Benchmarking Mathematica 13 across machines

Question

Many people consider the performance of MMA when choosing their computers. I feel we need some up-to-date benchmark results across hardware and systems for the major version 13, which will help a lot of people not even limited to users here.

So if you agree, please run the following and post the result as an answer.

Quit[]

then:

1+1;
Needs["Benchmarking`"]
Benchmark[]

Running Quit[] ensures the benchmark is evaluated on a fresh kernel, thus avoiding artificially high results due to caching.

It feels as though this is partially a duplicate of this question, which is the benchmarking thread that got the most traction: mathematica.stackexchange.com/questions/234881/… — Carl Lange, Apr 25, 2022 at 7:24
@CarlLange But maybe it still makes sense to do this in a new thread for major versions? — gwr, Apr 25, 2022 at 9:36
Does this answer your question? Benchmarking with Mathematica v.12 for up to date comparison across different machines — bbgodfrey, Apr 29, 2022 at 19:02
@bbgodfrey I do not understand the reason behind marking this as duplicate. Maybe I am too naive about it, but the post to which you are referring clearly discusses V12 whilst this one is for V13. Could you please elaborate a bit? Many thanks! — bmf, Apr 29, 2022 at 23:20
I think these benchmark posts for major versions are valuable and should not be closed as duplicates of older Q&A. However I also think these should be "Community Wiki", in line with other "share your results/tips" style posts. I am going to 1) close and reopen this post to clear out the votes and 2) mark this and the other v12 post as community wikis. It would help if we could standardize on the output format — this one asks for Benchmark[] whereas the other asks for BenchmarkReport[]. Having consistency across versions would help future readers. — rm -rf, May 1, 2022 at 14:23

xiaohuamao · Accepted Answer · 2022-04-25 04:59:51Z

3

iMac (Mid 2020) 27-inch, 3.6GHz 10-Core Intel i9-10910, 128GB RAM, Radeon 5700, macOS Monterey with MMA 13.0.1

single-core

{"System" -> "Mac OS X x86 (64-bit)", 
 "BenchmarkName" -> "WolframMark", "FullVersionNumber" -> "13.0.1", 
 "Date" -> "April 25, 2022", "BenchmarkResult" -> 4.664, 
 "TotalTime" -> 2.968, "Results" -> {{"Data Fitting", 0.211}, 
   {"Digits of Pi", 0.222}, {"Discrete Fourier Transform", 0.165}, 
   {"Eigenvalues of a Matrix", 0.254}, {"Elementary Functions", 0.16}, 
   {"Gamma Function", 0.319}, {"Large Integer Multiplication", 0.306}, 
   {"Matrix Arithmetic", 0.108}, {"Matrix Multiplication", 0.073}, 
   {"Matrix Transpose", 0.151}, {"Numerical Integration", 0.34}, 
   {"Polynomial Expansion", 0.049}, {"Random Number Sort", 0.346}, 
   {"Singular Value Decomposition", 0.145}, {"Solving a Linear System", 
    0.119}}}

multi-core (LaunchKernels[]; before the test)

{"System" -> "MacOSX-x86-64", "BenchmarkName" -> "WolframMark", 
 "FullVersionNumber" -> "13.0.1", "Date" -> "April 25, 2022", 
 "BenchmarkResult" -> 10.574, "TotalTime" -> 39.27}

answered Apr 25, 2022 at 4:59

community wiki

xiaohuamao

1

$\begingroup$ How many kernels were launched? On PC the benchmark includes this info. $\endgroup$
– Edmund
Apr 25, 2022 at 11:17
$\begingroup$ @Edmund I presume LaunchKernels[] launches all by default, which is 10 for this machine and is also what I saw in the activity monitor. $\endgroup$
– xiaohuamao
Apr 25, 2022 at 13:06
$\begingroup$ My 6 Intel cores beat your 10 Mac cores. :-) "BenchmarkResult" -> 7.801, "TotalTime" -> 31.938} $\endgroup$
– Edmund
Apr 25, 2022 at 13:12
3

$\begingroup$ @Edmund Seems that you've got a good machine :) Why not post an answer? BTW, mine is intel as well. $\endgroup$
– xiaohuamao
Apr 26, 2022 at 0:08
2

$\begingroup$ But that's not what's going on. Instead, the wall clock time to run WolframMark with LaunchKernels[] is multiples higher than without it, indicating it's not speeding things up, it's slowing them down. I think what's happening is that, when you execute LaunchKernels[], MMA runs several WolframMarks simultaneously (which is why it takes so much longer) and then does some sort of sum on the individual scores (which is why you get an artificially high score). $\endgroup$
– theorist
Jun 11, 2022 at 4:18

| Show 1 more comment

Stuart Poss · Accepted Answer · 2022-07-03 02:21:42Z

Dell Alienware Aurora-R13 Desktop Device name Stonefish-Alienware-Aurora-R13 Processor 12th Gen Intel(R) Core(TM) i9-12900KF 3.19 GHz Installed RAM 64.0 GB (63.8 GB usable) System type 64-bit operating system, x64-based processor

Quit[]

1 + 1;
Needs["Benchmarking`"]
Benchmark[]

