EPTThe Big Programming Thread - Page 747
| Forum Index > General Forum |
Thread Rules 1. This is not a "do my homework for me" thread. If you have specific questions, ask, but don't post an assignment or homework problem and expect an exact solution. 2. No recruiting for your cockamamie projects (you won't replace facebook with 3 dudes you found on the internet and $20) 3. If you can't articulate why a language is bad, don't start slinging shit about it. Just remember that nothing is worse than making CSS IE6 compatible. 4. Use [code] tags to format code blocks. | ||
|
Djagulingu
Germany3605 Posts
| ||
|
Manit0u
Poland17581 Posts
On July 30 2016 06:11 Djagulingu wrote: Anybody here has experience with OAuth2? Can anyone suggest me resources for that shit? Isn't that sufficient? There are ready-made libraries for that. http://oauth.net/2/ Do you need help with anything specific pertaining to this? | ||
|
Mr. Wiggles
Canada5894 Posts
I'm currently working on a high availability virtualized storage appliance, for example. Most of my work's done in C, C++, and Python. https://en.m.wikipedia.org/wiki/System_programming | ||
|
Nesserev
Belgium2760 Posts
| ||
|
Djagulingu
Germany3605 Posts
On July 30 2016 18:17 Manit0u wrote: Isn't that sufficient? There are ready-made libraries for that. http://oauth.net/2/ Do you need help with anything specific pertaining to this? Didn't even know that this thing existed. One more question to the vast amount of good programmers in TL: (How) can I build https without having a domain name first? For testing purposes, that is. | ||
|
Deleted User 101379
4849 Posts
On August 01 2016 14:59 Djagulingu wrote: Didn't even know that this thing existed. One more question to the vast amount of good programmers in TL: (How) can I build https without having a domain name first? For testing purposes, that is. Basically, you just generate a self-signed certificate and then you can use https. The browser will still complain because it's not a trusted certificate, but you can't change that without spending money for an actual signed certificate. https://www.sslshopper.com/article-how-to-create-and-install-an-apache-self-signed-certificate.html | ||
|
Nesserev
Belgium2760 Posts
| ||
|
Shield
Bulgaria4824 Posts
| ||
|
RoomOfMush
1296 Posts
On August 02 2016 04:08 Shield wrote: So, when processors get faster and GPUs get better, do you think C++ will be relevant in the next decades as well? I think C++11 and 14 have been useful. There is C++17 soon as well. Processors are not going to get much faster anytime soon. They reached physical limits some time ago. The problem with making processors faster is that the amount of electricity you need grows enormously and the heat it generates makes it very difficult to cool. Unless they invent some new kind of materials or completely change the way processors work they are not going to get much faster. That is the reason why we are heading more and more towards multi-core processors. Running 4 of them in parallel is much cheaper than running 1 which is 4 times faster. But regardless of that, I dont see why C++ would not be relevant anymore. What I could imagine is a new programming language with strong build in mutli-threading and synchronization features which make working with multi-core processors easier and less error prone. But I cant really imagine how that would look like at the moment. | ||
|
bangsholt
Denmark138 Posts
But predicting the future is always hard  | ||
|
Khalum
Austria831 Posts
On August 02 2016 04:56 bangsholt wrote: Except that those strong multithreading and synchronization features already exists in the sense of "just" using the Actor model (or message queues) so the need for a language to capitalize on this seems rather unneeded. But predicting the future is always hard Let me introduce you to Rust! | ||
tofucake
Hyrule19188 Posts
On August 02 2016 04:46 RoomOfMush wrote: Processors are not going to get much faster anytime soon. They reached physical limits some time ago. The problem with making processors faster is that the amount of electricity you need grows enormously and the heat it generates makes it very difficult to cool. Unless they invent some new kind of materials or completely change the way processors work they are not going to get much faster. most of this is not true | ||
|
Shield
Bulgaria4824 Posts
Edit: They actually plan to do 5 nm as well. https://en.wikipedia.org/wiki/5_nanometer | ||
|
RoomOfMush
1296 Posts
Care to write a little more than that? | ||
|
tofucake
Hyrule19188 Posts
