I'm guessing you don't work on this stuff do you...
I never worked on the CPUs (I was chipsets), but I certainly have plenty of experience with unintended effects.
* That it trains on other data is a non issue. * That it allows privileged access as part of a prediction is a mild (at best) issue. * That it doesn't have a prediction fail recovery mechanism (e.g. zero the cache for the failed prediction) is a minor issue.
HOWEVER When these component issues combine together it is a *huge* flaw.
How can this happen? * designers are idiots? -no, if they were idiots they certainly wouldn't have a working part at all. This stuff is _complex_
* different smart designers worked on different parts, but in isolation? -ding! winner. It is highly likely that there was more than one team involved and as a result each validated their block works, and they validated that the system works. No one person saw the *whole* picture and as a result this vulnerability exists.
Now, once a designed block is done, the industry standard is to leave it the fsck alone! So this block of VHDL likely was re-used and tweaked for process changes, but never actually fully re-factored, so no one ever saw the big picture at all.
That is how bugs like this come to be.
If you want to beat people up about it, it's not the engineers that should be beaten, it is management that keeps the engineers under such schedule pressure that there is *never* a window to review and refactor something unless it's flatly broken. Beat up senior management for how they're handling this whole thing...
But leave the guys in the trenches out of it. I guarantee you they were doing their best.
For real. This has shown that these code monkeys know zero about computer architecture. This isn't a flaw in an implementation, this is a flaw in a fundamental principle of CPU design.
I would like to issue you two challenges. 1) design a CPU from scratch that is even 1/100th as powerful as a current gen CPU. 2) Starting with x86 PIII (where this issue first existed IIRC) design a multicore part even half as good as a modern CPU.
This. For quite a while I have been tasked with customer support for *one* customer and primary QA at my current gig.
"But it's only one customer!" -management "Yeah and when they interrupt the super complex test case I was running I have to very nearly start over." -me
6 months later I have an offer at a new company on the way, where I can focus on what I'm genuinely good at.
This is a double edged sword. If management would take it at face value and say "okay, this is when people are productive, so lets make that time easier for them" it'd be okay.
We all know management instead will say: "Shit! they're only 'working' 12.5 hours a week but we're paying 40?" ignoring that a lot of the not work time is meetings/overhead and the games/social media time is often a case of "This problem is thorny, I'm going to take a break and come back to it." and while on said break our brain is actually still background processing the task.
Nevermind that the USAF already has biofuels in test that are grown from mushrooms and JP-4 compatible. Not sure how cost effective they are right now, but cost effectiveness can change just based on scarcity of what it's replacing...
Because it's not terminating at both LAX and JFK, it's actually JFK to AUS (whatever the Melbourne code is) and does a Fuel and unload at LAX for the west coast destinations.
there is a technical issue here too... We switched source repositories from ClearCase to Perforce around 5 years ago... I think that while you're right that EOL shouldn't be an excuse, there may be serious issues to resurrecting the older codebase... hence why I suggested similar spec new CPUs and MBs at cost as an option.
There is very strong evidence that (for example) applications which do a lot of I/O, like databases, will have a measurable hit.
Actually we're past that, we're now at: There is empirical data that applications that do a lot of Kernel calls (such as disk I/O like databases) will see a large impact.
It affects much older CPUs, but anything older than 5 years is EOL and not supported.
I am one of the most Intel Fanboi's here (just look at my post history) and even I can't absolve any of this, other than I can accept the flaw is so old no one thought about it in modern CPUs... but now that it's discovered it needs to be handled like the FDIV bug: uC patch what can be patched with no or trivial (less than 2%) performance issue. If it can't be patched it needs to be replaced.
Not sure how to handle the EOL issue, but I think that an at cost Intel board and similar spec CPU would be one viable way. Since there is a > 50% margin (some are as high as 65%) on this stuff from a manufacturer standpoint this should equate to getting a replacement mobo and cpu of spec similar to a 5 year old machine for about $120, and it would still have better overall specs.
I agree with you. I think the best course of action is to refund the purchase, then offer either a small cash payout of good will (noting at the same time said fine print) or a larger credit payout towards future plays. In fact, knowing how lotteries work they could offer to pay out everyone at face value in credit and they'd likely regain virtually all of the spent funds.
not in this case. This is more akin to a scratcher, except it's generated at purchase instead of pre-gen'd and bundled. I'm pretty sure that the best outcome is a refund and possibly a token payout to all purchasers of error'd tickets.
unfortunately I *can't* get new batteries for my Nexus 5's only the cheapo knockoffs now. I buy the highest quality ones I can but it's still a crap-shoot.
As to the ability to change the battery, on the nexus 5 I can do a battery swap in 5 min, a screen swap in 38.
perhaps it's heated to simply melt the snow off? (unlikely, I know). Also, who said the lack of gutters was a design oversight? I think it just shows the normal contempt for users that Apple has always shown.