Out[3]//InputForm=
{"MachineName" -> "stonefish-alien", 
"System" -> "Microsoft Windows (64-bit)", 
"BenchmarkName" -> "WolframMark", "FullVersionNumber" -> 
"13.0.1", "Date" -> "July 2, 2022", 
"BenchmarkResult" -> 3.687, "TotalTime" -> 3.754, 
"Results" -> {{"Data Fitting", 0.249}, 
{"Digits of Pi", 0.178}, {"Discrete Fourier Transform", 
0.347}, {"Eigenvalues of a Matrix", 0.297}, 
{"Elementary Functions", 0.358}, {"Gamma Function", 0.701}, 
{"Large Integer Multiplication", 0.236}, 
{"Matrix Arithmetic", 0.215}, {"Matrix Multiplication", 
0.097}, {"Matrix Transpose", 0.235}, 
{"Numerical Integration", 0.308}, {"Polynomial Expansion", 
0.029}, {"Random Number Sort", 0.088}, 
{"Singular Value Decomposition", 0.215}, 
{"Solving a Linear System", 0.201}}}

bmf · Accepted Answer · 2022-04-28 15:55:35Z

2

I am sitting on a

MacBook Pro (13-inch, M1, 2020) that has the Apple M1 chip and macOS Monterey 12.3.1

Needs["Benchmarking`"]
Benchmark[]

{"System" -> "Mac OS X ARM (64-bit)", 
 "BenchmarkName" -> "WolframMark", "FullVersionNumber" -> "13.0.0", 
 "Date" -> "April 28, 2022", "BenchmarkResult" -> 3.048, "TotalTime" -> 4.541, 
 "Results" -> {{"Data Fitting", 0.177}, {"Digits of Pi", 0.172}, 
   {"Discrete Fourier Transform", 0.318}, {"Eigenvalues of a Matrix", 0.261}, 
   {"Elementary Functions", 1.047}, {"Gamma Function", 0.232}, 
   {"Large Integer Multiplication", 0.2}, {"Matrix Arithmetic", 0.185}, 
   {"Matrix Multiplication", 0.398}, {"Matrix Transpose", 0.154}, 
   {"Numerical Integration", 0.326}, {"Polynomial Expansion", 0.039}, 
   {"Random Number Sort", 0.406}, {"Singular Value Decomposition", 0.364}, 
   {"Solving a Linear System", 0.262}}}

Also,

LaunchKernels[]

and do the test

{"MachineName" -> "8-node homogeneous cluster", "System" -> "MacOSX-ARM64", 
 "BenchmarkName" -> "WolframMark", "FullVersionNumber" -> "13.0.0", 
 "Date" -> "April 28, 2022", "BenchmarkResult" -> 9.175, "TotalTime" -> 36.208}

answered Apr 28, 2022 at 15:55

community wiki

bmf

$\begingroup$ These M1 numbers look bad, but they're not as bad as they could be. When I looked at this on an A12X, it was clear that the Wolfram code was - not using very good bignum code (slow Digits of Pi and Gamma Function) - mostly not using other cores for large vectors/matrices - probably not using SIMD for arrays of special functions. Bignums looks like it's on the way to being fixed. SIMD and multi-core look like they're mostly not yet touched. Likewise looks like no functions are routing to AMX. Probably worth trying again on 13.1 to see further progress. $\endgroup$
– Maynard Handley
Jul 8, 2022 at 23:41
$\begingroup$ Wolfram have always been about doing the job right, not quick and easy. So I expect they are taking their time to make sure bignums and SIMD are implemented optimally for ARM64. I'm not sure what the multi-core holdup is except if, perhaps, they want a generic framework that handles het cores (P and E) and/or accelerators (like AMX) optimally and aren't really bothering with moving packed vectors/matrices to multi-core until that framework is in place (which they also need, anyway, for Alder Lake et seq, and Intel's AMX). $\endgroup$
– Maynard Handley
Jul 8, 2022 at 23:44

Add a comment |

DrMrstheMonarch · Accepted Answer · 2022-04-29 11:08:41Z

I have posted benchmark results for 12.3.1 in this previous thread for my various machines, but this time I tried using Parallelize[]; and LaunchKernels[]; ahead and got varying results.

Macbook Pro M1 13'' 2020, 16GB Ram

Parallelize[];
Benchmark[]

{"MachineName" -> 
  "4-node homogeneous cluster", 
 "System" -> "MacOSX-ARM64", 
 "BenchmarkName" -> "WolframMark", 
 "FullVersionNumber" -> "12.3.1", 
 "Date" -> "April 29, 2022", 
 "BenchmarkResult" -> 6.639, 
 "TotalTime" -> 25.021}

Benchmark[]
{"MachineName" -> "laederlappen", "System" -> "Mac OS X ARM (64-bit)", 
 "BenchmarkName" -> "WolframMark", "FullVersionNumber" -> "12.3.1", 
 "Date" -> "April 29, 2022", "BenchmarkResult" -> 3.178, 
 "TotalTime" -> 4.355, "Results" -> {{"Data Fitting", 0.14}, 
   {"Digits of Pi", 0.171}, {"Discrete Fourier Transform", 0.298}, 
   {"Eigenvalues of a Matrix", 0.303}, {"Elementary Functions", 0.659}, 
   {"Gamma Function", 0.224}, {"Large Integer Multiplication", 0.188}, 
   {"Matrix Arithmetic", 0.104}, {"Matrix Multiplication", 0.183}, 
   {"Matrix Transpose", 0.139}, {"Numerical Integration", 0.766}, 
   {"Polynomial Expansion", 0.062}, {"Random Number Sort", 0.435}, 
   {"Singular Value Decomposition", 0.434}, {"Solving a Linear System", 
    0.249}}}

LaunchKernels[];
Benchmark[]
{"MachineName" -> "4-node homogeneous cluster", "System" -> "MacOSX-ARM64", 
 "BenchmarkName" -> "WolframMark", "FullVersionNumber" -> "12.3.1", 
 "Date" -> "April 29, 2022", "BenchmarkResult" -> 6.853, 
 "TotalTime" -> 24.239}

I have a feeling the benchmark isn't a reliable measure of machine ability.

