Lots of allocation when parsing float numbers

[BambOoxX]

Hello, I have more a question than a real issue.

When parsing strings composed of multiple float numbers, I get a lot of allocations I do not understand.
The following example

using Scanf, BenchmarkTools
s = "1.39086e+00  1.39391e+00  1.39544e+00  1.39239e+00  1.41680e+00  1.41375e+00"
@btime @scanf($s, "%13e%13e%13e%13e%13e%13e", Float32, Float32, Float32, Float32, Float32, Float32)

produces

6.560 μs (35 allocations: 1.88 KiB)

I can't really make sense of all these allocations. Could you please help me understand where they come from ?
For large datasets of the same kind, scanf becomes the bottleneck due to the number of allocations performed there.

Best regards.