aboutsummaryrefslogtreecommitdiff
path: root/benchmarks/ramspeed/pkg-descr
diff options
context:
space:
mode:
authorPawel Pekala <pawel@FreeBSD.org>2014-01-31 13:40:21 +0000
committerPawel Pekala <pawel@FreeBSD.org>2014-01-31 13:40:21 +0000
commit97fe3ab28b6675ce69c23347988532bf435f31c8 (patch)
treec4e7c1bc7a19727a91d24f518977ebe8893dd976 /benchmarks/ramspeed/pkg-descr
parent9283ec7f2a9b0907103e580e5baa29cb8b964ebe (diff)
Notes
Diffstat (limited to 'benchmarks/ramspeed/pkg-descr')
-rw-r--r--benchmarks/ramspeed/pkg-descr15
1 files changed, 15 insertions, 0 deletions
diff --git a/benchmarks/ramspeed/pkg-descr b/benchmarks/ramspeed/pkg-descr
new file mode 100644
index 000000000000..ef9396c1bc5e
--- /dev/null
+++ b/benchmarks/ramspeed/pkg-descr
@@ -0,0 +1,15 @@
+RAMspeed is a command line utility to measure cache and memory performance of
+computer systems. It offers 18 cache and memory benchmarks for i386 and amd64
+machines, though 6 only for alpha ones. There are *mark benchmarks such as
+INTmark, FLOATmark, MMXmark and SSEmark. They operate with linear (sequential)
+data streams passed through ALU, FPU, MMX and SSE units respectively.
+There are also *mem benchmarks such as INTmem, FLOATmem, MMXmem and SSEmem.
+These are supposed to illustrate how fast is actual read/write memory
+performance. There are also non-temporal versions of MMX and SSE benchmarks.
+They have been coded with special instructions to minimise cache pollution on
+memory reads and to eliminate it completely on memory writes. In addition, they
+operate with a built in aggressive data prefetching algorithm. In some cases,
+non-temporal MMXmark and SSEmark can deliver almost 100% of theoretical
+bandwidth while reading.
+
+WWW: http://alasir.com/software/ramspeed/