ftbench: last updates

[Ahmet Göksu]

Hi,
It is a late message, sorry for this.

BTW, have you checked whether replacing `CLOCK_REALTIME` with 
`CLOCK_MONOTONIC` gives better results? 

What platform do you actually develop on? What #if clause in 
function `get_time` of `ftbench.c` is run on your computer? 

I tried to change the CLOCK_MONOTONIC but my system is not even run this if clause. I am running on a mac. It is runnning only this part

...
#else    /* clock() accuracy has improved since glibc 2.18 */    return 1E6 * (double)clock() / (double)CLOCKS_PER_SEC;#endif /* _POSIX_TIMERS */

--

With 'bulk test' I mean that you don't time single iterations but do timings for 10 iterations in a group, say, thus avoiding issues with the granularity of the OS timing functions. 

I already changed the working way of the timing. I only start the benchmark at beginning and stop at the end.

Please separate this line slightly from the rest of the table
and print the *cumulated timing* (in 's', not 'µs') of all tests,
something like
 I wouldn't use '**' – this looks like a footnote, and people start 
searching where it belongs to. Instead, please use a simple 
itemization. 

* I suggest to rename column 'N' to 'Iterations' or '#Iterations' (or 
something similar). 

* The sentence about '(X | Y') is cryptic. Maybe simply start with 

done!

The links for `Baseline` and `Benchmark` to the `.txt` files are 
absolute and thus not portable. They must be relative. 

I didnt get this clear. Since objs dir and the source dir different, I needed to give abs path.

Best,

Goksu

goksu.in

Freetype Benchmark Results

Warning: Baseline and Benchmark have the same commit ID!

Info

Info	Baseline	Benchmark
Parameters	-c 1000 -w 100	-c 1000 -w 100
Commit ID	267f1914	267f1914
Commit Date	2023-08-30 18:15:08 +0300	2023-08-30 18:15:08 +0300
Branch	GSoC-2023-Ahmet	GSoC-2023-Ahmet

* Average time for single iteration. Smaller values are better.

If two values are given in the 'Iterations' column, those are belong to baseline and benchmark in order.

Total Results

Test	Iterations	* Baseline (µs)	* Benchmark (µs)	Difference (%)
Load	100000	635.3	591.8	6.9
Load_Advances (Normal)	100000	593.4	608.9	-2.6
Load_Advances (Fast)	100000	2.2	2.2	1.8
Load_Advances (Unscaled)	100000	2.1	2.0	1.0
Render	100000	933.3	935.3	-0.2
Get_Glyph	100000	677.8	679.8	-0.3
Get_Char_Index	94000	2.0	2.0	-1.0
Iterate CMap	1000	1.3	1.3	1.5
New_Face	1000	51.3	48.0	6.5
Embolden	100000	749.5	749.5	0.0
Stroke	58550 | 57600	2555.1	2567.4	-0.5
Get_BBox	100000	555.2	554.0	0.2
Get_CBox	100000	679.4	680.6	-0.2
New_Face & load glyph(s)	100000	112.7	111.4	1.1
Total duration for all tests:	80 s

Results for Roboto_subset.ttf

Test	Iterations	* Baseline (µs)	* Benchmark (µs)	Difference (%)
Load	120000	486.6	485.6	0.2
Load_Advances (Normal)	120000	489.4	485.1	0.9
Load_Advances (Fast)	120000	2.6	2.6	0.0
Load_Advances (Unscaled)	120000	2.4	2.4	0.0
Render	120000	885.6	881.9	0.4
Get_Glyph	120000	593.6	589.6	0.7
Get_Char_Index	94000	1.9	1.9	0.0
Iterate CMap	1000	1.3	1.3	0.0
New_Face	1000	44.3	41.6	6.1
Embolden	120000	640.0	639.3	0.1
Stroke	60000	2335.8	2323.6	0.5
Get_BBox	120000	148.7	138.6	6.8
Get_CBox	120000	594.0	590.8	0.5
New_Face & load glyph(s)	120000	114.3	104.7	8.4
Total duration for the font:	14 s

