To read all comments associated with this story, please click here.
Using the EKOPath 4 Compiler Suite for compiling most C, C++ and Fortran code for Intel 64 and AMD should result in significant performace improvements over binaries compared to the same source code compiled with GCC.
This means that 64-bit binaries (for programs where you have the source code) will get faster. FOSS software has the source code ... even for drivers and the core OS.
Therefore, any given Linux distribution (for example) now has the opportunity to make significantly faster 64-bit binaries, across the board, than they could previously afford.
Drivers, kernel, graphics layers (including the likes of Cairo, GTK and Qt) ... even media players and codec decoders ... should all get a significant performance boost. Then there are the applications themselves, which should likewise enjoy a performance boost.
It seems to me to be something to really look forward to for end-users (of Intel and AMD 64-bit FOSS systems).
Sorry I'm late to this conversation (actually I'm not sorry based on some of these other posts)
What I'd like to do is share my biased perspective in general. EKOPath has historically catered to a certain audience and as a result a specific type of code. (HPC which is mostly computationally complex stuff) The performance for that does carry over in most cases (not just EKOPath) to other types of code. I'm happy with a bit of hype, but I want people to have realistic expectations. Something like your cli tools are unlikely to see any gains. Something like mesa which was recently benchmarked could see FPS go from 295 to 301 (I don't remember the exact numbers, but you get the idea)
Keep in mind we're not perfect, but.... - If you're code isn't faster we're going to work with you to fix it. So the potential gains long term imho are there for anyone to at least try.
In the future multicore systems are going to be more commonplace. (Even in your mobile phone) The compiler infrastructure needs to be able to handle the different types of optimizations which are needed to ensure best performance. Languages like C++ may get more/less complex. Our goal whether it's for C++/Fortran/ or something new will be the same. 
Nice, going by all this it means that we may even see some nice gains in the CPU limited Gallium3D drivers correct?
Things are looking really good then for my next build if AMD can get Gigabyte to do an A75 Fusion mobo with Coreboot instead of BIOS or UEFI so I can FOSS while I FOSS!




Member since:
2010-07-07
Does this mean anything for us unwashed masses of end users, or is it purely something for the code monkeys to drool over?