Intel compiler suite knows both how to vectorize (for inner loops) and parallelize (for outer loops). What you are referring to here with OpenMP is parallelizing, which is great.
But you will still get 4x+ (depending on the CPU) of performance overall if the inner loops are vectorized. If it means you can get the same performance with 4x less hardware, that can make a huge difference to the cost and viability of a project.
I tested this extensively, including quite recently, and unfortunately where intel's C compiler as far back as 2008 produced excellent results (all the loops in my code vectorized even back then), GCC to this day fails to vectorize most of them. last time I tried with clang (last year), it fared even worse than GCC.
The big problem when it comes to using anything other than Fortran, C or C++ is that 20 years after the first MMX and SSE instruction sets have been added to CPUs, there are only a handful of compilers that known how to vectorize even loops that are hand-crafted to be vectorizable - and the ones that can do it are all commercially licenced (GCC might theoretically have some support for it, but in reality it doesn't vectorize most things).
And since most of the performance advancement in silicon has for a long time now been focused in SIMD units, that means that for any performance sensitive workload there are no feasible alternatives. If it has taken GCC 20 years to get not very far, how long will be be before much younger compilers get anywhere with this performance critical feature?
"Flexera points out that as much as 50 percent of code in commercial and IoT software products is open source."
And most of those products are violating the GPL.
"Even with the move toward more agile development and DevOps, vulnerabilities continue to take off..."
Agile (and possibly to a much lesser extent DevOps) are a part of the problem, much more than a part of the solution, even when they are implemented properly. The key problem with Agile in security terms is that the concept is entirely based around the lack of up-front design. While we have to accept that up-front design isn't always possible (how many clients today are capable of formulating exactly what they want their software to do?), the lack of up front design does away with the ideal stage of the development process at which to anticipate, highlight and design around the most obvious of security vulnerabilities.
While sticking to good coding practices is necessary condition to achieve reasonably exploit-free software, it is not a sufficient condition, and without up-front understanding of the big picture, it is impossible to have good awareness of where integration weakness might lead to an exploitable flaw.
There is a very good reason why industries in which reliability is paramount (aerospace, automotive, medical) do not (and never will) use agile software development methodologies. Sole purpose of agile is to facilitate delivering a lot of features based on unknowable requirements quickly. For some environments, this is exactly what you need; for others it is suicidal.
Material loss: Insurance (read the small print and make sure you get a policy that covers it).
Privacy loss: Encrypt the disk. Commercial and open source full disk encryption software is easily available.
Data loss: Backups. Plenty of affordable online backup/storage solutions are available, some specialising in specifically in backups (Backblaze, Crashplan), others that are more generic (Amazon Drive, Google Drive).
The problem is that Studios have been pissing people off and running their business based on very incorrect metrics. They talk about a drop in DVD sales, yet they make a decision to cancel a show before a season is even finished, sometimes before it even started airing, let alone before DVD box set of that season is available.
They make decisions to cancel based on "viewing figured" what are only based on live viewing, so DVD sales never feature into whether a show gets cancelled or not. The only legitimate way for people not in the same country as the studio to view the show in the years showrly after airing in the US is to by the DVD box set, yet all those millions of sales don't get counted.
Until they start looking at the profitability of the global sales they will keep falling foul of their own broken business models. This is why the like of Amazon and Netflix are winning over them, because their metrics for assessing the viability of their business model isn't fundamentally broken.
This already exists. It is called "dispersed storage", and is developed by a company called Cleversafe (which has since, IIRC, been bought by IBM.
Version 1.x was open source and you can still get the source from here:
https://fossies.org/linux/misc...
(the dsnet-* archives).
This is always considered controversial, but it's totally true! Good programming depends 99% on the skill of the programmer. A bad programmer will write bad programs in ANY language. Granted, some languages encourage good programming more than others, but it is all ultimately down to the skill of the developer.
Where there is no police, there is no speed limit. Any driver that tells you they never speed is lying to you, regardless of whether their speeding is intentional or accidental.
Speeding is a problem for cyclists and pedestrians who have to share the roads with impatient maniacs.
So, how do you justify these on motorways where cyclists and pedestrians are not allowed?
Compiler for what? JavaScript is interpreted, so no compilation is necessary.If you are writing an app 100% in JavaScript, you might want to "compile" your app into a single page containing all the CSS, JavaScript and HTML, but I wouldn't exactly call that compiling.
Mod parent up!
You cannot offer insurance against prosecution for doing something illegal. You can offer insurance that covers _investigation_ costs, but not to cover the fine or the impact of any sanctions if the claimant is conviced.
Their claim that the material "has been tested and documented by several prestigious institutions, laboratories and universities" is as laughable as it is vague.
Now, now, I'm sure it was one of those "prestigious non-accredited universities" that we keep getting emails about.:-)
The man has a point. Putting all additional tax on fuel would have multiple benefits:
1) No additional hardware to put into cars, no doubt at the vehicle owner's expense.
2) Fuel-efficient cars are encouraged while gas guzzling SUVs are discouraged.
3) Avoiding tax on fuel is much more difficult than hacking a GPS tracking device.
Bottom line - it would keep things simple. Alternatives would totally needlessly complicate things.
IMNSHO, the bunch of mugs who are willing to regularly work 80+ hour weeks for no overtime deserve everything they get. Others can only abuse you as much as you let them. If you are going to let them, then STFU.
Slashdot Mirror
Intel compiler suite knows both how to vectorize (for inner loops) and parallelize (for outer loops). What you are referring to here with OpenMP is parallelizing, which is great. But you will still get 4x+ (depending on the CPU) of performance overall if the inner loops are vectorized. If it means you can get the same performance with 4x less hardware, that can make a huge difference to the cost and viability of a project.
