multicdore-multithreading benchmarks

multicdore-multithreading benchmarks

[tmoran]

I have an N-cpu version of quicksort.  It speeds up nicely on my dual-core
machine.  Would some kind soul with a more-than-two processor machine be
willing to compile and run it for comparison?

Re: multicdore-multithreading benchmarks

[Alinabi]

I have a dual dual-core 64 bit machine (dual AMD Opteron). Where can I
get the code?

On Dec 18, 8:22 pm, tmo...@acm.org wrote:
> I have an N-cpu version of quicksort.  It speeds up nicely on my dual-core
> machine.  Would some kind soul with a more-than-two processor machine be
> willing to compile and run it for comparison?

Re: multicdore-multithreading benchmarks

[tmoran]

>I have a dual dual-core 64 bit machine (dual AMD Opteron). Where can I
>get the code?
home.comcast.net/~twmoran/trysortn.zip

Re: multicdore-multithreading benchmarks

[Alinabi]

Here are 5 different runs, for statistical purposes. Compiled with
gnatmake -O3.
Let me know if you want me to try other switches.

Run 1:
CPUs N= 1000        N= 10000      N= 100000     N= 1000000
 1    0.000142750    0.001998500   0.023578000   0.162093000
 2    0.000078750    0.001163750   0.012594250   0.103046500
 3    0.000078750    0.000876500   0.008899750   0.081707750
 4    0.000078750    0.001040500   0.009002500   0.085632000
 5    0.000079000    0.001277250   0.008339250   0.071098500
 6    0.000079000    0.001148750   0.008171250   0.071358750
 7    0.000079000    0.001434250   0.010027500   0.063942500
 8    0.000079000    0.001490750   0.008093750   0.064089000

Run 2:
CPUs N= 1000        N= 10000      N= 100000     N= 1000000
 1    0.000133500    0.002041500   0.023057000   0.160626250
 2    0.000074750    0.001211250   0.010813500   0.105832000
 3    0.000074250    0.000863750   0.007667750   0.070813000
 4    0.000074000    0.000876000   0.008855750   0.047772500
 5    0.000075250    0.000863000   0.008994750   0.049151250
 6    0.000074000    0.001062500   0.007130750   0.049553500
 7    0.000074250    0.001249750   0.007859750   0.051682000
 8    0.000074250    0.001496000   0.008086000   0.061584250

Run 3:
CPUs N= 1000        N= 10000      N= 100000     N= 1000000
 1    0.000145500    0.001793500   0.023460750   0.154006250
 2    0.000079250    0.001111250   0.012812750   0.137050500
 3    0.000143000    0.001504750   0.012376500   0.105685000
 4    0.000145750    0.001228000   0.011719000   0.095653250
 5    0.000079500    0.001309500   0.011715500   0.077465000
 6    0.000079750    0.001695000   0.011437000   0.078212500
 7    0.000144250    0.001776000   0.010947500   0.076057500
 8    0.000079750    0.001518000   0.011901750   0.075582500

Run 4:
CPUs N= 1000        N= 10000      N= 100000     N= 1000000
 1    0.000140250    0.001837250   0.020237750   0.152752500
 2    0.000077500    0.001121500   0.012037750   0.130276250
 3    0.000077750    0.001041750   0.007699500   0.077842750
 4    0.000077750    0.001173250   0.007399250   0.070188750
 5    0.000077750    0.000950000   0.007958250   0.069657000
 6    0.000077750    0.001154000   0.008442750   0.068289000
 7    0.000078000    0.001137500   0.009036750   0.065445000
 8    0.000078000    0.001197000   0.009599500   0.064271500

Run 5:
CPUs N= 1000        N= 10000      N= 100000     N= 1000000
 1    0.000143000    0.001761500   0.023593000   0.165574250
 2    0.000079250    0.001087250   0.011648750   0.149709000
 3    0.000079250    0.001099250   0.011416750   0.085586250
 4    0.000079250    0.000976250   0.010137750   0.086734250
 5    0.000079250    0.001110500   0.009749000   0.077554500
 6    0.000079250    0.001242750   0.010455250   0.070517250
 7    0.000079000    0.001301750   0.008186750   0.065894500
 8    0.000079250    0.001370000   0.008127750   0.065806000

On Dec 18, 9:50 pm, tmo...@acm.org wrote:
> >I have a dual dual-core 64 bit machine (dual AMD Opteron). Where can I
> >get the code?home.comcast.net/~twmoran/trysortn.zip