2 revs · Accepted Answer · 2022-05-01 15:13:02Z

Laptops don't fair to well against desktops but they are easier to carry.

Lenovo ThinkBook P15 Gen1, Intel Xeon W-10855M (max 5.1 GHz), 128GB DDR4 RAM, NVIDIA Quadro RTX 4000 Max-Q (8 GB), UHD 4k 15in

Needs["Benchmarking`"]
Benchmark[]

{"MachineName" -> "whaleshark", "System" -> "Microsoft Windows (64-bit)"
 , "BenchmarkName" -> "WolframMark", "FullVersionNumber" -> "13.0.1"
 , "Date" -> "May 1, 2022"
 , "BenchmarkResult" -> 3.483, "TotalTime" -> 3.974
 , "Results" -> {
    {"Data Fitting", 0.258}
    , {"Digits of Pi", 0.249}
    , {"Discrete Fourier Transform", 0.321}
    , {"Eigenvalues of a Matrix", 0.286}
    , {"Elementary Functions", 0.395}
    , {"Gamma Function", 0.362}
    , {"Large Integer Multiplication", 0.338}
    , {"Matrix Arithmetic", 0.242}
    , {"Matrix Multiplication", 0.19}
    , {"Matrix Transpose", 0.369}
    , {"Numerical Integration", 0.388}
    , {"Polynomial Expansion", 0.04}
    , {"Random Number Sort", 0.125}
    , {"Singular Value Decomposition", 0.2}
    , {"Solving a Linear System", 0.211}
   }
}

LaunchKernels[]
Benchmark[]

{"MachineName" -> "6-node homogeneous cluster", "System" -> "Windows-x86-64"
 , "BenchmarkName" -> "WolframMark", "FullVersionNumber" -> "13.0.1"
 , "Date" -> "May 1, 2022"
 , "BenchmarkResult" -> 7.601, "TotalTime" -> 32.781
}

2 revs · Accepted Answer · 2022-06-11 04:44:56Z

2019 27" iMac, 8-core 3.6/5.0 GHz i9-9900K (Coffee Lake), 32 GB RAM, Radeon 580X (8 GB).

MacOS Monterey, v. 12.4

MMA 13.0.1.

Score: 4.602

Run on a fresh kernel.

{"MachineName" -> "imac", "System" -> "Mac OS X x86 (64-bit)", 
 "BenchmarkName" -> "WolframMark", "FullVersionNumber" -> "13.0.1", 
 "Date" -> "June 10, 2022", "BenchmarkResult" -> 4.602, "TotalTime" -> 3.008, 
 "Results" -> {{"Data Fitting", 0.214}, {"Digits of Pi", 0.194}, 
   {"Discrete Fourier Transform", 0.218}, {"Eigenvalues of a Matrix", 0.233}, 
   {"Elementary Functions", 0.173}, {"Gamma Function", 0.252}, 
   {"Large Integer Multiplication", 0.251}, {"Matrix Arithmetic", 0.11}, 
   {"Matrix Multiplication", 0.102}, {"Matrix Transpose", 0.218}, 
   {"Numerical Integration", 0.33}, {"Polynomial Expansion", 0.049}, 
   {"Random Number Sort", 0.393}, {"Singular Value Decomposition", 0.138}, 
   {"Solving a Linear System", 0.133}}}

3 revs · Accepted Answer · 2022-07-03 10:35:17Z

Dell Latitude 5411 15.5" Notebook - Full HD - 1920 x 1080 - Core i5 i5-10400H 10th Gen 2.6GHz Quad-core (4 Core) - 8GB RAM - 256GB SSD:


Processor	QuadCore Intel(R) Core(TM) i5-10400H CPU @ 2.60GHz
Chipsets	Intel Comet Point-H WM490, Intel Comet Lake-H
System Memory	7789 MB (DDR4 SDRAM)
Memory module	Micron 8ATF1G64HZ-3G2J1 8 GB (1 rank, 16 banks) DDR4-3200 (1600 MHz)
Storage	SSD Toshiba KIOXIA 256GB KXG60ZNV256G NVMe
OS Name	Microsoft Windows 10 Pro x64

{"System" -> "Microsoft Windows (64-bit)", "BenchmarkName" -> "WolframMark", 
 "FullVersionNumber" -> "13.0.1", "Date" -> "April 29, 2022", 
 "BenchmarkResult" -> 2.487, "TotalTime" -> 5.566, 
  "Results" -> {
   {"Data Fitting", 0.319}, 
   {"Digits of Pi", 0.271}, 
   {"Discrete Fourier Transform", 0.581},
   {"Eigenvalues of a Matrix", 0.308}, 
   {"Elementary Functions", 0.683}, 
   {"Gamma Function", 0.387},
   {"Large Integer Multiplication", 0.375}, 
   {"Matrix Arithmetic", 0.358}, 
   {"Matrix Multiplication", 0.308},
   {"Matrix Transpose", 0.703}, 
   {"Numerical Integration", 0.407}, 
   {"Polynomial Expansion", 0.049},
   {"Random Number Sort", 0.156}, 
   {"Singular Value Decomposition", 0.288}, 
   {"Solving a Linear System", 0.373}}}