I tested this extensively, including quite recently, and unfortunately where intel's C compiler as far back as 2008 produced excellent results (all the loops in my code vectorized even back then), GCC to this day fails to vectorize most of them. last time I tried with clang (last year), it fared even worse than GCC.
The big problem when it comes to using anything other than Fortran, C or C++ is that 20 years after the first MMX and SSE instruction sets have been added to CPUs, there are only a handful of compilers that known how to vectorize even loops that are hand-crafted to be vectorizable - and the ones that can do it are all commercially licenced (GCC might theoretically have some support for it, but in reality it doesn't vectorize most things). And since most of the performance advancement in silicon has for a long time now been focused in SIMD units, that means that for any performance sensitive workload there are no feasible alternatives. If it has taken GCC 20 years to get not very far, how long will be be before much younger compilers get anywhere with this performance critical feature?
"Flexera points out that as much as 50 percent of code in commercial and IoT software products is open source." And most of those products are violating the GPL.
"Even with the move toward more agile development and DevOps, vulnerabilities continue to take off..."
Agile (and possibly to a much lesser extent DevOps) are a part of the problem, much more than a part of the solution, even when they are implemented properly. The key problem with Agile in security terms is that the concept is entirely based around the lack of up-front design. While we have to accept that up-front design isn't always possible (how many clients today are capable of formulating exactly what they want their software to do?), the lack of up front design does away with the ideal stage of the development process at which to anticipate, highlight and design around the most obvious of security vulnerabilities.
While sticking to good coding practices is necessary condition to achieve reasonably exploit-free software, it is not a sufficient condition, and without up-front understanding of the big picture, it is impossible to have good awareness of where integration weakness might lead to an exploitable flaw.
There is a very good reason why industries in which reliability is paramount (aerospace, automotive, medical) do not (and never will) use agile software development methodologies. Sole purpose of agile is to facilitate delivering a lot of features based on unknowable requirements quickly. For some environments, this is exactly what you need; for others it is suicidal.
Material loss: Insurance (read the small print and make sure you get a policy that covers it).
Privacy loss: Encrypt the disk. Commercial and open source full disk encryption software is easily available.
Data loss: Backups. Plenty of affordable online backup/storage solutions are available, some specialising in specifically in backups (Backblaze, Crashplan), others that are more generic (Amazon Drive, Google Drive).
The problem is that Studios have been pissing people off and running their business based on very incorrect metrics. They talk about a drop in DVD sales, yet they make a decision to cancel a show before a season is even finished, sometimes before it even started airing, let alone before DVD box set of that season is available.
They make decisions to cancel based on "viewing figured" what are only based on live viewing, so DVD sales never feature into whether a show gets cancelled or not. The only legitimate way for people not in the same country as the studio to view the show in the years showrly after airing in the US is to by the DVD box set, yet all those millions of sales don't get counted.
Until they start looking at the profitability of the global sales they will keep falling foul of their own broken business models. This is why the like of Amazon and Netflix are winning over them, because their metrics for assessing the viability of their business model isn't fundamentally broken.
This already exists. It is called "dispersed storage", and is developed by a company called Cleversafe (which has since, IIRC, been bought by IBM. Version 1.x was open source and you can still get the source from here: https://fossies.org/linux/misc... (the dsnet-* archives).
In which case you are saying that taxation is stealing. You may have a point, there, mind you.
He'd be SOOO much less boned if he only owed them $2.8bn...
If it were the other way around - in UK you can file small claims court claims online. :-)
This is always considered controversial, but it's totally true! Good programming depends 99% on the skill of the programmer. A bad programmer will write bad programs in ANY language. Granted, some languages encourage good programming more than others, but it is all ultimately down to the skill of the developer.
Where there is no police, there is no speed limit. Any driver that tells you they never speed is lying to you, regardless of whether their speeding is intentional or accidental.
Dude! I like the way you're thinking! :-D
Speeding is a problem for cyclists and pedestrians who have to share the roads with impatient maniacs. So, how do you justify these on motorways where cyclists and pedestrians are not allowed?
Dude, you've been watching Urotsukidoji too much.
Compiler for what? JavaScript is interpreted, so no compilation is necessary.If you are writing an app 100% in JavaScript, you might want to "compile" your app into a single page containing all the CSS, JavaScript and HTML, but I wouldn't exactly call that compiling.
... idle sensationalism to me... But then again, I am known to be cynical...
Mod parent up! You cannot offer insurance against prosecution for doing something illegal. You can offer insurance that covers _investigation_ costs, but not to cover the fine or the impact of any sanctions if the claimant is conviced.
Those sound like the words of someone who speaks from experience. :-)
What gets me is why aren't the ISPs transparently proxying SMTP...
Their claim that the material "has been tested and documented by several prestigious institutions, laboratories and universities" is as laughable as it is vague.
Now, now, I'm sure it was one of those "prestigious non-accredited universities" that we keep getting emails about.
I'm just waiting for the day when God turns up and claims he has prior art to the patented gene.
Who cares where the drugs are made? All the "generic viagra" spam proves that this is already done on a large scale. As it should be!
The man has a point. Putting all additional tax on fuel would have multiple benefits: 1) No additional hardware to put into cars, no doubt at the vehicle owner's expense. 2) Fuel-efficient cars are encouraged while gas guzzling SUVs are discouraged. 3) Avoiding tax on fuel is much more difficult than hacking a GPS tracking device. Bottom line - it would keep things simple. Alternatives would totally needlessly complicate things.
IMNSHO, the bunch of mugs who are willing to regularly work 80+ hour weeks for no overtime deserve everything they get. Others can only abuse you as much as you let them. If you are going to let them, then STFU.