summaryrefslogtreecommitdiff
path: root/benchmarks/hint/DESCR
diff options
context:
space:
mode:
authorzuntum <zuntum>2001-10-31 22:52:58 +0000
committerzuntum <zuntum>2001-10-31 22:52:58 +0000
commitec263708b4436f2f6fe8ff2c5ddd38bc49dbd322 (patch)
treeb3c17888184b1b17ed1b1fa28f1697197bb0b4f6 /benchmarks/hint/DESCR
parent26e767802169bc1a013a3a64740d0a84afa59755 (diff)
downloadpkgsrc-ec263708b4436f2f6fe8ff2c5ddd38bc49dbd322.tar.gz
Move pkg/ files into package's toplevel directory
Diffstat (limited to 'benchmarks/hint/DESCR')
-rw-r--r--benchmarks/hint/DESCR18
1 files changed, 18 insertions, 0 deletions
diff --git a/benchmarks/hint/DESCR b/benchmarks/hint/DESCR
new file mode 100644
index 00000000000..673f746a507
--- /dev/null
+++ b/benchmarks/hint/DESCR
@@ -0,0 +1,18 @@
+HINT (Hierarchical INTegration) is a computer benchmarking tool
+developed at the Scalable Computing Laboratory of the Ames Laboratory,
+a US Department of Energy laboratory. Unlike traditional benchmarks,
+HINT neither fixes the size of the problem nor the calculation time.
+Instead, it uses a rigorously defined measure called QUIPS (QUality
+Improvements Per Second).
+
+The output of HINT is data which can be plotted (the gnuplot graphics
+package is useful for plotting HINT output). The plot shows the speed
+of the computation as the size of the problem increases. The HINT
+graph shows the peak processor speed and the effects of different
+levels of cache and memory on processing speed. See
+http://www.scl.ameslab.gov/Projects/HINT/ for more information and
+an on-line database of HINT results.
+
+HINT also outputs a single number which is the integral of the HINT
+curve. This number, called NetQUIPS, can be used for general
+comparison between machines.