I get that you need party alignment to (effectively) mount a campaign. That said (and I am not a direct constituent as I'm almost as far away from you as possible while still being in the lower 48) how do you feel about the DNCs shenanigans in the last cycle?
It has been evident for quite some time now that our nominally 2 party system is really just an oligarchy of the elite and is no longer (for quite some while) a government by, of, and for the people. How do you propose to sow seeds of change?
I accept that you as one person, in office, would be unable to effect sweeping changes, any thoughts of how to effect small ones?
Will you make a vow upon your testicles (or any other suitable body part) to *NOT* be involved in any obviously scandalous affairs? Mind you I don't give two fucks about threesomes or dating guys, or whatever, what I want is a vow that you won't do something like: * go off about how being gay is so horrible, then get busted in gay sex * say how perfect your marriage is and that you'd never stray, then get busted having an affair.
In other words: Will you promise to say what you do and do what you say? If you say: "My marriage is sexless and my wife knows I have a mistress" then coolio, go have a mistress.
And, as phantomfive wrote about crime: Sounds great, what are the plans? How are you going to take on "Big prisons"?
And in closing: Cheers, I hope you win, and I hope that you can sow the seeds of a new government.
ooh, #3 and #4 are deal killers for me as well... #3: I'm a single dad, I plan my outings days in advance at least, this includes buying the tix. #4 I have kids (in light of above statement Duh!), no way we're all risking sitting apart from each other. #4a, perhaps you can just use multiple cards in one transaction? Given the other limitations, I'm not sure this would work though.
If you read his posts you'd know sever things about him: 1) he *assumes* everything is on the record and monitored at all levels while in country. 2) he has completely different kit for in and out of DPRK.
Given that context having his phone or any other device hacked doesn't actually impact him as he's operating on the assumption that such compromise has already happened.
I assume (no overt post saying so, but given his other posts it seems likely) that once back in Japan he dumps everything DPRK related into a box. and doesn't touch it again until the next trip back.
Slashdot Mirror
I'm guessing you don't work on this stuff do you...
I never worked on the CPUs (I was chipsets), but I certainly have plenty of experience with unintended effects.
* That it trains on other data is a non issue.
* That it allows privileged access as part of a prediction is a mild (at best) issue.
* That it doesn't have a prediction fail recovery mechanism (e.g. zero the cache for the failed prediction) is a minor issue.
HOWEVER
When these component issues combine together it is a *huge* flaw.
How can this happen?
* designers are idiots?
-no, if they were idiots they certainly wouldn't have a working part at all. This stuff is _complex_
* different smart designers worked on different parts, but in isolation?
-ding! winner. It is highly likely that there was more than one team involved and as a result each validated their block works, and they validated that the system works. No one person saw the *whole* picture and as a result this vulnerability exists.
Now, once a designed block is done, the industry standard is to leave it the fsck alone! So this block of VHDL likely was re-used and tweaked for process changes, but never actually fully re-factored, so no one ever saw the big picture at all.
That is how bugs like this come to be.
If you want to beat people up about it, it's not the engineers that should be beaten, it is management that keeps the engineers under such schedule pressure that there is *never* a window to review and refactor something unless it's flatly broken. Beat up senior management for how they're handling this whole thing...
But leave the guys in the trenches out of it. I guarantee you they were doing their best.
I hope they do. He's a disingenuous ass.
For real. This has shown that these code monkeys know zero about computer architecture. This isn't a flaw in an implementation, this is a flaw in a fundamental principle of CPU design.
I would like to issue you two challenges.
1) design a CPU from scratch that is even 1/100th as powerful as a current gen CPU.
2) Starting with x86 PIII (where this issue first existed IIRC) design a multicore part even half as good as a modern CPU.
and the smart slaves use out of band devices for comms, social media, etc. all while remaining logged in at their proscribed station.
This.
For quite a while I have been tasked with customer support for *one* customer and primary QA at my current gig.
"But it's only one customer!" -management
"Yeah and when they interrupt the super complex test case I was running I have to very nearly start over." -me
6 months later I have an offer at a new company on the way, where I can focus on what I'm genuinely good at.
This is a double edged sword.
If management would take it at face value and say "okay, this is when people are productive, so lets make that time easier for them" it'd be okay.
We all know management instead will say: "Shit! they're only 'working' 12.5 hours a week but we're paying 40?" ignoring that a lot of the not work time is meetings/overhead and the games/social media time is often a case of "This problem is thorny, I'm going to take a break and come back to it." and while on said break our brain is actually still background processing the task.
the policy talks about "All personal devices" not just phones.
So... Nope, ditch the watch too.
Water can't hold the amount of O2 required. Fluorinert is the "water" you're thinking of. No human (that I know of) has tried, but rodents have.
Nevermind that the USAF already has biofuels in test that are grown from mushrooms and JP-4 compatible. Not sure how cost effective they are right now, but cost effectiveness can change just based on scarcity of what it's replacing...