{"System" -> "Microsoft Windows (64-bit)", "BenchmarkName" -> "WolframMark", 
 "FullVersionNumber" -> "13.1.0", "Date" -> "July 3, 2022", 
 "BenchmarkResult" -> 2.463`, 
 "TotalTime" -> 5.619`, "Results" -> {
   {"Data Fitting", 0.347`},
   {"Digits of Pi", 0.276`},
   {"Discrete Fourier Transform", 0.583`},
   {"Eigenvalues of a Matrix", 0.331`},
   {"Elementary Functions", 0.633`},
   {"Gamma Function", 0.381`},
   {"Large Integer Multiplication", 0.375`},
   {"Matrix Arithmetic", 0.374`},
   {"Matrix Multiplication", 0.296`},
   {"Matrix Transpose", 0.687`},
   {"Numerical Integration", 0.468`},
   {"Polynomial Expansion", 0.053`},
   {"Random Number Sort", 0.162`},
   {"Singular Value Decomposition", 0.288`},
   {"Solving a Linear System", 0.365`}}}

10 revs · Accepted Answer · 2023-03-08 02:23:32Z

CPU: i5-12500h
OS : Ubuntu 22.04
RAM : 16G
Score : 5.62

{"MachineName" -> "cvgmt", "System" -> "Linux x86 (64-bit)", 
 "BenchmarkName" -> "WolframMark", "FullVersionNumber" -> 
  "13.0.1", "Date" -> "May 15, 2022", 
 "BenchmarkResult" -> 5.62, "TotalTime" -> 2.463, 
 "Results" -> {{"Data Fitting", 0.153}, 
   {"Digits of Pi", 0.162}, {"Discrete Fourier Transform", 
    0.227}, {"Eigenvalues of a Matrix", 0.261}, 
   {"Elementary Functions", 0.12}, {"Gamma Function", 0.237}, 
   {"Large Integer Multiplication", 0.227}, 
   {"Matrix Arithmetic", 0.046}, {"Matrix Multiplication", 
    0.137}, {"Matrix Transpose", 0.173}, 
   {"Numerical Integration", 0.256}, {"Polynomial Expansion", 
    0.027}, {"Random Number Sort", 0.06}, 
   {"Singular Value Decomposition", 0.228}, 
   {"Solving a Linear System", 0.149}}}

LaunchKernels[]

{"MachineName" -> "12-node homogeneous cluster", 
 "System" -> "Linux-x86-64", "BenchmarkName" -> "WolframMark", 
 "FullVersionNumber" -> "13.0.1", "Date" -> "April 30, 2022", 
 "BenchmarkResult" -> 14.951, 
  "TotalTime" -> 33.33}

2 revs · Accepted Answer · 2023-03-08 03:10:22Z

Running 13th Gen Intel on 24 Cores. "BenchmarkResult" -> 4.309

For some reason I am doing worse than :

Apple M1 Max (10Cores) "BenchmarkResult" -> 4.65
Linux 12th Gen Intel(R) Core(TM) i7-12700K (12 Cores) "BenchmarkResult" -> 6.08

Any thoughts/tips?

Full details:

CPU: Processor 13th Gen Intel (R) Core (TM) i9 - 13900 KF, 3000 Mhz, 24 Core (s), 32 Logical Processor (s)

OS : Windows 11

RAM : 64GB

1 + 1;
Needs["Benchmarking`"]
Benchmark[]

{"MachineName" -> "IntroductionToProbability", "System" ->
"Microsoft Windows (64-bit)", "BenchmarkName" -> "WolframMark",  
"FullVersionNumber" -> "13.2.1", "Date" -> "March 8, 2023",
"BenchmarkResult" -> 4.309, "TotalTime" -> 3.212,   "Results" ->
{{"Data Fitting", 0.178}, {"Digits of Pi", 0.146}, {"Discrete Fourier
Transform", 0.209},     {"Eigenvalues of a Matrix", 0.391},
{"Elementary Functions", 0.139}, {"Gamma Function", 0.196},    
{"Large Integer Multiplication", 0.213}, {"Matrix Arithmetic", 0.507},
{"Matrix Multiplication", 0.076},     {"Matrix Transpose", 0.227},
{"Numerical Integration", 0.337}, {"Polynomial Expansion", 0.043},    
{"Random Number Sort", 0.08}, {"Singular Value Decomposition", 0.312},
{"Solving a Linear System", 0.158}}}

The running score of i7-12700K is the running score on Linux. — cvgmt, Mar 8 at 2:22
@cvgmt Thanks I will edit to clarify. Would you expect Linux to run fast than Windows? — IntroductionToProbability, Mar 8 at 3:09

Wicher · Accepted Answer · 2023-05-03 21:58:15Z

Core i7-6700 @3.4GHz, Windows 10, 32GB, MMA 13.1 Score 1.767

{"MachineName" -> "desktop-23at6mr", "System" -> "Microsoft Windows (64-bit)", "BenchmarkName" -> "WolframMark", "FullVersionNumber" -> "13.1.0", "Date" -> "May 3, 2023", "BenchmarkResult" -> 1.767, "TotalTime" -> 7.834, "Results" -> {{"Data Fitting", 0.94}, {"Digits of Pi", 0.414}, {"Discrete Fourier Transform", 0.809}, {"Eigenvalues of a Matrix", 0.785}, {"Elementary Functions", 0.517}, {"Gamma Function", 0.394}, {"Large Integer Multiplication", 0.397}, {"Matrix Arithmetic", 0.395}, {"Matrix Multiplication", 0.317}, {"Matrix Transpose", 0.56}, {"Numerical Integration", 1.213}, {"Polynomial Expansion", 0.141}, {"Random Number Sort", 0.193}, {"Singular Value Decomposition", 0.372}, {"Solving a Linear System", 0.387}}}

