Граф коммитов

72 Коммитов

Автор SHA1 Сообщение Дата
Marko Radmilac 596d3287f9 Cleanup warnings in Linux 2015-09-03 09:24:03 -07:00
Frank Seide 795972137a Merge remote-tracking branch 'origin' into fseide/bs
Conflicts:
	DataReader/HTKMLFReader/ssematrix.h
	Math/Math/CUDAPageLockedMemAllocator.cpp
	Math/Math/Matrix.cpp
	Math/Math/NoGPU.cpp
	Math/Math/ValueQuantizer.h
2015-09-02 10:12:15 -07:00
Frank Seide 3c42fdc7ac some fixes to make gcc happy (one still missing);
somehow git lost 'Makefile'--readding, hoping it won't confuse the merge process
2015-09-02 08:43:31 -07:00
Frank Seide b369013c06 added BS to Linux Makefile--with lots of fallout that needs to be fixed 2015-09-01 13:44:20 -07:00
Amit Agarwal 6df22d9ecb Merge branch 'master' of https://git01.codeplex.com/cntk into amitaga/mergeHTKMLFReaders 2015-09-01 13:27:19 -07:00
Amit Agarwal 385fc0bb0a Merge the linux and windows forks of the HTKMLFReader fork into one. 2015-09-01 13:09:57 -07:00
Scott Cyphers 1036728875 Merge remote-tracking branch 'cntk/master' 2015-09-01 13:57:24 -04:00
Amit Agarwal 64f5fa7c0b Fixed CPUONLY build 2015-09-01 00:06:38 -07:00
Frank Seide 356aee0654 added BrainScript to Linux Makefile INCLUDEPATH 2015-08-31 21:06:28 -07:00
Scott Cyphers 2a498966fa Merge remote-tracking branch 'cntk/master' 2015-08-28 13:36:11 -04:00
Amit Agarwal cac9fad69c Addressed code review feedback 2015-08-17 14:48:06 -07:00
Scott Cyphers c6e29cf341 Cleanups, README improvements, missing Makefile dependency
Remove the no longer needed kaldi Makefiles.

Fix some pre-configure references in the README that were missed, add
a missing step to the build instructions, remove some obsolete
information, be more specific about the acml library that works, and
reword a few sentences.

Add math library dependencies to the kaldi plugins so that make -j
doesn't try to build them before the math library is finished.
2015-08-11 16:58:18 -04:00
Amit Agarwal 2670a56551 Merge branch 'master' of https://git01.codeplex.com/cntk into amitaga/parallelTraining
Conflicts:
	MachineLearning/CNTK/SGD.h
	Makefile
2015-08-11 10:36:15 -07:00
Scott Cyphers f43874cd84 Add a configure script for setting build parameters
Add a configure script for initializing build parameters, either
for in or out of source builds. The script generates a Config.make
in the build directory, and, for out of source builds, a trampoline
Makefile.

Make the build-and-test script to do an out of source build.

Add Config.make to .gitignore, as well as emacs temporary file patterns.
2015-08-07 13:03:27 -04:00
Scott Cyphers 30edcc0dac Use original build targets when PREFIX is not supplied. 2015-08-06 15:28:56 -04:00
Scott Cyphers 0950188608 Remove site-specific paths from Makefile.
Change configuration to build to a specific PREFIX directory, where
a Config.make is located that contains build and site-specific information.
This also makes it easy to check just how an earlier build was configured.

Update the instructions in README.
2015-08-06 12:41:57 -04:00
Scott Cyphers b6109fe5d3 Simplified unified CPU/GPU/Kaldi Makefile
Modularize build specifications for each target, where each target
adds what it needs to paths.

Add rpath to cntkmath and plugins so they do not need LD_LIBRARY_PATH.

Remove object files from cntk that were already in cntkmath.

Organize build targets into UNIX-like bin and lib directories under a
configuration-specific directory.  Have .gitignore ignore these
directories.

Make it easy to keep sources in alphabetic order for easier comparison
with the Windows project definition.
2015-08-05 18:46:31 -04:00
Amit Agarwal 7583752ccf Implementation of data-parallel SGD with 1bit quantization of gradients for reduced communication costs 2015-08-05 09:23:33 -07:00
Scott Cyphers e8de73a144 Add the appropriate math -rpath to shared libraries link. 2015-07-30 13:35:50 -04:00
Marko Radmilac 5d9337dcf8 Implement minibatch prefetching
This change implements a read-ahead (prefetch) of one minibatch on a separate thread to ensure that main thread always has enough work. This is done through coarse-level parallelism, where GetMinibatch is called from a separate thread and result is cached for main thread to consume. The synchronization is a simple producer/consumer model with a single mutex and a conditional variable.

This change also modifies how we measure time it takes to read and compute a minibatch worth of data.
2015-07-15 11:57:06 -07:00
Alexey Kamenev 752b35ced3 Add lib path for NVML, remove redundant lib references 2015-07-02 10:53:26 -07:00
Marko Radmilac 43e903515c Create common makefile, and remove cpu and gpu specific makefiles. Also do some cleanup. 2015-06-24 13:59:05 -07:00