Defining : name 'BenchmarkToools' - how?

Specific Domains Optimization (Mathematical)
1

Hi,
newbie trying to understand how to define the ‘BenchmarkToools’ ?
I installed ‘BenchmarkToools’ with using/add Pkg.
now in jupyter notebook : getting msg:
NameError: name ‘BenchmarkToools’ is not defined
So does one have to set a variable like b = BenchmarkToools(x) or how does this work ?
Thanks

2

Is it simply that you are misspelling the name, with three ‘o’ characters instead of two? Try rerunning with
using BenchmarkTools.

2 Likes
3

Hi @John_Gibson,
correct, it was misspeled in the above line - oops thanks -
now entering BenchmarkTools, it just outputs BenchmarkTools…
so, where to go from there to measure some script time exec ?

4

I get results using @time …but not using @benchmark … :face_with_raised_eyebrow:

5

I’ll point you to the Benchmarks manual and suggest that you post minimal working examples of any issues you are having. This PSA explains how and why.
https://discourse.julialang.org/t/psa-make-it-easier-to-help-you/14757/81

3 Likes
6 
julia> using BenchmarkTools

julia> @benchmark randn(1000)
BenchmarkTools.Trial: 
  memory estimate:  7.94 KiB
  allocs estimate:  1
  --------------
  minimum time:     2.196 μs (0.00% GC)
  median time:      2.470 μs (0.00% GC)
  mean time:        2.590 μs (3.68% GC)
  maximum time:     76.003 μs (92.82% GC)
  --------------
  samples:          10000
  evals/sample:     9
7

oh, of course @John_Gibson , I was so into the basic course that I forgot to mention my code : … :see_no_evil: :speak_no_evil:

a = rand(10^7)
sum(a)
@time sum(a)

I think I got it right now with the added line, thanks to the doc you’ve just provided me, Great thanks for that ! have a great end-of-Week-end ! :hugs:

@benchmark sum(a)       (had tried s/t else that was not working for sure)

its all working fine now…thanks guys

2 Likes
8

I’ll test it with rand function too…good idea !

9

Glad to hear you are up and running!

1 Like
10 

BenchmarkTools.Trial: 
  memory estimate:  7.94 KiB
  allocs estimate:  1
  --------------
  minimum time:     1.865 μs (0.00% GC)
  median time:      2.065 μs (0.00% GC)
  mean time:        2.892 μs (8.88% GC)
  maximum time:     227.362 μs (98.09% GC)
  --------------
  samples:          10000
  evals/sample:     10
11

just in the middle of the Parallel julia ac. course but hope to be able to use all Pc Cores for my future scripts… :upside_down_face:

12

Just to add that the closest equivalent to time in BenchmarkTools is @btime, which will just output the runtime and memory allocations.

2 Likes
13

Hi wisers…

  • While with Julia Konsole 1

returns 2 threads

…when restarting jupyter from Konsole2 after that

Threads.nthreads()

still returns 1 thread.… :speak_no_evil:

How does one have Jupyter to take this julia modif into account ?
thanks

14

This is quite a different question from the one this thread is about - it’s usually preferable to ask separate questions to make things easier to find for future users.

In this case, someone already did start a thread about this specific question: Enable multiple Cores for Jupyter Lab - #7 by Gaussia

1 Like