Alec Graves · Accepted Answer · 2023-07-29 23:48:00Z

Galaxy Book Pro 360

$Version 13.3.0 for Linux x86 (64-bit) (June 12, 2023)
11th Gen Intel(R) Core(TM) i7-1165G7 @ 2.80GHz
16 GB RAM
Parabola GNU/Linux-libre
kernel: 6.4.0-gnu-1-vanilla

Score: 5.163

Fresh kernel:

$ wolframscript -c 'Needs["Benchmarking`"]; Benchmarking`Benchmark[]'
{MachineName -> monad, System -> Linux x86 (64-bit), BenchmarkName -> WolframMark,
FullVersionNumber -> 13.3.0, Date -> July 29, 2023,
BenchmarkResult -> 5.163, TotalTime -> 2.681, 
Results -> {{Data Fitting, 0.152}, {Digits of Pi, 0.222}, {Discrete Fourier Transform, 0.161}, {Eigenvalues of a Matrix, 0.203}, {Elementary Functions, 0.111}, {Gamma Function, 0.324}, {Large Integer Multiplication, 0.309}, {Matrix Arithmetic, 0.061}, {Matrix Multiplication, 0.152}, {Matrix Transpose, 0.171}, {Numerical Integration, 0.308}, {Polynomial Expansion, 0.047}, {Random Number Sort, 0.116}, {Singular Value Decomposition, 0.176}, {Solving a Linear System, 0.168}}}

After LaunchKernels[] / LaunchKernels[8]:

$ wolframscript -c 'LaunchKernels[]; Needs["Benchmarking`"]; Benchmarking`Benchmark[]'
{MachineName -> 4-node homogeneous cluster, System -> Linux-x86-64, BenchmarkName -> WolframMark, FullVersionNumber -> 13.3.0, Date -> July 29, 2023,
BenchmarkResult -> 9.395, TotalTime -> 17.68}

$ wolframscript -c 'LaunchKernels[8]; Needs["Benchmarking`"]; Benchmarking`Benchmark[]'
{MachineName -> 8-node homogeneous cluster, System -> Linux-x86-64, BenchmarkName -> WolframMark, FullVersionNumber -> 13.3.0, Date -> July 29, 2023, 
BenchmarkResult -> 10.405, TotalTime -> 31.928}

3 revs · Accepted Answer · 2023-07-29 23:50:52Z

Wolfram Cloud

$Version 13.3.0 for Linux x86 (64-bit) (June 3, 2023)
2-core x86-64??
32 GB RAM
Linux x86 (64-bit)??
kernel??

$ wolframscript -c 'Print[ToString /@ {$ProcessorType, $ProcessorCount, $MachineType, $OperatingSystem, $System}]' -o
{x86-64, 2, PC, Unix, Linux x86 (64-bit)}

Score: 1.432

Fresh kernel:

$ wolframscript -o -c 'Needs["Benchmarking`"]; Benchmarking`Benchmark[]'
{MachineName -> wolframcloud-prd-cmp-4j-16,
System -> Linux x86 (64-bit),
BenchmarkName -> WolframMark,
FullVersionNumber -> 13.3.0,
Date -> July 29, 2023, 
BenchmarkResult -> 1.432, 
TotalTime -> 9.663, Results -> {{Data Fitting, 0.437}, {Digits of Pi, 0.39}, {Discrete Fourier Transform, 0.9}, {Eigenvalues of a Matrix, 0.486}, {Elementary Functions, 0.87}, {Gamma Function, 0.842}, {Large Integer Multiplication, 0.61}, {Matrix Arithmetic, 0.458}, {Matrix Multiplication, 0.843}, {Matrix Transpose, 1.267}, {Numerical Integration, 1.}, {Polynomial Expansion, 0.242}, {Random Number Sort, 0.341}, {Singular Value Decomposition, 0.499}, {Solving a Linear System, 0.478}}}

LaunchKernels[] not supported.

2 revs · Accepted Answer · 2023-07-29 23:51:07Z

Asus ROG Ally running Parabola linux

$Version 13.3.0 for Linux x86 (64-bit) (June 12, 2023)
AMD Ryzen Z1 Extreme
12 GB RAM
Parabola GNU/Linux-libre
kernel: 5.15.88-gnu-1-lts

Score: 5.233

Fresh Quit[] kernel:

{"MachineName" -> "monad", "System" -> "Linux x86 (64-bit)", "BenchmarkName" -> "WolframMark", "FullVersionNumber" -> "13.3.0", "Date" -> "July 28, 2023", "BenchmarkResult" -> 5.233, 
 "TotalTime" -> 2.645, "Results" -> {{"Data Fitting", 0.192}, {"Digits of Pi", 0.178}, {"Discrete Fourier Transform", 0.259}, {"Eigenvalues of a Matrix", 0.207}, 
   {"Elementary Functions", 0.194}, {"Gamma Function", 0.231}, {"Large Integer Multiplication", 0.21}, {"Matrix Arithmetic", 0.042}, {"Matrix Multiplication", 0.131}, 
   {"Matrix Transpose", 0.236}, {"Numerical Integration", 0.315}, {"Polynomial Expansion", 0.05}, {"Random Number Sort", 0.066}, {"Singular Value Decomposition", 0.158}, 
   {"Solving a Linear System", 0.176}}}

After LaunchKernels[]:

