shared-memory

In this thread the OP is suggested to use mmap() instead of shmget() to get shared memory in Linux. I visited this page and this page to get some documentation, but the second one gives an obscure example regarding mmap() . Being almost a newbie, and needing to share some information (in text form) between two processes, should I use the shmget() method or mmap() ? And why? Sergey L. Both methods are viable. mmap method is a little bit more restrictive then shmget , but easier to use. shmget is the old System V shared memory model and has the widest support. mmap / shm_open is the new POSIX

I have a similar question as before in boost::interprocess Containers of containers NOT in shared memory and How to I create a boost interprocess vector of interprocess containers but this time I like to use my class, which uses a scoped_allocator, also on the heap and shared memory. The solution to my first question was to use a template class with an allocator type In my second previous question it turned out that using a scoped_allocator together with a container of containers within the shared memory makes live easier. Now I like to have both, is this possible? Attached a example with a

Is there a way in PHP to use "out of session" variables, which would not be loaded/unloaded at every connexion, like in a Java server ? Please excuse me for the lack of accuracy, I don't figure out how to write it in a proper way. The main idea would be to have something like this : <?php ... // $variablesAlreadyLoaded is kind of "static" and shared between all PHP threads // No need to initialize/load/instantiate it. $myVar = $variablesAlreadyLoaded['aConstantValueForEveryone']; ... ?> I already did things like this using shmop and other weird things, but if there is a "clean" way to do this

I'm trying to implement a reduction along the row direction of a 2D matrix. I'm starting from a code I found on stackoverflow (thanks a lot Robert!) thrust::max_element slow in comparison cublasIsamax - More efficient implementation? The above link shows a custom kernel that performs reduction on a single row. It divides the input row into many rows and each row has 1024 threads. Works very well. For the 2D case, everything's the same except that now there's a y grid dimension. So each block's y dimension is still 1. The problem is that when I try to write data onto the shared memory within

Is it possible to share variables and arrays between all PHP processes without duplicating them ? Using memcached, I think PHP duplicates the used memory: $array = $memcache->get('array'); $array will contain a copy from memcached. So my idea is, there could be a static variable that was already defined, and shared between all processes. By default its simply not possible. Every solution will always copy the content into the current scope, because if not, there is no way to access it. I dont know, what exactly want to do, but maybe you can do that "outside", for example as a gearman job, and

I am implementing two processes on a LynxOS SE (POSIX conformant) system that will communicate via shared memory. One process will act as a "producer" and the other a "consumer". In a multi-threaded system my approach to this would be to use a mutex and condvar (condition variable) pair, with the consumer waiting on the condvar (with pthread_cond_wait ) and the producer signalling it (with pthread_cond_signal ) when the shared memory is updated. How do I achieve this in a multi-process, rather than multi-threaded, architecture? Is there a LynxOS/POSIX way to create a condvar/mutex pair that