www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - D Wins the Shootout

reply Ant <Someone nowhere.com> writes:
http://shootout.alioth.debian.org/great/benchmark.php?test=all&lang=all&sort=fullcpu
Jun 07 2005
parent reply Kyle Furlong <ky220 umail.ucsb.edu> writes:
Yes, but it is the only language with all the benchmarks coded.

Ant wrote:
 http://shootout.alioth.debian.org/great/benchmark.php?test=all&l
ng=all&sort=fullcpu 
 

Jun 07 2005
next sibling parent reply "Ben Hinkle" <bhinkle mathworks.com> writes:
"Kyle Furlong" <ky220 umail.ucsb.edu> wrote in message 
news:d84lvd$2r07$1 digitaldaemon.com...
 Yes, but it is the only language with all the benchmarks coded.

 Ant wrote:
 http://shootout.alioth.debian.org/great/benchmark.php?test=all&lang=all&sort=fullcpu


It's also interesting to compare D with other languages across the implemented benchmarks: http://shootout.alioth.debian.org/great/benchmark.php?test=all&lang=dlang&lang2=icpp&sort=fullcpu
Jun 07 2005
next sibling parent reply Chris Sauls <ibisbasenji gmail.com> writes:
Ben Hinkle wrote:
 It's also interesting to compare D with other languages across the 
 implemented benchmarks:
 http://shootout.alioth.debian.org/great/benchmark.php?test=all&lang=dlang&lang2=icpp&sort=fullcpu

From the looks of that one, memory use is the one crux of the current D. -- Chris Sauls
Jun 07 2005
parent "Ben Hinkle" <bhinkle mathworks.com> writes:
"Chris Sauls" <ibisbasenji gmail.com> wrote in message 
news:d84pdb$2tki$1 digitaldaemon.com...
 Ben Hinkle wrote:
 It's also interesting to compare D with other languages across the 
 implemented benchmarks:
 http://shootout.alioth.debian.org/great/benchmark.php?test=all&lang=dlang&lang2=icpp&sort=fullcpu

From the looks of that one, memory use is the one crux of the current D. -- Chris Sauls

It's interesting to note the memory use column seems to be heavily dependent on the startup footprint. For example the "harmonic" test is just void main(char[][] args) { int i = 0, n = args.length > 1 ? atoi(args[1]) : 10000000; double partialSum = 0.0; for (int i=1; i<=n; i++) partialSum += 1.0/i; writefln("%0.9f",partialSum); } and yet it uses 1.92 times the memory of the C++ version. I assume that's due to startup footprint. I guess it's only a rough measurement since the python memory usage for the same test is 28 times the C++.
Jun 07 2005
prev sibling parent reply Hasan Aljudy <hasan.aljudy gmail.com> writes:
Ben Hinkle wrote:
 "Kyle Furlong" <ky220 umail.ucsb.edu> wrote in message 
 news:d84lvd$2r07$1 digitaldaemon.com...
 
Yes, but it is the only language with all the benchmarks coded.

Ant wrote:

http://shootout.alioth.debian.org/great/benchmark.php?test=all&lang=all&sort=fullcpu


It's also interesting to compare D with other languages across the implemented benchmarks: http://shootout.alioth.debian.org/great/benchmark.php?test=all&lang=dlang&lang2=icpp&sort=fullcpu

sorry for sounding like an idiot, but, can someone decipher that chart for me? :P
Jun 07 2005
parent "Ben Hinkle" <ben.hinkle gmail.com> writes:
"Hasan Aljudy" <hasan.aljudy gmail.com> wrote in message 
news:d855mu$7gb$1 digitaldaemon.com...
 Ben Hinkle wrote:
 "Kyle Furlong" <ky220 umail.ucsb.edu> wrote in message 
 news:d84lvd$2r07$1 digitaldaemon.com...

Yes, but it is the only language with all the benchmarks coded.

Ant wrote:

http://shootout.alioth.debian.org/great/benchmark.php?test=all&lang=all&sort=fullcpu


It's also interesting to compare D with other languages across the implemented benchmarks: http://shootout.alioth.debian.org/great/benchmark.php?test=all&lang=dlang&lang2=icpp&sort=fullcpu

sorry for sounding like an idiot, but, can someone decipher that chart for me? :P

larger numbers are bad. So if the plot has bars extending to the left (the D side) then D was performing worse than Intel C++ (black for memory, white for time). The table of numbers below the chart has the ratio of the two languages. So for example if you have a number 30 for the memory usage of a test foo then it means foo took up 30 more memory implemented in D than C++.
Jun 07 2005
prev sibling next sibling parent reply clayasaurus <clayasaurus gmail.com> writes:
Kyle Furlong wrote:
 Yes, but it is the only language with all the benchmarks coded.
 
 Ant wrote:
 
 http://shootout.alioth.debian.org/great/benchmark.php?test=all&l
ng=all&sort=fullcpu 


no, SML MLton has all the benchmarks coded.
Jun 07 2005
parent Kyle Furlong <ky220 umail.ucsb.edu> writes:
Yes, I didnt see that one at first, thanks for the correction.

clayasaurus wrote:
 Kyle Furlong wrote:
 
 Yes, but it is the only language with all the benchmarks coded.

 Ant wrote:

 http://shootout.alioth.debian.org/great/benchmark.php?test=all&l
ng=all&sort=fullcpu 


no, SML MLton has all the benchmarks coded.