Results for Arial_subset.ttf

Test	Iterations	* Baseline (µs)	* Benchmark (µs)	Difference (%)
Load	95000	892.1	713.2	20.1
Load_Advances (Normal)	95000	753.4	788.9	-4.7
Load_Advances (Fast)	95000	2.0	2.1	-5.0
Load_Advances (Unscaled)	95000	1.9	1.9	0.0
Render	95000	1057.4	1037.1	1.9
Get_Glyph	95000	812.5	779.6	4.0
Get_Char_Index	94000	2.4	2.4	0.0
Iterate CMap	1000	1.4	1.3	7.1
New_Face	1000	65.3	51.8	20.7
Embolden	95000	874.6	835.5	4.5
Stroke	57000	3458.1	3425.7	0.9
Get_BBox	95000	764.6	719.4	5.9
Get_CBox	95000	816.4	766.2	6.1
New_Face & load glyph(s)	95000	110.7	111.9	-1.1
Total duration for the font:	18 s

Results for TimesNewRoman_subset.ttf

Test	Iterations	* Baseline (µs)	* Benchmark (µs)	Difference (%)
Load	95000	836.4	821.0	1.8
Load_Advances (Normal)	95000	818.6	825.4	-0.8
Load_Advances (Fast)	95000	2.1	2.0	4.8
Load_Advances (Unscaled)	95000	1.9	1.9	0.0
Render	95000	1200.3	1188.2	1.0
Get_Glyph	95000	900.1	906.8	-0.7
Get_Char_Index	94000	1.8	1.8	0.0
Iterate CMap	1000	1.3	1.3	0.0
New_Face	1000	53.2	52.8	0.8
Embolden	95000	1024.7	1031.6	-0.7
Stroke	42750	1036.1	1042.9	-0.7
Get_BBox	95000	882.8	883.7	-0.1
Get_CBox	95000	905.2	908.9	-0.4
New_Face & load glyph(s)	95000	135.7	133.9	1.3
Total duration for the font:	20 s

Results for Tahoma_subset.ttf

Test	Iterations	* Baseline (µs)	* Benchmark (µs)	Difference (%)
Load	95000	468.9	510.9	-9.0
Load_Advances (Normal)	95000	467.9	516.5	-10.4
Load_Advances (Fast)	95000	2.2	2.1	4.5
Load_Advances (Unscaled)	95000	2.0	1.9	5.0
Render	95000	769.0	827.5	-7.6
Get_Glyph	95000	558.6	607.6	-8.8
Get_Char_Index	94000	1.8	1.9	-5.6
Iterate CMap	1000	1.3	1.3	0.0
New_Face	1000	46.8	47.6	-1.7
Embolden	95000	625.4	665.5	-6.4
Stroke	66500 | 61750	3020.9	3125.7	-3.5
Get_BBox	95000	509.9	557.3	-9.3
Get_CBox	95000	560.1	613.1	-9.5
New_Face & load glyph(s)	95000	101.2	104.5	-3.3
Total duration for the font:	14 s

Results for Verdana_subset.ttf

Test	Iterations	* Baseline (µs)	* Benchmark (µs)	Difference (%)
Load	95000	492.5	428.1	13.1
Load_Advances (Normal)	95000	437.8	428.8	2.1
Load_Advances (Fast)	95000	2.3	2.2	4.3
Load_Advances (Unscaled)	95000	2.1	2.1	0.0
Render	95000	754.2	741.8	1.6
Get_Glyph	95000	524.4	515.6	1.7
Get_Char_Index	94000	1.9	1.9	0.0
Iterate CMap	1000	1.3	1.3	0.0
New_Face	1000	46.8	46.0	1.7
Embolden	95000	583.0	575.8	1.2
Stroke	66500	2924.4	2919.1	0.2
Get_BBox	95000	469.9	471.0	-0.2
Get_CBox	95000	521.2	523.9	-0.5
New_Face & load glyph(s)	95000	101.4	101.9	-0.5
Total duration for the font:	14 s

Freetype Benchmark