Recommended CPU for timetabling

[y]

4min 15s - AMD Athlon 5350 (2.05 GHz - precompiled FET 5.14.3 with Windows 7 32bit - metered by y)

Lousy performance for a 4-core processor! It must be because FET uses only one core all of the time. Compared with the Intel Celeron G1840 2.80 GHz, it looks awful, but for tasks using all processors (e.g. mining), the AMD performs better.

[Volker Dirr]

Yes, FET use single core only. I will add your other relsut also into the first post now. Thank you for report.

[Volker Dirr]

I added a new result:
9min 08s - Broadcom BCM2837 Raspberry Pi 3 - ARM Cortex-A53 (1.2 GHz - Raspbian - metered by Volker Dirr)

[Benahmed Abdelkrim]

Sorry for french Words.
   Je suis d'accord avec mr Volker, j'ajoute que le plus important avec FET, ce n'est pas le nombre de core CPU, mais c'est le cache mémoire, la ram et la fréquence de l'horloge (exprimé en GHz ). En plus de des activités placées en premier par fet.
  Donc en résumé le temps met par fet pour produire un tableau dépend de deux choses:
    1- les caractéristiques citées plus haut.
    2- le phénomène aléatoire du placement des activités.
Ce dernier point en peut le vérifier facilement; avec le même tableau, en aura des temps de productions déférents!


I agree with mr Volker , I add that the most important with FET , it is not the number of core CPUs , but the cache, the ram and the clock frequency (expressed in GHz). In addition to the activities placed first by fet .
  So in short time places per fect for an array depends on two things :
    1- the characteristics listed above.
    2- the random phenomenon of investment activities .
This last point can be checked easily ; with the same array, will have time to deferent productions!

[Volker Dirr]

Hi

anyone with an AMD Ryzen out there? I am interested in the results of that CPU (see first message of this topic). Maybe you can do a benchmark? Of course all other CPUs are also interesting. Let me know if you have got new results.

Thank you!

[Volker Dirr]

I added 2 new results in the first message of this topic:
1min 34s - AMD Ryzen 1800x (3.60GHz - precompiled - Windows 10 - metered by tobse2056)
1min 40s - Intel i5 4460 (3.20 GHz - precompiled FET with Windows 10  - metered by Volker Dirr)

Please let me now your results. (Also intresting will be AMD Ryzen 7 1700, Intel Atom, Intel Celeron, Intel Pentium, ...)

[Volker Dirr]

3 new results by Rumtata:
(AMD Phenom II X4 955 @3,6GHz- Windows 10- 2m 2s)
(Intel Xeon E3 1230 v3 @3,7GHz Windows 10 - 1m 28s)
(Dell Venue 8 Pro [Tablet] Intel Atom Z3735G@1,33GHz-Windows 8- 7m 12s) battery only (without cable power supply)

Thank you.
Please check the first message in this topic to see the full list.

[Volker Dirr]

I added 3 new results in the first message of this topic:
1min 05s - Intel i7 4770k, 4.8 GHz, precompiled, Windows 10, Addi
1min 29s - Intel i3 6100, 3.7GHz, precompiled, Windows10, Salatsauce45
1min 50s - Intel Xeon E5 1620v2, 3.7 Ghz, precompiled, Windows 7, DarkWing13

by the way:
a big thank you to the guys from the "pcgameshardware"-forum.
( http://extreme.pcgameshardware.de/prozessoren/473676-ryzen-besiter-bitte-um-einen-kurzen-benchmark.html )

[Volker Dirr]

I added a new result in the first message of this topic:
1min 34s - Intel Pentium G4560, 3.5 GHz, precompiled, Windows 10

[Volker Dirr]

i added a new result in the first post of this topic:
1min 41s - AMD Ryzen 1600, 3.2 GHz,  precompiled, Windows 10, Liviu Lalescu

[Volker Dirr]

I added a new result with the enhanced Pi 3 (See first post of this topic). It was released 6 month ago. In my opinion the cheapest way to make timetables, since you need to pay only ~35$ for this single-board computer. (Board + CPU + GPU + RAM)

7min 45s - Broadcom BCM2837B Raspberry Pi 3+ - ARM Cortex-A53 (1.4 GHz - Raspbian - metered by Volker Dirr)

[Volker Dirr]

New Raspberry Pi 4 just released. Anyone bought it and can run the benchmark? I guess it will be around 20% up to 2 times faster.
Also the new Ryzen 3xxx with Zen 2 will be intresting (will be released 7.7.). Hope someone can run a test with that also.
Of course all other CPUs are also intresting. Let me know your results.

[Volker Dirr]

I added new results of an Ryzen 2700x in the first post. Metered by Liviu.
He measured it with the old precompiled windows version and a new Linux version (so 10 years newer compiler). So the Linux version is much faster.

hmm...
It is maybe time to think about a new benchmark and/or separate the results into different tables. There are a lot of reasons for a new benchmark. The precompiled version is pretty old (old compiler). There is at least one big improvement in the algorithm that might give much better results for CPUs with smaller cache size (removing the foreach stuff some years ago). We also should maybe care about a multi core test. Also the benchmark looks too easy. It should be more difficult.
Let me know your ideas about a new benchmark.

But as long as there is no new benchmark version:
I am still interested in other CPUs. Especially AMD Ryzen 3xxx and Raspberry Pi 4.

[Volker Dirr]

I just added results for the new raspberry pi 4:
4min 36s - Broadcom BCM2711, Raspberry Pi 4, ARM Cortex-A72, 1.5 GHz, Raspbian, Volker Dirr

[cipman]

Intel i5-9400F @ 2.90 GHz, on Ubuntu 19.04, FET 5.14.3 compiled from source --> 56 seconds from 10 trial runs (1 run was 55 secs) - printscreen at https://i.imgur.com/ebU0EGD.png

Not as fast as I've expected (45-48 secs) but I chose this CPU for a silent PC configuration and it is dead silent running 6 FET instances on all 6 cores with CPU temp at 68-70 deg C in 25 deg C ambient. So I'm more than happy with the result.

Looking forward for a Ryzen 5 3600 speed test (although I'm not an AMD fan - yet).