ARROW-17281: [C++] Fix cache size reporting on Windows

[cyb70289]

No description provided.

[github-actions]

https://issues.apache.org/jira/browse/ARROW-17281

[pitrou]

LGTM, did you get a chance to try on an actual machine?

[cyb70289]

I tested on an i7-1165 laptop with 4 cores x 2 hyperthreads.
It reports four L1 cache with 48K in size, and another four L1 cache with 32K in size.
Original code reports totally 320K, the PR reports 48K.

[kou]

+1

[ursabot]

Benchmark runs are scheduled for baseline = 71af8d8 and contender = b83da3f. b83da3f is a master commit associated with this PR. Results will be available as each benchmark for each run completes.
Conbench compare runs links:
[Finished ⬇️0.0% ⬆️0.0%] ec2-t3-xlarge-us-east-2
[Finished ⬇️0.27% ⬆️0.07%] test-mac-arm
[Finished ⬇️1.09% ⬆️0.0%] ursa-i9-9960x
[Finished ⬇️0.71% ⬆️0.04%] ursa-thinkcentre-m75q
Buildkite builds:
[Finished] b83da3fc ec2-t3-xlarge-us-east-2
[Finished] b83da3fc test-mac-arm
[Finished] b83da3fc ursa-i9-9960x
[Finished] b83da3fc ursa-thinkcentre-m75q
[Finished] 71af8d8f ec2-t3-xlarge-us-east-2
[Finished] 71af8d8f test-mac-arm
[Finished] 71af8d8f ursa-i9-9960x
[Finished] 71af8d8f ursa-thinkcentre-m75q
Supported benchmarks:
ec2-t3-xlarge-us-east-2: Supported benchmark langs: Python, R. Runs only benchmarks with cloud = True
test-mac-arm: Supported benchmark langs: C++, Python, R
ursa-i9-9960x: Supported benchmark langs: Python, R, JavaScript
ursa-thinkcentre-m75q: Supported benchmark langs: C++, Java