Lines Matching full:our

52 small size and our low sampling rate. This means it’s more efficient to use the
60 our probability of sampling an allocation of any size is, as well as our
68 sample all bytes within the allocation if we sample bytes at our sampling rate.
76 We can see from the chart below that if we 16X our sampling rate from 32KiB to
95 about and it’s useful as a gauge of how wrong on average we might be with our
111 can expect for the things that end up in our heap profile. It’s important to
121 Benchmarking of the STL distributions on a Pixel 4 reinforces our approach of
144 and immediately if the allocation size was greater than our sampling rate. This
149 an allocation equal in size to our sampling rate ~63% of the time, rather than
151 byte smaller than our sampling rate, and one a byte larger. This is still
156 the memory being used, but also how it’s allocated. If our sampling rate were
159 probability 63%. Again this doesn’t bias our results, but it means that our
164 mean our method is insensitive to how the memory is split amongst allocations,