They haven't reached their physical limits at all. Processors today are made with 10 nm dies, and the past couple years with 14nm. Current methods have allowed manufacturers to shrink the process down even further to 7nm, and they are working on making 5nm economical (prototypes already exist). There have also been experimental transistors made from graphene that are only a few atoms thick, and due to being graphene and not doped silicon, this gets around some of the size issues. Combine that with other research in super conductors (now possible at -70C, when before they were restricted to the -270C range) and quantum states (yay qubits) means that Moore's Law will continue, and we will see faster and faster processors for a long while yet. And as for "more fasterness = more hotyness", that's wrong. Intel and AMD both operate on tick/tock releases. A tick year is one where processors gain more speed, and a tock year is one where processors are made more efficient and use less power. Take a look at the Intel Pentium G4400T, which runs at 1.5 volts and 35 watts, and clocks 2.9 GHz over 2 cores with 2 threads each, and is 64 bit, and is a 14 nm cast. Now compare that to the single core, single thread 32 bit Pentium 4, which draws 1.6 to 1.7 volts at 51.6 watts for 1.3 GHz and is a 180 nm cast. Small = less energy = less heat, plus all the advances in cooling makes for some awesome possibilities. The 4400T is vastly superior in speed, processing power, and energy consumption for the exact reasons you say it should be worse. | ||
|
RoomOfMush
1296 Posts
On August 02 2016 09:25 tofucake wrote: They haven't reached their physical limits at all. Processors today are made with 10 nm dies, and the past couple years with 14nm. Current methods have allowed manufacturers to shrink the process down even further to 7nm, and they are working on making 5nm economical (prototypes already exist). There have also been experimental transistors made from graphene that are only a few atoms thick, and due to being graphene and not doped silicon, this gets around some of the size issues. Combine that with other research in super conductors (now possible at -70C, when before they were restricted to the -270C range) and quantum states (yay qubits) means that Moore's Law will continue, and we will see faster and faster processors for a long while yet. And as for "more fasterness = more hotyness", that's wrong. Intel and AMD both operate on tick/tock releases. A tick year is one where processors gain more speed, and a tock year is one where processors are made more efficient and use less power. Take a look at the Intel Pentium G4400T, which runs at 1.5 volts and 35 watts, and clocks 2.9 GHz over 2 cores with 2 threads each, and is 64 bit, and is a 14 nm cast. Now compare that to the single core, single thread 32 bit Pentium 4, which draws 1.6 to 1.7 volts at 51.6 watts for 1.3 GHz and is a 180 nm cast. Small = less energy = less heat, plus all the advances in cooling makes for some awesome possibilities. The 4400T is vastly superior in speed, processing power, and energy consumption for the exact reasons you say it should be worse. Although it is true that they manage to build smaller transistors, that does not really lead to faster clock speeds. The size of CPU's is roughly the same over the years. This means smaller transistors => more transistors. The power consumption you safe from having smaller transistors is equialized by having more of them. The clock speeds are more or less stagnant over the last couple of years. You rarely see CPU's above 3.8GHz and hardly any above 4.6GHz. And it is unlikely it will go higher than that anytime soon. Building faster CPU's is just not energy efficient anymore. Although I dont have any papers on hand, there are plenty of blogs which can be found all over the web, for example: https://www.technologyreview.com/s/421186/why-cpus-arent-getting-any-faster/ https://www.comsol.com/blogs/havent-cpu-clock-speeds-increased-last-years/ And yes, clock speed is not the only important factor, but its a factor with a big impact and it is easy to measure. I know there is lots of other things that can be done to improve the "speed" of the CPU by reducing the number of cycles needed for certain actions or improving the branch prediction and what not, but those things have limits too and every little improvement comes with diminishing returns. (By the way, everything I said is about the good old standard CPU's. There is a big chance a new exotic technology could shake things up) | ||
|
Nyxisto
Germany6287 Posts
| ||
|
Nesserev
Belgium2760 Posts
| ||
|
Deleted User 101379
4849 Posts
On August 02 2016 13:42 Nyxisto wrote: What are the opinions about Go around here? It doesn't fill any niche that isn't already covered by an established language and it's not significantly enough better (if it's better at all) than the established languages for the niches it does fill to warrant learning it. | ||
|
Doodsmack
United States7224 Posts
On July 23 2016 01:14 enigmaticcam wrote: That depends on how you use SQL server. Your typical DBA manages a SQL server mostly at an infrastructure level, which would all be proprietary to Microsoft. But at an application level, learning to collate data and understand how to talk to a database is absolutely relevant regardless of the language or implementation. Also not all databases support internet based applications. There are a lot of internal applications in the corporate world that require a database layer, and SQL Server thrives there. Thanks for the info all. I only have a basic understand of all this but what are some database technologies that may be used more for internet based applications (including cloud because I know there's Azure SQL). Are these technologies relational? I am learning T-SQL and feel pretty comfortable querying a relational db. | ||
| ||