Jun 07 2005
prev sibling parent reply =?iso-8859-1?q?Knud_S=F8rensen?= <12tkvvb02 sneakemail.com> writes:
On Tue, 07 Jun 2005 10:36:49 -0700, Kyle Furlong wrote:

 Yes, but it is the only language with all the benchmarks coded.
 
 Ant wrote:
 http://shootout.alioth.debian.org/great/benchmark.php?test=all&l
ng=all&sort=fullcpu 


Hi I have made at a program to find the test where D is performing pure. For cpu time the 10 worst different test is 0.02941 objinst(1500000) 0.03846 message(150) 0.06452 echo(40000) 0.08333 except(100000) 0.12360 regexmatch(12000) 0.23529 prodcons(100000) 0.25000 ary(5000) 0.32143 hello(150) 0.40000 random(600000) 0.41176 matrix(600) for memory it is 0.00338 fibo(12) 0.00529 except(100000) 0.01717 ackermann(7) 0.03078 hash2(150) 0.03909 sieve(1200) 0.06250 fasta(2500000) 0.06566 ary(5000) 0.07426 nestedloop(12) 0.07937 except(200000) 0.08371 strcat(40000) for code lines it is 0.04000 wordfreq(10) 0.05263 wc(1500) 0.07143 reversefile(10) 0.10345 lists(12) 0.16327 process(1000) 0.16667 nestedloop(12) 0.16667 sumcol(1000) 0.22472 pidigits(200) 0.28302 message(100) 0.29412 objinst(1000000) The number in front is (best non d benchmark)/(d benchmark)
Jun 08 2005
parent reply Will <Will_member pathlink.com> writes:
In article <pan.2005.06.08.10.10.30.70738 sneakemail.com>,
=?iso-8859-1?q?Knud_S=F8rensen?= says...
On Tue, 07 Jun 2005 10:36:49 -0700, Kyle Furlong wrote:

 Yes, but it is the only language with all the benchmarks coded.
 
 Ant wrote:
 http://shootout.alioth.debian.org/great/benchmark.php?test=all&l
ng=all&sort=fullcpu 


Hi I have made at a program to find the test where D is performing pure. For cpu time the 10 worst different test is 0.02941 objinst(1500000) 0.03846 message(150) 0.06452 echo(40000) 0.08333 except(100000) 0.12360 regexmatch(12000) 0.23529 prodcons(100000) 0.25000 ary(5000) 0.32143 hello(150) 0.40000 random(600000) 0.41176 matrix(600) for memory it is 0.00338 fibo(12) 0.00529 except(100000) 0.01717 ackermann(7) 0.03078 hash2(150) 0.03909 sieve(1200) 0.06250 fasta(2500000) 0.06566 ary(5000) 0.07426 nestedloop(12) 0.07937 except(200000) 0.08371 strcat(40000) for code lines it is 0.04000 wordfreq(10) 0.05263 wc(1500) 0.07143 reversefile(10) 0.10345 lists(12) 0.16327 process(1000) 0.16667 nestedloop(12) 0.16667 sumcol(1000) 0.22472 pidigits(200) 0.28302 message(100) 0.29412 objinst(1000000) The number in front is (best non d benchmark)/(d benchmark)

Are the values to the left C/C++? ..and no period is missing to the right??
Jun 10 2005
parent =?iso-8859-1?q?Knud_S=F8rensen?= <12tkvvb02 sneakemail.com> writes:
On Fri, 10 Jun 2005 16:23:22 +0000, Will wrote:

 In article <pan.2005.06.08.10.10.30.70738 sneakemail.com>,
 =?iso-8859-1?q?Knud_S=F8rensen?= says...
On Tue, 07 Jun 2005 10:36:49 -0700, Kyle Furlong wrote:

 Yes, but it is the only language with all the benchmarks coded.
 
 Ant wrote:
 http://shootout.alioth.debian.org/great/benchmark.php?test=all&l
ng=all&sort=fullcpu 


Hi I have made at a program to find the test where D is performing pure. For cpu time the 10 worst different test is 0.02941 objinst(1500000) 0.03846 message(150) 0.06452 echo(40000) 0.08333 except(100000) 0.12360 regexmatch(12000) 0.23529 prodcons(100000) 0.25000 ary(5000) 0.32143 hello(150) 0.40000 random(600000) 0.41176 matrix(600) for memory it is 0.00338 fibo(12) 0.00529 except(100000) 0.01717 ackermann(7) 0.03078 hash2(150) 0.03909 sieve(1200) 0.06250 fasta(2500000) 0.06566 ary(5000) 0.07426 nestedloop(12) 0.07937 except(200000) 0.08371 strcat(40000) for code lines it is 0.04000 wordfreq(10) 0.05263 wc(1500) 0.07143 reversefile(10) 0.10345 lists(12) 0.16327 process(1000) 0.16667 nestedloop(12) 0.16667 sumcol(1000) 0.22472 pidigits(200) 0.28302 message(100) 0.29412 objinst(1000000) The number in front is (best non d benchmark)/(d benchmark)

Are the values to the left C/C++? ..and no period is missing to the right??

Lets take cpu time for objinst(1500000) The best test is clean with 0.05 secs. D completes the same test in 1.7 secs. Which gives 0.05/1.7 = 0.02941
Jun 12 2005