serban / kmeans Goto Github PK

I have a .txt which include 750,000 obj, every obj has 2 coords, the K is 5, i calculate the result in less than 1 second... it's so faster than what i think. Is that normal?

my GPU is 1060
Performing **** Regular Kmeans (CUDA version) ****
Input file: Image_data/anchor_boxes.txt
numObjs = 758896
numCoords = 2
numClusters = 5
threshold = 0.0010
Loop iterations = 1
I/O time = 0.3839 sec
Computation timing = 0.0988 sec

Hi, I am planning to run this program on a dataset sized almost 100GB on my server(more than 200GB of mem).
Could you please tell how to implement is cause I constantly getting a 'segmentation fault' error message when the memory exceed 4GB.
I have checked the BLAS and LAPACK libraries are all 64 bits version.
omp_kmeans is also compiled using a 64bit gcc compiler.

Thank you for your kindness.

Using -O3 changes the performance on the CPU dramatically in my case, especially the openmp version.

Hi,
I am planning to run cuda_main executable which is generated from Makefile. My input file contains 102760448 rows and 1 column. When I ran the cuda_main program, I am getting the following error :
lineLen = 128
File ../fc6weights_flattened.txt numObjs = 102760448
File ../fc6weights_flattened.txt numCoords = 0
CUDA Error 11: invalid argument

I have a doubt why it is showing numCoords as 0, ideally it should have shown as 1.

Can you please help me find what am I missing ?