Skip to content

Commit

Permalink
Update instructions.
Browse files Browse the repository at this point in the history
  • Loading branch information
devinamatthews committed Jul 25, 2017
1 parent cef4626 commit 7dc3540
Show file tree
Hide file tree
Showing 2 changed files with 14 additions and 1 deletion.
7 changes: 6 additions & 1 deletion optimizing_gemm/Makefile
Original file line number Diff line number Diff line change
@@ -1,8 +1,13 @@
SHELL=/bin/bash

#OpenBLAS
#BLAS_LIBS=-lopenblas
#MKL
BLAS_LIBS=-lmkl_intel_lp64 -lmkl_sequential -lmkl_core

CXX=g++
CXXFLAGS=-std=c++11 -O3 -march=native -fopenmp -DNDEBUG -I$(HOME)/miniconda/envs/sss/include -Wno-depracated-declarations
LDFLAGS=-fopenmp -lopenblas -Wl,-rpath,$(HOME)/miniconda/envs/sss/lib
LDFLAGS=-fopenmp $(BLAS_LIBS) -Wl,-rpath,$(HOME)/miniconda/envs/sss/lib

all: run

Expand Down
8 changes: 8 additions & 0 deletions optimizing_gemm/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -33,8 +33,16 @@ Prerequisites:
- OpenBLAS (may also be changed in the Makefile)
- Eigen v3
- gnuplot
- ghostscript
- Intel or AMD x86-64 processor with AVX (for the last example only, FMA support is required)

In an anaconda environment, the additional requirements can be installed with:

```bash
conda install -c menpo -c bioconda eigen=3.2.7 gnuplot=4.6 openblas
conda install -c conda-forge ghostscript
```

The various example implementations of GEMM are contained in the files `my_dgemm_<n>.cxx` where `n` is from 0 to 8. To compile and automatically run and plot the first "triple-loop" example, run the command:

```
Expand Down

0 comments on commit 7dc3540

Please sign in to comment.