String::format benchmark results

Vicente Romero vicente.romero at oracle.com
Wed Mar 27 02:40:18 UTC 2019


This version [1] includes two new tables for floats, the numbers are not 
so great for them. There was also a SOE during execution of test 
testFormat0080float0080String, see last row of `Table 7`, that's why 
there is no information in the last half of the table. I will 
investigate the reason for the error,

Vicente

[1] 
http://cr.openjdk.java.net/~vromero/string_format_benchmark_results/v3/benchmarkResults_intrinsics_string_format_v3.html

On 3/26/19 3:01 PM, Vicente Romero wrote:
> Hello,
>
> I have published another version [1], same benchmark, same builds, 
> this one include the standard errors of the measurements. The 
> variability of some of the data is significantly large (46% in one case)
>
> Thanks,
> Vicente
>
> [1] 
> http://cr.openjdk.java.net/~vromero/string_format_benchmark_results/v2/benchmarkResults_intrinsics_string_format_v2.html
>
> On 3/25/19 6:06 PM, Vicente Romero wrote:
>> Hi all,
>>
>> It seems like I overwrote the benchmark for String::format, thanks 
>> Alex, please check the results at [1]
>>
>> Thanks,
>> Vicente
>>
>> [1] 
>> http://cr.openjdk.java.net/~vromero/string_format_benchmark_results/v1/benchmarkResults_intrinsics_string_format_v1.html
>



More information about the amber-dev mailing list