive to defects that DRAM/SRAM and the like are designed with redundant logic on the die to boost yields so it's almost a certainty that the DRAM in your machine right now has various bits of logic disabled because of defects.

Edit: I should also add that "flawed" processors are rampant in the GPU world as well. Why do GPU manufacturers design GPU families where family members have differing number of pipelines and whatnot rather than just varying clock speeds? One reason is better market segmentation but the other is that it increases wafer yields. The dies that have too many defects to be the "top-of-line" GPU get tested as some lesser version with fewer pipelines or whatever and binned as such if they pass.

[Strazos]

Bleh, anywhere a person can read about the production process? Like, an article for someone who is not an electrical engineering graduate student?

[Trippy]

Bleh, anywhere a person can read about the production process? Like, an article for someone who is not an electrical engineering graduate student?

http://www.intel.com/education/makingchips/

[Yegolev]

When someone at work asks me how ______ stopped working or went wrong, I often say "Hey, I'm surprised these things work at all."

[schild]

When someone at work asks me how ______ stopped working or went wrong, I often say "Hey, I'm surprised these things work at all."

At my job I reply with "It's the internet. Half black magic, half user error."

[Sky]

Trippy, I know a bit about the processor process :P My point was that I don't buy the low-binned stuff, I always buy the high bins for exactly that reason. You can't do that with a PS3, because you can't tell which ones will make it into consoles. Or you can: the low bins.

“There are a lot of chips with six cores operational, and we’ve been thinking about whether we should really throw all of those away. We also have a separate part number for chips with all eight cores good. The stuff that’s going to be for medical imaging, aerospace and defense and data uses eight cores,” Mr. Reeves said.

http://www.xbitlabs.com/news/cpu/display/20060713074825.html

I just googled that one at random, I've read several articles saying the same thing. Actually, that one was pretty positive because it didn't mention that they expect them to fail over their lifetime, too. So you get a 6-core PS3 that fails down to a 5-core, say. Remove one for security and you've got 4-core, what if a game designer is programming with 6 free cores in mind and he has 33% the power he thought he had?

Should be interesting, but I'll stay on the sidelines until the high bins actually make it into consoles. I watched this same crap with Moto and IBM and PowerPC, the better chips all went to industrial uses.

[Trippy]

You seem to be equating wafer yield rates with in-the-field failure rates and it doesn't work like that. All chips have the potential for failing in the field but the chances of that happening to a CPU that's been properly tested and burned-in and that hasn't been abused (e.g. overclocked to heck or improperly cooled) is very very low.

[HaemishM]

Unless you are talking about first-gen Sony Hardware. They are not known for having good first-gen hardware durability.

[Sky]

Ok, I found the article I was referring to. Darn you to heck for making me dig :P

Electronic News: Do any of those cores ever go bad, so that you start out with seven and you wind up with six or five?
Reeves: There’s a reliability failure rate for all chip types. By definition, reliability failure is one point circuit that has failed. If it happens to be in an SPE, it will knock out one of the cores. We have electronic fuses now, rather than laser fuses, which you can only blow when you’re doing wafer tests. Electronic fuses you blow electrically. If you really want to be focused on reliability and up-time availability, you can design one of these chips to self-detect. You can ship it with eight cores working, blow one of them, and from a user perspective you would have self-healed it in the field.

Electronic News: But would it be as fast as the chip with eight cores?
Reeves: Yes, because the Playstation 3 only uses seven of them. You’d have a spare. That isn’t implemented in Cell, but it could be. We implemented that same strategy for IBM systems. If you take a logic hit on a chip, you don’t have any impact on performance because there is enough redundancy built in.

Electronic News: What happens if one of the cores blows on the Sony Playstation 3 if there are only seven to start with?
Reeves: It’s just like a reliability failure on your TV or DVD recorder. If it’s within warranty, you send it back. If it’s not, your game doesn’t work anymore. You’ll always have choices about how reliable you want to make a chip with burn-in. Most chips that go into the consumer marketplace on things such as camcorders or DVD players aren’t burned in. But you can add burn-in and improve reliability 5x to 10x. It’s extra cost. Certainly, a company like Sony adds that in.

http://www.edn.com/article/CA6350202.html

Add in what Haemmy said about first-gen Sony consoles...

[Trippy]

Unless you are talking about first-gen Sony Hardware. They are not known for having good first-gen hardware durability.

That was an overheating issue, just like on the Xbox 360 and I agree that first-gen hardware often has issues that get worked out later. If the design can't cool the CPU properly then yes you would expect an increase in CPU failures but that's still independent of the wafer yields.

[Trippy]

Ok, I found the article I was referring to. Darn you to heck for making me dig :P

Shouldn't have been that hard -- it's linked to in your Xbit article you linked. It still doesn't change my point about yields and in-the-field failure rates.

Electronic News: Do any of those cores ever go bad, so that you start out with seven and you wind up with six or five?
Reeves: There’s a reliability failure rate for all chip types. By definition, reliability failure is one point circuit that has failed. If it happens to be in an SPE, it will knock out one of the cores.

Yup chips can fail in the field, we've already established that.

We have electronic fuses now, rather than laser fuses, which you can only blow when you’re doing wafer tests. Electronic fuses you blow electrically. If you really want to be focused on reliability and up-time availability, you can design one of these chips to self-detect. You can ship it with eight cores working, blow one of them, and from a user perspective you would have self-healed it in the field.

This is the part that might be confusing you. When an SPE fails it's not because a fuse has blown -- the fuse is only blown if there's logic on the chip that does that explicitly which it doesn't have. Maybe someday it might when 8 SPE yields are high enough and Sony or IBM can justify the expense of redesigning the chip.

Electronic News: What happens if one of the cores blows on the Sony Playstation 3 if there are only seven to start with?
Reeves: It’s just like a reliability failure on your TV or DVD recorder. If it’s within warranty, you send it back. If it’s not, your game doesn’t work anymore. You’ll always have choices about how reliable you want to make a chip with burn-in. Most chips that go into the consumer marketplace on things such as camcorders or DVD players aren’t burned in. But you can add burn-in and improve reliability 5x to 10x. It’s extra cost. Certainly, a company like Sony adds that in.

If an SPE dies you send it in for repairs. Just like any other hardware failure. And burn-in reduces the chances of failure in the field.