{"MachineName" -> "8-node homogeneous cluster", "System" -> "Linux-x86-64", "BenchmarkName" -> "WolframMark", "FullVersionNumber" -> "13.3.0", "Date" -> "July 28, 2023", 
 "BenchmarkResult" -> 10.783, "TotalTime" -> 30.809}

2 revs · Accepted Answer · 2023-07-29 23:51:31Z

Steam deck running parabola linux :)

$Version 13.3.0 for Linux x86 (64-bit) (June 12, 2023)
AMD Custom APU 0405
16GB RAM
Parabola GNU/Linux-libre
kernel: 5.15.88-gnu-1-lts

Score: 2.933

Fresh Quit[] kernel:

{"MachineName" -> "monad", "System" -> "Linux x86 (64-bit)", 
 "BenchmarkName" -> "WolframMark", "FullVersionNumber" -> "13.3.0", 
 "Date" -> "July 27, 2023", "BenchmarkResult" -> 2.933, "TotalTime" -> 4.72, 
 "Results" -> {{"Data Fitting", 0.279}, {"Digits of Pi", 0.244}, 
   {"Discrete Fourier Transform", 0.373}, {"Eigenvalues of a Matrix", 0.36}, 
   {"Elementary Functions", 0.382}, {"Gamma Function", 0.304}, 
   {"Large Integer Multiplication", 0.312}, {"Matrix Arithmetic", 0.137}, 
   {"Matrix Multiplication", 0.319}, {"Matrix Transpose", 0.441}, 
   {"Numerical Integration", 0.545}, {"Polynomial Expansion", 0.089}, 
   {"Random Number Sort", 0.158}, {"Singular Value Decomposition", 0.353}, 
   {"Solving a Linear System", 0.424}}}

After LaunchKernels[]:

{"MachineName" -> "4-node homogeneous cluster", "System" -> "Linux-x86-64", 
 "BenchmarkName" -> "WolframMark", "FullVersionNumber" -> "13.3.0", 
 "Date" -> "July 27, 2023", "BenchmarkResult" -> 4.578, "TotalTime" -> 36.287}

3 revs · Accepted Answer · 2023-07-31 05:08:20Z

$Version 13.3.0 for Microsoft Windows (64-bit) (June 3, 2023)
AMD Ryzen 7 4800H with Radeon Graphics 2.90 GHz
16GB RAM
Windows 11 Insider Preview 22H2
ASUS TUF Gaming A17 (2019)

Score: 2.77

Fresh Quit[] kernel:

{
  "MachineName" -> "tuf", 
  "System" -> "Microsoft Windows (64-bit)", 
  "BenchmarkName" -> "WolframMark", 
  "FullVersionNumber" -> "13.3.0", 
  "Date" -> "July 31, 2023", 
  "BenchmarkResult" -> 2.77, 
  "TotalTime" -> 4.998, 
  "Results" -> {
    {"Data Fitting", 0.435}, 
    {"Digits of Pi", 0.245}, 
    {"Discrete Fourier Transform", 0.452}, 
    {"Eigenvalues of a Matrix", 0.49}, 
    {"Elementary Functions", 0.545}, 
    {"Gamma Function", 0.3}, 
    {"Large Integer Multiplication", 0.283}, 
    {"Matrix Arithmetic", 0.195}, 
    {"Matrix Multiplication", 0.22}, 
    {"Matrix Transpose", 0.46}, 
    {"Numerical Integration", 0.615}, 
    {"Polynomial Expansion", 0.07}, 
    {"Random Number Sort", 0.109}, 
    {"Singular Value Decomposition", 0.272}, 
    {"Solving a Linear System", 0.307}
  }
}

After LaunchKernels[]:

{
  "MachineName" -> "8-node homogeneous cluster", 
  "System" -> "Windows-x86-64", 
  "BenchmarkName" -> "WolframMark", 
  "FullVersionNumber" -> "13.3.0", 
  "Date" -> "July 31, 2023", 
  "BenchmarkResult" -> 6.993, 
  "TotalTime" -> 47.503
}

3 revs · Accepted Answer · 2023-08-05 16:26:23Z

$Version 13.3.0 for Linux x86 (64-bit) (June 12, 2023)
13th Gen Intel(R) Core(TM) i9-13900K
32GB RAM
Parabola GNU/Linux-libre
kernel: 5.15.88-gnu-1-lts

Score: 8.435

Fresh Quit[] kernel:

{"MachineName" -> "monad", "System" -> "Linux x86 (64-bit)", 
 "BenchmarkName" -> "WolframMark", 
 "FullVersionNumber" -> "13.3.0", 
 "Date" -> "July 21, 2023", 
 "BenchmarkResult" -> 8.435, 
 "TotalTime" -> 1.641, 
 "Results" -> {{"Data Fitting", 0.104}, {"Digits of Pi", 0.131}, {"Discrete Fourier Transform", 0.171}, 
   {"Eigenvalues of a Matrix", 0.154}, {"Elementary Functions", 0.051}, {"Gamma Function", 0.184}, 
   {"Large Integer Multiplication", 0.175}, {"Matrix Arithmetic", 0.015}, {"Matrix Multiplication", 0.061}, 
   {"Matrix Transpose", 0.1}, {"Numerical Integration", 0.199}, {"Polynomial Expansion", 0.024}, 
   {"Random Number Sort", 0.023}, {"Singular Value Decomposition", 0.131}, {"Solving a Linear System", 0.118}}}

After LaunchKernels[] (8 kernels):

{"MachineName" -> "8-node homogeneous cluster",
 "System" -> "Linux-x86-64",
 "BenchmarkName" -> "WolframMark",
 "FullVersionNumber" -> "13.3.0",
 "Date" -> "July 21, 2023",
 "BenchmarkResult" -> 25.221,
 "TotalTime" -> 13.172}

