It would be helpful to give more information as to why you say "This
doesn't work"?
But, I'm guessing that you are not seeing a decline in times when using
StartTimer/StopTimer.
It is likely that is because the copy function is fast, and frequent calls
to StartTimer/StopTimer involve some error/overhead. So they are
counteracting each other. Note that the documentation says the StopTimer
can be used for "complex initialization", and I am guessing that a 1k copy
does not count as complex. If you change the buffer size from 1024 to
1024*16 then you will see that the StartTimer/StopTimer version is in fact
faster.
PeterGo's solution works fine, as does bench marking a larger buffer. But
if your goal is to compare BenchmarkFilter3 and BenchmarkFilter4, then you
could reasonably just ignore the the overhead, since it will be the same
for both functions.
On Wednesday, July 7, 2021 at 11:29:30 AM UTC-4 tapi...@gmail.com wrote:
> This doesn't work:
>
> func BenchmarkFilter3(b *testing.B) {
> data := buildOrginalData()
> data2 := make([]int, len(data))
> b.ResetTimer()
> for i := 0; i < b.N; i++ {
> b.StopTimer()
> copy(data2, data)
> b.StartTimer()
> _ = Filter3(data2)
> }
> }
>
> On Wednesday, July 7, 2021 at 11:28:13 AM UTC-4 tapi...@gmail.com wrote:
>
>>
>> For example, I don't want the time consumed for "copy(data2, data)" being
>> counted.
>> How to achieve this?
>>
>> func BenchmarkFilter3(b *testing.B) {
>> data := buildOrginalData()
>> data2 := make([]int, len(data))
>> b.ResetTimer()
>> for i := 0; i < b.N; i++ {
>> copy(data2, data)
>> _ = Filter3(data2)
>> }
>> }
>>
>
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/c1f156d5-e582-4932-88f8-a2df6c987d85n%40googlegroups.com.
