Results

1. ()

PARAMS = { flag, {block0, part0}, {block1, part1} }. flag = {"-Xptxas -dlcm=ca", "-Xptxas -dlcm=cg" } # compiler flag.

people.maths.ox.ac.uk/gilesm/talks/mrsc11.pdf- 66k - 11 Apr 2011 - Cached

2. ()

PARAMS = { flag, {block0, part0}, {block1, part1} }. flag = {"-Xptxas -dlcm=ca", "-Xptxas -dlcm=cg" } # compiler flag.

people.maths.ox.ac.uk/gilesm/talks/sandia_1.pdf- 101k - 2 Feb 2012 - Cached

3. ()

PARAMS = { flag, {block0, part0}, {block1, part1} }. flag = {"-Xptxas -dlcm=ca", "-Xptxas -dlcm=cg" } # compiler flag.

people.maths.ox.ac.uk/gilesm/talks/ed_4.pdf- 109k - 7 Jan 2012 - Cached

4. CUDA_Course_profiling2

Quick & easy step:Turn on L1 cache by using-Xptxas -dlcm=ca. Cache line utilization.

people.maths.ox.ac.uk/~gilesm/cuda/2019/CUDA_Course_profiling2.pdf- 1466k - 19 Jul 2019 - Cached

5. doi.10.1112/comjnl/bxr062

GTX260 -O3 -arch=sm_13 -Xptxas -Xptxas=-v-dlcm=ca -use_fast_math. Tesla C2050 -O3 -arch=sm_20 -Xptxas -Xptxas=-v-dlcm=ca -use_fast_math.

people.maths.ox.ac.uk/gilesm/files/OP2-CJ.pdf- 576k - 4 Mar 2013 - Cached

6. https://people.maths.ox.ac.uk/gilesm/files/JPDC-OP2.pdf

dlcm=ca-use fast math compiler flags.

people.maths.ox.ac.uk/gilesm/files/JPDC-OP2.pdf- 936k - 23 May 2012 - Cached