dhrystone and whetstone benchmarks

Post here first, or if you can't find a relevant section!
User avatar
Pito
Posts: 94
Joined: Tue Dec 24, 2019 1:53 pm

Re: dhrystone and whetstone benchmarks

Post by Pito »

I have got BlackF407 board with 407ZET (144pin) on it (plus 512kB 10ns sram soldered in).
I do upload via J-Link (clone) with J-flash utility (set to SWD @4MHz).
Then I have to unplug/plug the USB cable, the RST does not start the stuff after upload..
It seems USB code needs some fix.
Last edited by Pito on Sun Jan 26, 2020 10:35 am, edited 1 time in total.
Pukao Hats Cleaning Services Ltd.
Bingo600
Posts: 86
Joined: Sat Dec 21, 2019 3:56 pm

Re: dhrystone and whetstone benchmarks

Post by Bingo600 »

Black Ali F407
F407VE-Black.png
F407VE-Black.png (57.37 KiB) Viewed 5129 times

Code: Select all

##########################################
Single Precision C Whetstone Benchmark
Calibrate
       0.10 Seconds          1   Passes (x 100)
       0.51 Seconds          5   Passes (x 100)
       2.57 Seconds         25   Passes (x 100)

Use 97  passes (x 100)

          Single Precision C/C++ Whetstone Benchmark
Loop content                  Result              MFLOPS      MOPS   Seconds

N1 floating point     -1.12475013732910156        81.327              0.023
N2 floating point     -1.12274742126464844        57.431              0.227
N3 if then else        1.00000000000000000                 167.325    0.060
N4 fixed point        12.00000000000000000                 315.000    0.097
N5 sin,cos etc.        0.49909299612045288                   1.292    6.246
N6 floating point      0.99999982118606567        59.255              0.883
N7 assignments         3.00000000000000000                  71.990    0.249
N8 exp,sqrt etc.       0.75110614299774170                   1.654    2.181

MWIPS                                             97.332              9.966
Bingo600
Posts: 86
Joined: Sat Dec 21, 2019 3:56 pm

Re: dhrystone and whetstone benchmarks

Post by Bingo600 »

Pito wrote: Sun Jan 26, 2020 10:27 am I have got BlackF407 board with 407ZET (144pin) on it (plus 512kB 10ns sram soldered in).
I do upload via J-Link (clone) with J-flash utility (set to SWD @4MHz).
Then I have to unplug/plug the USB cable, the RST does not start the stuff after upload..
It seems USB code needs some fix.
I have Jlink's too ;)
But mostly i use ST-Linkv2 or BMP , BMP is so easy (I'd expect even more on Win)

I made a BMP out of a BluePill here
https://www.eevblog.com/forum/microcont ... msg2823502

Binary
https://www.eevblog.com/forum/microcont ... msg2823454

BMP SWD pinout
https://www.eevblog.com/forum/microcont ... 7954;image


Try it out ;)

Though i agree the J-link is nice ....

/Bingo
User avatar
Pito
Posts: 94
Joined: Tue Dec 24, 2019 1:53 pm

Re: dhrystone and whetstone benchmarks

Post by Pito »

When looking at the historical whetstone result data - the big CPUs get higher MWIPS, but much lower MFLOPs at the same clock.

Also I do not understand you got 100 MWIPS @96MHz (411) and @168MHz (407) we get 98.
There must be a subtle bug in the code somewhere.
The historical results show nice linear dependency MWIPS on clock.


I do assume we should see something like 150 MWIPS @168MHz (407).

I would suggest small changes in the code, wait..

In the original code it was double

Code: Select all

double  theseSecs = 0.0;
double  startSecs = 0.0;
double  secs = 0.0;
Also you may use micros() to get better resolution

Code: Select all

void getSecs()
{
  theseSecs =  micros() / 1000000.0;
  return;
}
I doubt it helps, however, you may try..

PS: is the ART enabled in the STM core ????
Pukao Hats Cleaning Services Ltd.
User avatar
fpiSTM
Posts: 1746
Joined: Wed Dec 11, 2019 7:11 pm
Answers: 91
Location: Le Mans
Contact:

Re: dhrystone and whetstone benchmarks

Post by fpiSTM »

For DISCO F407, I think yes.
https://github.com/stm32duino/Arduino_C ... t.cpp#L191

For the black no.
Bingo600
Posts: 86
Joined: Sat Dec 21, 2019 3:56 pm

Re: dhrystone and whetstone benchmarks

Post by Bingo600 »

fpiSTM wrote: Sun Jan 26, 2020 11:20 am For DISCO F407, I think yes.
https://github.com/stm32duino/Arduino_C ... t.cpp#L191

For the black no.
@fpiSTM

My disco F407 is old --- Like bought within 1yr from 1'st release ... Prob no ART in that one.
Also makes the timings consistent w. My Disco & Pito's Black F407

Where do i find the PLL settings for the PILLF401 - I have no idea what speed my F411 (selected as F401) is running ??

/Bingo

One could use a few support routines ..

int Board_Init_Parms(int *HSx_Speed; int HSE_Enabled, int *PLL_Speed, int *ART_Enabled) or the likes ....
Maybe passing a struct would be better.

Interesting things could be :
What clocktype are we running : Direct , HSE , HSI
What is the PLL Speed in MHz (0) for No PLL
What ART level are we running ??
ABP1 / ABP2

Well give me all :-)
STM32-Clocks.png
STM32-Clocks.png (76.51 KiB) Viewed 5113 times
/Bingo
/Bingo
User avatar
fpiSTM
Posts: 1746
Joined: Wed Dec 11, 2019 7:11 pm
Answers: 91
Location: Le Mans
Contact:

Re: dhrystone and whetstone benchmarks

Post by fpiSTM »

All clock config are in the variant, in SystemClock_Config.
Bingo600
Posts: 86
Joined: Sat Dec 21, 2019 3:56 pm

Re: dhrystone and whetstone benchmarks

Post by Bingo600 »

Seems like the PILLF401 runs 84MHz

WUTTT ... I Can't upload a 98KB image
Is there a limit of 100K on pics ??
That must be raised a bit ...

Code
https://github.com/stm32duino/Arduino_C ... ariant.cpp

Clk84
PILLF4xx-84MHz.png
PILLF4xx-84MHz.png (64.79 KiB) Viewed 5111 times
Clk96
PILLF4xx-96MHz.png
PILLF4xx-96MHz.png (65.53 KiB) Viewed 5111 times
Can the F4 run w. 2 Flash waitstates @96MHz (I'm lazy) ;)

/Bingo
User avatar
Pito
Posts: 94
Joined: Tue Dec 24, 2019 1:53 pm

Re: dhrystone and whetstone benchmarks

Post by Pito »

All F4xx ever have got the ART.
It has to be enabled, then it does not matter how many ws you have (the ART accelerate the flash access).
You may overclock to 2ws @96MHz, it may or may not work.
Pukao Hats Cleaning Services Ltd.
Bingo600
Posts: 86
Joined: Sat Dec 21, 2019 3:56 pm

Re: dhrystone and whetstone benchmarks

Post by Bingo600 »

Pito wrote: Sun Jan 26, 2020 12:11 pm All F4xx ever have got the ART.
It has to be enabled, then it does not matter how many ws you have (the ART accelerate the flash access).
You may overclock to 2ws @96MHz, it may or may not work.
Seems like OLD F4's doesn't , hence the check

https://github.com/stm32duino/Arduino_C ... #L188-L193

/Bingo
Post Reply

Return to “General discussion”