I will stay away from any Zeppelin-Die-based CPUs until AMD provides a proper explanation and fix for the "gcc segmentation fault" bug that haunts Ryzen CPUs.
Umm... they already did both of those.
No, they did neither. Please provide a link to that proper explanation and description of the fix you say exists.
What AMD currently does is rather "have some poor guy test N CPUs to find some non-buggy exemplars amongst them and ship those to RMA demanding customers". And as one can see from the reports of affected people, even recently delivered RMA-exchange CPUs (manufactured in CW 30) were still affected:
https://docs.google.com/spread...
Their production going forward has solved the problem.
Can you provide a link to any credible source stating this - or the nature of the change to the production process? And if this was true, then why has AMD returned still-defective CPUs even to people who RMA'd because of exactly this bug in the last few weeks?
They aren't doing a recall because it would bankrupt them.
That does seem plausible, but personally I don't want to be involved in an RMA-until-you-get-lucky game.
Phronix reported that AMD has debugged the problem and that all CPUs (1600-1800x for example) manufactured after week 30 (30 July) will not have this problem.
Can you provide a link? I am a regular Phoronix reader, but did not spot such a report. (And AMD returned CPUs manufactured earlier to some that RMA'd, and some of these _still_ had the defect.)
As much as I appreciate the price/performance of Ryzen based CPUs and the competition it has sparked, I will stay away from any Zeppelin-Die-based CPUs until AMD provides a proper explanation and fix for the "gcc segmentation fault" bug that haunts Ryzen CPUs.
After many months of not admitting any bug existence to the affected users, AMD finally admitted there is one, yet they neither recalled affected CPUs nor do they tell us how to distinguish affected from unaffected CPUs - so even if you buy a Ryzen today, you can still buy one unsuitable for ordinary gcc compilation tasks.
Given that they cannot (or do not want to) say which CPUs are affected, and given that AMD did never explain a root cause of this bug and how it is fixed, I do not believe they actually have fixed it or know how to fix it. Even CPUs that were manufactured in calender week 25 of 2017 have turned out to be affected!
Why should I believe they fixed this for "Threadripper"? Sure, they know by now how to test individual CPU exemplars for the bug, and might deliver unaffected ones to the press for reviews. Does that tell me they will do the same testing for the exemplars delivered to the mass market? No.
Currently, even most of the large-budget block busters are still either shot on cameras with less than 4k resolution, or post-produced in 2k. And talking about streaming, movies are streamed at bandwidths so ricidculously low that every complex scenery turns into mush.
So why pay extra for a 4k upscaling that your player or TV can do just as well?
I for one am willing to pay for actual, high-bandwidth 4k content - of which very, very little is on offer.
- Had an understanding of the root cause of the bug, and told the public what it was and how they solved it
- Told everyone how to distinguish affected CPUs from unaffected CPUs (without a multi-hour run of some test-script that not AMD, but desparate affected buyers implemented and made available)
- Recalled all the defective CPUs and replaced them with working ones, including CPUs sold as part of computers
What you describe is rather the bare minimum they owe customers going through lots of trouble due to a defect product they were sold.
AMD has never explained (and might not even know) the root cause of this bug. It is merely hear-say that newer CPUs are probably not affected.
So yes, if you buy a Ryzen now you could still get one that is affected. Better make sure you run the test scripts for this bug for multiple hours before assuming you got a flawless CPU.
AMD in many countries sends people who are affected by this bug off to their computer dealer if they did not buy a boxed retail CPU, but bought a whole computer with the Ryzen CPU as part of it.
Which basically means they are screwed: Computer dealers totally lack the knowledge or incentive to even understand the nature of this bug, especially given the fact that AMD has never explained which of their CPUs are affected and which are not.
So at best, your computer dealer will give you another computer in return, with a CPU in it that likely has the same defect. At worst, the computer dealer will just tell you that "Linux isn't supported" because many press articles wrongly claimed that this bug occurs only when using Linux.
Two of my friends are currently going through this service nightmare, and I have postponed my plans for Ryzen/Threadripper/Epyc computers until I know how this will turn out for them.
What I certainly won't do is becoming passenger in a car that is not really self-driving, but totally dependent on permanent online transmissions, hyper-accurate maps (which of course will never really be up-to-date) and other weird external influences.
If the industry can sell me a car that it really autonomous, as in: "Has sensors to look around and that is all it needs", then I'm buying.
