ARM tinyBench Timings

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

ARM tinyBench Timings

KenDickey
 
Many thanks again to all for getting Cuis Smalltalk running on CogSpur and to all the stellar VM great work !!

A couple of "quick-look" measurements using tynyBenchmarks

=======================
ACER Chromebook CB5-311
NVIDIA Tegra K1 Quad-Core ARM Cortex-A15 "r3"
2.3 GHz

CogSpur VM
 Squeak 5.1
      980,000,000 bytecodes/sec; 66,000,000 sends/sec
 
 Cuis 2917
    1,083,024,854 bytecodes/sec; 64,289,750 sends/sec

Stack VM
      258,847,320 bytecodes/sec; 13,014,759 sends/sec

======================================
C.H.I.P. $9 Computer [Next Thing Co]
Allwinner R8 Single-Core ARM Cortex-A8
1 GHz

tinyBenchmarks for CHIP:
CogSpur  253,716,551 bytecodes/sec; 16,853,816 sends/sec
Stack     55,315,471 bytecodes/sec;  3,348,667 sends/src


CogSpur is definitely worthwhile for the ARM.  :)


Thanks again to all for doing this !!!
-KenD
-KenD
Reply | Threaded
Open this post in threaded view
|

Re: ARM tinyBench Timings

timrowledge


> On 25-08-2016, at 6:58 AM, KenD <[hidden email]> wrote:
>
>
> Many thanks again to all for getting Cuis Smalltalk running on CogSpur and to all the stellar VM great work !!
>
> A couple of "quick-look" measurements using tynyBenchmarks
>
> =======================
> ACER Chromebook CB5-311
> NVIDIA Tegra K1 Quad-Core ARM Cortex-A15 "r3"
> 2.3 GHz
>
> CogSpur VM
> Squeak 5.1
>      980,000,000 bytecodes/sec; 66,000,000 sends/sec

Nice - about 4X my Pi3. Wonder how it fares with the bigger benchmarks.

>
>
> ======================================
> C.H.I.P. $9 Computer [Next Thing Co]
> Allwinner R8 Single-Core ARM Cortex-A8
> 1 GHz
>
> tinyBenchmarks for CHIP:
> CogSpur  253,716,551 bytecodes/sec; 16,853,816 sends/sec

Similar to a Pi3.


tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
"#define QUESTION ((bb) || !(bb))  - Shakespeare."


Reply | Threaded
Open this post in threaded view
|

Re: ARM tinyBench Timings

timrowledge
In reply to this post by KenDickey

I was about to recommend that Ken try the shootout benchmarks as a bigger and better performance test and I find that I can’t find a URL to offer him!

Where on earth should we find the code?

tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
Useful random insult:- One chip short of a cookie.


Reply | Threaded
Open this post in threaded view
|

Re: ARM tinyBench Timings

timrowledge


> On 25-08-2016, at 7:03 PM, tim Rowledge <[hidden email]> wrote:
>
> I was about to recommend that Ken try the shootout benchmarks as a bigger and better performance test and I find that I can’t find a URL to offer him!
>
> Where on earth should we find the code?

Ah-hah! Found it at last.

In repository
MCHttpRepository
        location: 'http://source.squeak.org/VMMaker'
        user: 'squeak'
        password: ‘squeak'

package CogBenchmarks-eem.1


tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
Strange OpCodes: IA: Illogical And


Reply | Threaded
Open this post in threaded view
|

Re: ARM tinyBench Timings

timrowledge
In reply to this post by timrowledge

Ken ran the shootout benchmarks on his Acer ARM 2.1GHz chromebook, I ran them on my 3.4GHz i7 iMac and 1.2GHz Pi 3.

Pi3 (actual fresh run is almost twice as fast as last time I recorded numbers!)
nbody 36.8
btree 25.5
redux 42.8
threadring 58.2

Acer ARM chromebook
8
8.6
18
19

iMac i7
3.2
2.1
4.4
6.5

So roughly speaking a Pi3 is 10% of iMac performance, acer c’book is ~30%. Not bad!

tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
Useful random insult:- A room temperature IQ.