Same machine running Windows 11 Pro:

$Version 13.3.0 for Microsoft Windows (64-bit) (June 12, 2023)

Score: 5.739

{"MachineName" -> "pandora", "System" -> "Microsoft Windows (64-bit)", "BenchmarkName" -> "WolframMark",
 "FullVersionNumber" -> "13.3.0", "Date" -> "July 30, 2023", "BenchmarkResult" -> 5.739, "TotalTime" -> 2.412,
 "Results" -> {{"Data Fitting", 0.176}, {"Digits of Pi", 0.144}, {"Discrete Fourier Transform", 0.2},
   {"Eigenvalues of a Matrix", 0.248}, {"Elementary Functions", 0.131}, {"Gamma Function", 0.193},
   {"Large Integer Multiplication", 0.203}, {"Matrix Arithmetic", 0.093}, {"Matrix Multiplication", 0.064},
   {"Matrix Transpose", 0.163}, {"Numerical Integration", 0.298}, {"Polynomial Expansion", 0.031},
   {"Random Number Sort", 0.074}, {"Singular Value Decomposition", 0.251}, {"Solving a Linear System", 0.143}}}

LaunchKernels[]

{"MachineName" -> "8-node homogeneous cluster", "System" -> "Windows-x86-64", "BenchmarkName" -> "WolframMark",
 "FullVersionNumber" -> "13.3.0", "Date" -> "July 30, 2023", "BenchmarkResult" -> 19.74, "TotalTime" -> 16.829}

Not sure why there is such a large discrepancy. I am guessing there are significant performance differences in the C runtime between Windows and Linux. Perhaps it comes down to the OS effectively allocating processes across the heterogeneous CPU architecture of the i9 as well.

Same computer running WSL2 (Ubuntu 22.04)

os: 22.04.2 LTS (Jammy Jellyfish)
kernel: 5.15.90.1-microsoft-standard-WSL2

Score: 7.335

$ wolframscript -c 'Needs["Benchmarking`"]; Benchmarking`Benchmark[]'
{MachineName -> pandora, System -> Linux x86 (64-bit), BenchmarkName -> WolframMark, FullVersionNumber -> 13.3.0, Date -> August 5, 2023, BenchmarkResult -> 7.335, TotalTime -> 1.887, Results -> {{Data Fitting, 0.14}, {Digits of Pi, 0.146}, {Discrete Fourier Transform, 0.149}, {Eigenvalues of a Matrix, 0.158}, {Elementary Functions, 0.078}, {Gamma Function, 0.219}, {Large Integer Multiplication, 0.182}, {Matrix Arithmetic, 0.027}, {Matrix Multiplication, 0.072}, {Matrix Transpose, 0.105}, {Numerical Integration, 0.238}, {Polynomial Expansion, 0.112}, {Random Number Sort, 0.031}, {Singular Value Decomposition, 0.118}, {Solving a Linear System, 0.112}}}

LaunchKernels[]:

$ wolframscript -c 'LaunchKernels[]; Needs["Benchmarking`"]; Benchmarking`Benchmark[]'
InputForm[{MachineName -> 8-node homogeneous cluster, System -> Linux-x86-64, BenchmarkName -> WolframMark, FullVersionNumber -> 13.3.0, Date -> August 5, 2023, BenchmarkResult -> 24.476, TotalTime -> 13.573}]

To summarize, for modern Intel processors in 2023 and version 13.3, Wolfram Language (WL) on native Linux performs at 1.5x WL on Windows 11, and WL on WSL2 performs at 1.3x WL on Windows 11. Also, WL under native Linux operates 15% faster than WL on WSL.

Nasser · Accepted Answer · 2023-05-03 22:34:36Z

0

window 10 pro, 128 GB RAM. Mathematica V 13.2.1

Final score is 4.857

Mathematica graphics

answered May 3 at 22:34

community wiki

Nasser

Add a comment |

Ytrewq · Accepted Answer · 2023-07-29 13:00:36Z

Amd 5900X, wsl2 ubuntu

$ wolframscript
Wolfram Language 13.3.0 Engine for Linux x86 (64-bit)
Copyright 1988-2023 Wolfram Research, Inc.

In[1]:= Needs["Benchmarking`"]
In[2]:= Benchmark[]
Out[2]//InputForm=
{"MachineName" -> "desktopxxxxxxx", "System" -> "Linux x86 (64-bit)", "BenchmarkName" -> "WolframMark",
 "FullVersionNumber" -> "13.3.0", "Date" -> "July 29, 2023", "BenchmarkResult" -> 5.659, "TotalTime" -> 2.446,
 "Results" -> {{"Data Fitting", 0.192}, {"Digits of Pi", 0.167}, {"Discrete Fourier Transform", 0.213},
   {"Eigenvalues of a Matrix", 0.24}, {"Elementary Functions", 0.101}, {"Gamma Function", 0.219},
   {"Large Integer Multiplication", 0.214}, {"Matrix Arithmetic", 0.025}, {"Matrix Multiplication", 0.083},
   {"Matrix Transpose", 0.21}, {"Numerical Integration", 0.324}, {"Polynomial Expansion", 0.081},
   {"Random Number Sort", 0.038}, {"Singular Value Decomposition", 0.169}, {"Solving a Linear System", 0.17}}}

In[3]:= LaunchKernels[]
In[4]:= Benchmark[]
Out[4]//InputForm=
{"MachineName" -> "8-node homogeneous cluster", "System" -> "Linux-x86-64", "BenchmarkName" -> "WolframMark",
 "FullVersionNumber" -> "13.3.0", "Date" -> "July 29, 2023", "BenchmarkResult" -> 12.738, "TotalTime" -> 26.081}

Navvye · Accepted Answer · 2023-07-31 06:22:14Z

On Apple M2 - 2023, 8GB RAM. Performs way better than I expected!