Oh, and sure, I totally don't want a car that any half-competent script-kid-hacker can manipulate to drive into the next tree - which is equivalent to "I don't want a car that is "online", ever.
"I've never gone in debt" is saying "I've never attempted a project/investment larger than I could support at the time."
And permanently being in debt is saying "My attempts into projects/investments I could not support when I started them were unsuccessful, thus I spend the rest of my life working for the profit of my creditors."
It always amazes me how used US citizens are to the idea that being in debt is something totally normal - even if it is debt of the government lended from foreign countries.
I for one like to work for money I can spend on goods and services, rather than paying interest.
Just out of curiousity I searched for a (not at all famous) paper I published in the 1990s. Found it on Sci-Hub by just entering a few key words in a search engine - conveniently retrievable.
I remembered that at some point in time, a state-funded institution did officially archive my paper. But it took me about half an hour to finally find it, buried behind multiple retrieval forms and links, with no chance to find it had I not looked up its entire, exact title before.
No question, Sci-Hub did the better job of keeping my little contribution to the world's knowledge available to the public.
I'm so glad I can honestly tell that billion people I do not want to communicate with: "Sorry, but I am not on Facebook, and certainly not using their messenger".
Facebook should really separate completely from the open Internet and keep its users in a walled, electro-fenced garden. The Internet is just a better place with Facebook users staying amongst their kind.
Every 65mm or 70mm analog film I have ever seen was in terms of resolution and signal/noise ratio _way_ below the quality of modern digital cameras, even below the decade old "3.4k" Arri cameras.
I wonder what others pretend to have seen, but movies like "The hateful eight", if anything, proved that analog film is outdated and bad. The whole indoor lighting they had to apply in "the hateful eight" to make the analog film catch a usable picture was terrible, it looked completely unrealistic (I mean seriously: A dark hut in a snow-storm, and magically bright lights shine from the ceiling all the time... WTF?!?)
I do not know whether "Dunkirk" is a movie worth watching, but I am absolutely sure if it is, it is certainly not because of image quality. And sure as hell it will hardly stretch the limits of a Bluray disc in terms of quality - certainly no reason to buy an UHD Bluray of it.
In Germany, there are not just "minor" and "adults" before the law, but fine-grained differences are made depending on age, see https://de.wikipedia.org/wiki/... for details.
Even childred at the age of 7 have (very limited) rights to do business transactions on their own, and a person at the age of 15 already has a lot of rights.
And just because parents have a tendency to just not believe in mundane causes of death, such as traffic accidents, does not mean the police is incapable of determining whether something was an accident or suicide - especially, when like in this case, lots of witnesses were around and simply not a single indication to suicide was present.
When your child dies, then the living people who communicated with your child retain their right of communication privacy, and your curiousity as a parent does not quite outweigh that right.
Imagine how pissed you would be if your girl-friend is run over by a truck and immediately afterwards, her parents are starting to read (and possibly circulate) all the juicy details of your prior conversations.
Such would not be necessary: The police already did officially investigate the death of that girl, and the police would have had the possibility to look into her Facebook based communication-
The result of the police investigation, however, was that her death was an accident - a result that her parents obviously are not willing to accept.
and isn't a virtual home all they every wanted? They are working in virtual jobs, spend their time with virtual friends - so why bother with reality at all?
Like over 90% of the earths population I, too, live in a country that signed the SI treaty - and the "nutritional information" printed on food here is obliged to present the food energy in Joule. (Many still also print "calorie" values in addition, but that is optional.)
The most bizarre thing about "calories" is that most people who talk about them use numbers that are off by a factor of 1000 - just as if people were talking about "bytes" while in fact they mean kilobytes.
You could always specify a connector using a multi-mode optical fiber. As you can see from the almost zero prices of (even longer) TOSlink cables, such cables are very cheap - and there almost no limits to the data rates possible on optical fibers.
The problem with Thunderbolt is that they wanted to make the ports cheap and burden the price X times on the buyer of the cables, later. Won't fly.
Slashdot Mirror
I will stay away from any Zeppelin-Die-based CPUs until AMD provides a proper explanation and fix for the "gcc segmentation fault" bug that haunts Ryzen CPUs.
Umm... they already did both of those.
No, they did neither. Please provide a link to that proper explanation and description of the fix you say exists.