Because it's not terminating at both LAX and JFK, it's actually JFK to AUS (whatever the Melbourne code is) and does a Fuel and unload at LAX for the west coast destinations.
there is a technical issue here too...
We switched source repositories from ClearCase to Perforce around 5 years ago...
I think that while you're right that EOL shouldn't be an excuse, there may be serious issues to resurrecting the older codebase...
hence why I suggested similar spec new CPUs and MBs at cost as an option.
myself included...
Correct.
But as GP noted, this is likely for ease of code mgt on MS's part.
There is very strong evidence that (for example) applications which do a lot of I/O, like databases, will have a measurable hit.
Actually we're past that, we're now at:
There is empirical data that applications that do a lot of Kernel calls (such as disk I/O like databases) will see a large impact.
It affects much older CPUs, but anything older than 5 years is EOL and not supported.
I am one of the most Intel Fanboi's here (just look at my post history) and even I can't absolve any of this, other than I can accept the flaw is so old no one thought about it in modern CPUs... but now that it's discovered it needs to be handled like the FDIV bug: uC patch what can be patched with no or trivial (less than 2%) performance issue. If it can't be patched it needs to be replaced.
Not sure how to handle the EOL issue, but I think that an at cost Intel board and similar spec CPU would be one viable way. Since there is a > 50% margin (some are as high as 65%) on this stuff from a manufacturer standpoint this should equate to getting a replacement mobo and cpu of spec similar to a 5 year old machine for about $120, and it would still have better overall specs.
I agree with you.
I think the best course of action is to refund the purchase, then offer either a small cash payout of good will (noting at the same time said fine print) or a larger credit payout towards future plays.
In fact, knowing how lotteries work they could offer to pay out everyone at face value in credit and they'd likely regain virtually all of the spent funds.
not in this case. This is more akin to a scratcher, except it's generated at purchase instead of pre-gen'd and bundled.
I'm pretty sure that the best outcome is a refund and possibly a token payout to all purchasers of error'd tickets.
unfortunately I *can't* get new batteries for my Nexus 5's only the cheapo knockoffs now.
I buy the highest quality ones I can but it's still a crap-shoot.
As to the ability to change the battery, on the nexus 5 I can do a battery swap in 5 min, a screen swap in 38.
I'm rather fearful my pixel won't be so easy.
Well, you don't have my vote because I'm in the wrong state, but to quote Fry from Futureama:
Shut up and take my money!
I'll be donating what I can when I get home from work.
perhaps it's heated to simply melt the snow off?
(unlikely, I know).
Also, who said the lack of gutters was a design oversight? I think it just shows the normal contempt for users that Apple has always shown.
I get that you need party alignment to (effectively) mount a campaign.
That said (and I am not a direct constituent as I'm almost as far away from you as possible while still being in the lower 48) how do you feel about the DNCs shenanigans in the last cycle?
It has been evident for quite some time now that our nominally 2 party system is really just an oligarchy of the elite and is no longer (for quite some while) a government by, of, and for the people.
How do you propose to sow seeds of change?
I accept that you as one person, in office, would be unable to effect sweeping changes, any thoughts of how to effect small ones?
Will you make a vow upon your testicles (or any other suitable body part) to *NOT* be involved in any obviously scandalous affairs? Mind you I don't give two fucks about threesomes or dating guys, or whatever, what I want is a vow that you won't do something like:
* go off about how being gay is so horrible, then get busted in gay sex
* say how perfect your marriage is and that you'd never stray, then get busted having an affair.
In other words: Will you promise to say what you do and do what you say?
If you say: "My marriage is sexless and my wife knows I have a mistress" then coolio, go have a mistress.
And, as phantomfive wrote about crime: Sounds great, what are the plans?
How are you going to take on "Big prisons"?
And in closing:
Cheers, I hope you win, and I hope that you can sow the seeds of a new government.
he's recursively dismayed at misspelling "arrrrr!"
ooh, #3 and #4 are deal killers for me as well...
#3: I'm a single dad, I plan my outings days in advance at least, this includes buying the tix.
#4 I have kids (in light of above statement Duh!), no way we're all risking sitting apart from each other.
#4a, perhaps you can just use multiple cards in one transaction? Given the other limitations, I'm not sure this would work though.
If the theaters like AMC don't support it, how am I supposed to go to the movies?
I would totally be in if this becomes as easy as Netflix though.
If you read his posts you'd know sever things about him:
1) he *assumes* everything is on the record and monitored at all levels while in country.
2) he has completely different kit for in and out of DPRK.
Given that context having his phone or any other device hacked doesn't actually impact him as he's operating on the assumption that such compromise has already happened.
I assume (no overt post saying so, but given his other posts it seems likely) that once back in Japan he dumps everything DPRK related into a box. and doesn't touch it again until the next trip back.