MachineName: amairahs-macbook-air
System: Mac OS X ARM (64-bit)
BenchmarkName: WolframMark
FullVersionNumber: 13.3.0
Date: July 31, 2023
BenchmarkResult: 4.197
TotalTime: 3.298

Results:
  {"Data Fitting", 0.153}
  {"Digits of Pi", 0.147}
  {"Discrete Fourier Transform", 0.268}
  {"Eigenvalues of a Matrix", 0.214}
  {"Elementary Functions", 0.6}
  {"Gamma Function", 0.191}
  {"Large Integer Multiplication", 0.166}
  {"Matrix Arithmetic", 0.083}
  {"Matrix Multiplication", 0.189}
  {"Matrix Transpose", 0.136}
  {"Numerical Integration", 0.272}
  {"Polynomial Expansion", 0.035}
  {"Random Number Sort", 0.355}
  {"Singular Value Decomposition", 0.275}
  {"Solving a Linear System", 0.214}

murray · Accepted Answer · 2023-07-31 19:18:35Z

Quit[]
Needs["Benchmarking`"]
Benchmark[]

 {"MachineName" -> "Mac Studio M1 Max 64GB RAM", 
  "System" -> "Mac OS X ARM (64-bit)", 
  "BenchmarkName" -> "WolframMark", 
  "FullVersionNumber" -> "13.3.0", 
  "Date" -> "July 31, 2023", 
  "BenchmarkResult" -> 4.583, "TotalTime" -> 3.02, 
  "Results" -> {{"Data Fitting", 0.183}, 
    {"Digits of Pi", 0.162}, 
    {"Discrete Fourier Transform", 0.226}, 
    {"Eigenvalues of a Matrix", 0.252}, 
    {"Elementary Functions", 0.298}, 
    {"Gamma Function", 0.209}, 
    {"Large Integer Multiplication", 0.181}, 
    {"Matrix Arithmetic", 0.055}, 
    {"Matrix Multiplication", 0.119}, 
    {"Matrix Transpose", 0.096}, 
    {"Numerical Integration", 0.359}, 
    {"Polynomial Expansion", 0.05}, 
    {"Random Number Sort", 0.378}, 
    {"Singular Value Decomposition", 0.284}, 
    {"Solving a Linear System", 0.168}}}

 Quit[]
 Needs["Benchmarking`"]
 LaunchKernels[]
 Benchmark[]

 {"MachineName" -> "8-node homogeneous cluster", 
  "System" -> "MacOSX-ARM64", "BenchmarkName" -> 
  "WolframMark", "FullVersionNumber" -> "13.3.0", 
  "Date" -> "July 31, 2023", 
  "BenchmarkResult" -> 14.612, "TotalTime" -> 22.736}

Jorge · Accepted Answer · 2023-09-23 00:05:50Z

I'm running Mathematica 13.3.1 in my new laptop and the performance is not as good as expected. Anyone have some suggestions?

Asus Zenbook 14x

13th Gen Intel(R) Core(TM) i7-13700H 2.40 GHz
16 GB RAM (15.6 GB usable)
Windows 11 Home (64-bit)

In a fresh kernel:

Score 2.576

{"MachineName" -> "perlanegra", "System" -> "Microsoft Windows (64-bit)", "BenchmarkName" -> "WolframMark", 
 "FullVersionNumber" -> "13.3.1", "Date" -> "September 22, 2023", "BenchmarkResult" -> 2.576, "TotalTime" -> 5.374, 
 "Results" -> {{"Data Fitting", 0.328}, {"Digits of Pi", 0.461}, {"Discrete Fourier Transform", 0.338}, 
   {"Eigenvalues of a Matrix", 0.545}, {"Elementary Functions", 0.288}, {"Gamma Function", 0.328}, 
   {"Large Integer Multiplication", 0.302}, {"Matrix Arithmetic", 0.172}, {"Matrix Multiplication", 0.153}, 
   {"Matrix Transpose", 0.347}, {"Numerical Integration", 0.56}, {"Polynomial Expansion", 0.114}, {"Random Number Sort", 0.222}, 
   {"Singular Value Decomposition", 0.409}, {"Solving a Linear System", 0.807}}}

Although, in the multicore running the things gets better.

Score 12.415

{"MachineName" -> "14-node homogeneous cluster", "System" -> "Windows-x86-64", "BenchmarkName" -> "WolframMark", 
 "FullVersionNumber" -> "13.3.1", "Date" -> "September 22, 2023", "BenchmarkResult" -> 12.415, "TotalTime" -> 46.827}

Stack Exchange Network

Benchmarking Mathematica 13 across machines

21 Answers 21

Your Answer

Not the answer you're looking for? Browse other questions tagged
system-performance
version-13
benchmarking
or ask your own question.

Linked

Hot Network Questions

Benchmarking Mathematica 13 across machines

21 Answers 21

Your Answer

Sign up or log in

Post as a guest

Not the answer you're looking for? Browse other questions tagged system-performanceversion-13benchmarking or ask your own question.

Linked

Related

Hot Network Questions

Not the answer you're looking for? Browse other questions tagged
system-performance
version-13
benchmarking
or ask your own question.