What AMD currently does is rather "have some poor guy test N CPUs to find some non-buggy exemplars amongst them and ship those to RMA demanding customers". And as one can see from the reports of affected people, even recently delivered RMA-exchange CPUs (manufactured in CW 30) were still affected: https://docs.google.com/spread...
Their production going forward has solved the problem.
Can you provide a link to any credible source stating this - or the nature of the change to the production process?
And if this was true, then why has AMD returned still-defective CPUs even to people who RMA'd because of exactly this bug in the last few weeks?
They aren't doing a recall because it would bankrupt them.
That does seem plausible, but personally I don't want to be involved in an RMA-until-you-get-lucky game.
Phronix reported that AMD has debugged the problem and that all CPUs (1600-1800x for example) manufactured after week 30 (30 July) will not have this problem.
Can you provide a link? I am a regular Phoronix reader, but did not spot such a report. (And AMD returned CPUs manufactured earlier to some that RMA'd, and some of these _still_ had the defect.)
As much as I appreciate the price/performance of Ryzen based CPUs and the competition it has sparked, I will stay away from any Zeppelin-Die-based CPUs until AMD provides a proper explanation and fix for the "gcc segmentation fault" bug that haunts Ryzen CPUs.
After many months of not admitting any bug existence to the affected users, AMD finally admitted there is one, yet they neither recalled affected CPUs nor do they tell us how to distinguish affected from unaffected CPUs - so even if you buy a Ryzen today, you can still buy one unsuitable for ordinary gcc compilation tasks.
Given that they cannot (or do not want to) say which CPUs are affected, and given that AMD did never explain a root cause of this bug and how it is fixed, I do not believe they actually have fixed it or know how to fix it. Even CPUs that were manufactured in calender week 25 of 2017 have turned out to be affected!
Why should I believe they fixed this for "Threadripper"?
Sure, they know by now how to test individual CPU exemplars for the bug, and might deliver unaffected ones to the press for reviews. Does that tell me they will do the same testing for the exemplars delivered to the mass market? No.
More information on this bug via https://forum.level1techs.com/... and https://community.amd.com/thre...
Currently, even most of the large-budget block busters are still either shot on cameras with less than 4k resolution, or post-produced in 2k. And talking about streaming, movies are streamed at bandwidths so ricidculously low that every complex scenery turns into mush.
So why pay extra for a 4k upscaling that your player or TV can do just as well?
I for one am willing to pay for actual, high-bandwidth 4k content - of which very, very little is on offer.
"Bloody stellar" I would call if they:
- Had an understanding of the root cause of the bug, and told the public what it was and how they solved it
- Told everyone how to distinguish affected CPUs from unaffected CPUs (without a multi-hour run of some test-script that not AMD, but desparate affected buyers implemented and made available)
- Recalled all the defective CPUs and replaced them with working ones, including CPUs sold as part of computers
What you describe is rather the bare minimum they owe customers going through lots of trouble due to a defect product they were sold.
AMD has never explained (and might not even know) the root cause of this bug. It is merely hear-say that newer CPUs are probably not affected.
So yes, if you buy a Ryzen now you could still get one that is affected. Better make sure you run the test scripts for this bug for multiple hours before assuming you got a flawless CPU.
AMD in many countries sends people who are affected by this bug off to their computer dealer if they did not buy a boxed retail CPU, but bought a whole computer with the Ryzen CPU as part of it.
Which basically means they are screwed: Computer dealers totally lack the knowledge or incentive to even understand the nature of this bug, especially given the fact that AMD has never explained which of their CPUs are affected and which are not.
So at best, your computer dealer will give you another computer in return, with a CPU in it that likely has the same defect. At worst, the computer dealer will just tell you that "Linux isn't supported" because many press articles wrongly claimed that this bug occurs only when using Linux.
Two of my friends are currently going through this service nightmare, and I have postponed my plans for Ryzen/Threadripper/Epyc computers until I know how this will turn out for them.
What I certainly won't do is becoming passenger in a car that is not really self-driving, but totally dependent on permanent online transmissions, hyper-accurate maps (which of course will never really be up-to-date) and other weird external influences.
If the industry can sell me a car that it really autonomous, as in: "Has sensors to look around and that is all it needs", then I'm buying.
Oh, and sure, I totally don't want a car that any half-competent script-kid-hacker can manipulate to drive into the next tree - which is equivalent to "I don't want a car that is "online", ever.
