I have some new data that suggests there are issues inherent to pmap
and possibly other parallelism with Clojure on older Intel quad+ core
machines.
I added a noop loop to the benchmark. It looks like this:
(defn noops [n]
(when (> n 0)
(recur (- n 1
Running those in parallel is also n
> It seems very weird that my version of fac changes performance
> characteristics on my machine and not yours (OS/hardware dependent?).
> Can you tell your hardware configuration, esp. number of physical and
> logical cores?
It's an early Mac Pro with two dual-core Xeon 5150s, 5gb RAM, Mac OS
10.
Hi Zak,
It seems very weird that my version of fac changes performance
characteristics on my machine and not yours (OS/hardware dependent?).
Can you tell your hardware configuration, esp. number of physical and
logical cores? I am planning next to leave out using pmap and just try
to run the thing
ka, I ran some more tests, including partition-work and your version
of fac. I also ran some code from http://shootout.alioth.debian.org in
both C and Java.
On these 10-element sequences, partition-work seems to be a few tens
of milliseconds slower than partition-all. It does look generally
useful
Hi Zak,
I tried your example on my i7 (4 physical cores, 8 logical); here are
the results -
1:298 user=> (time (do (doall (map fac (take 10 (repeat 5
nil))
"Elapsed time: 54166.665145 msecs"
1:300 user=> (time (do (doall (pmap fac (take 10 (repeat 5
nil))
"Elapsed time: 27418.263
Heinz - playing with the size of the number doesn't have much effect,
except that when it becomes very small, parallelization overhead
eventually exceeds compute time.
Lee - Parallel GC slowed it down by 3 seconds on the four core
benchmark.
--
You received this message because you are subscribe
Zak,
This may not be your main issue and I haven't done enough testing with my own
code to know if it's even my main issue, but I've found that things appear to
go better for me on multicore machines if I invoke java with the
-XX:+UseParallelGC option.
-Lee
On May 30, 2010, at 12:31 PM, Zak
On May 30, 2010, at 18:31 , Zak Wilson wrote:
> I'm running Clojure code on an early Mac Pro with OS X 10.5 and Java
> 1.6. It has two dual-core Xeon 5150s and 5GB of memory.
Just a idea, two dual cores != 4 cores. Parallelism on more then one CPU is
always slower then on one cpu with multiple c
I'm running Clojure code on an early Mac Pro with OS X 10.5 and Java
1.6. It has two dual-core Xeon 5150s and 5GB of memory.
I'm not getting the performance I expected despite top reporting 390%
steady-state CPU use, so I wrote some trivial tests to see if I was
actually getting the benefit of all