... to let the responsible VW executives breathe the lovely fumes from their oh-so-clean Diesels. I'm sure they could modify some VW Diesel for this.
Even nedit still works well (if you do not fancy UTF-8, like slashdot) :-)
"I've never gone in debt" is saying "I've never attempted a project/investment larger than I could support at the time."
And permanently being in debt is saying "My attempts into projects/investments I could not support when I started them were unsuccessful, thus I spend the rest of my life working for the profit of my creditors."
It always amazes me how used US citizens are to the idea that being in debt is something totally normal - even if it is debt of the government lended from foreign countries.
I for one like to work for money I can spend on goods and services, rather than paying interest.
I for one will wait for the discount prices to be expected at the end of the current mining hype.
Lots of AMD GPUs will then be sold on the second-hand market, also putting pressure on the prices of new GPUs.
Just out of curiousity I searched for a (not at all famous) paper I published in the 1990s. Found it on Sci-Hub by just entering a few key words in a search engine - conveniently retrievable.
I remembered that at some point in time, a state-funded institution did officially archive my paper. But it took me about half an hour to finally find it, buried behind multiple retrieval forms and links, with no chance to find it had I not looked up its entire, exact title before.
No question, Sci-Hub did the better job of keeping my little contribution to the world's knowledge available to the public.
I'm so glad I can honestly tell that billion people I do not want to communicate with: "Sorry, but I am not on Facebook, and certainly not using their messenger".
Facebook should really separate completely from the open Internet and keep its users in a walled, electro-fenced garden. The Internet is just a better place with Facebook users staying amongst their kind.
Every 65mm or 70mm analog film I have ever seen was in terms of resolution and signal/noise ratio _way_ below the quality of modern digital cameras, even below the decade old "3.4k" Arri cameras.
I wonder what others pretend to have seen, but movies like "The hateful eight", if anything, proved that analog film is outdated and bad. The whole indoor lighting they had to apply in "the hateful eight" to make the analog film catch a usable picture was terrible, it looked completely unrealistic (I mean seriously: A dark hut in a snow-storm, and magically bright lights shine from the ceiling all the time... WTF?!?)
I do not know whether "Dunkirk" is a movie worth watching, but I am absolutely sure if it is, it is certainly not because of image quality. And sure as hell it will hardly stretch the limits of a Bluray disc in terms of quality - certainly no reason to buy an UHD Bluray of it.
Industry favors "cheap and docile" over "expensive and of opinion".
In Germany, there are not just "minor" and "adults" before the law, but fine-grained differences are made depending on age, see https://de.wikipedia.org/wiki/... for details.
Even childred at the age of 7 have (very limited) rights to do business transactions on their own, and a person at the age of 15 already has a lot of rights.
And just because parents have a tendency to just not believe in mundane causes of death, such as traffic accidents, does not mean the police is incapable of determining whether something was an accident or suicide - especially, when like in this case, lots of witnesses were around and simply not a single indication to suicide was present.
When your child dies, then the living people who communicated with your child retain their right of communication privacy, and your curiousity as a parent does not quite outweigh that right.
Imagine how pissed you would be if your girl-friend is run over by a truck and immediately afterwards, her parents are starting to read (and possibly circulate) all the juicy details of your prior conversations.
Such would not be necessary: The police already did officially investigate the death of that girl, and the police would have had the possibility to look into her Facebook based communication-
The result of the police investigation, however, was that her death was an accident - a result that her parents obviously are not willing to accept.
and isn't a virtual home all they every wanted? They are working in virtual jobs, spend their time with virtual friends - so why bother with reality at all?
Like over 90% of the earths population I, too, live in a country that signed the SI treaty - and the "nutritional information" printed on food here is obliged to present the food energy in Joule. (Many still also print "calorie" values in addition, but that is optional.)
The most bizarre thing about "calories" is that most people who talk about them use numbers that are off by a factor of 1000 - just as if people were talking about "bytes" while in fact they mean kilobytes.
You could always specify a connector using a multi-mode optical fiber. As you can see from the almost zero prices of (even longer) TOSlink cables, such cables are very cheap - and there almost no limits to the data rates possible on optical fibers.
The problem with Thunderbolt is that they wanted to make the ports cheap and burden the price X times on the buyer of the cables, later. Won't fly.
or at least very obsolete, given that the official SI unit for food energy is the Joule.