Dylan Tuttle

@hzongaro Done! Here they are: [log_countPositives_exactIndex_1.txt](https://github.com/user-attachments/files/19291339/log_countPositives_exactIndex_1.txt) [log_countPositives_exactIndex_2.txt](https://github.com/user-attachments/files/19291340/log_countPositives_exactIndex_2.txt) [log_hasNegatives_exactIndex.txt](https://github.com/user-attachments/files/19291341/log_hasNegatives_exactIndex.txt)

Updated logs for the shift-based residual processing algorithm: [log_countPositives_shift.txt](https://github.com/user-attachments/files/19455255/log_countPositives_shift.txt) [log_hasNegatives_shift.txt](https://github.com/user-attachments/files/19455256/log_hasNegatives_shift.txt) (logs for last week's longer but faster algorithm discussed in scrum this morning can be found in [this comment](https://github.com/eclipse-openj9/openj9/pull/21121#issuecomment-2730120363))

Reverted to non-exact residual handling. Logs: [log_countPositives_nonExact.txt](https://github.com/user-attachments/files/19592520/log_countPositives_nonExact.txt) [log_hasNegatives_nonExact.txt](https://github.com/user-attachments/files/19592521/log_hasNegatives_nonExact.txt)

I seem to have a couple of problems at the moment. For some reason, in cases with negative bytes, my implementation seems to be unable to return an index larger...

I fixed the JCL crash I mentioned in my [previous comment](https://github.com/eclipse-openj9/openj9/pull/21121#issuecomment-2803077381), turns out I neglected to add a jump to the end label in the no negatives case which was...

Thanks to some help from @hzongaro, we figured out that the max index of 128 was also due to a bug in the unit test, which means we are all...

Logs for the newest version: [log_countPositives_20250421.txt](https://github.com/user-attachments/files/19835079/log_countPositives_20250421.txt) [log_hasNegatives_20250421.txt](https://github.com/user-attachments/files/19835077/log_hasNegatives_20250421.txt)

At @BradleyWood's suggestion, I've added SIMD encoding for `PMOVMSKB` to specify that we want the 16 byte version. If that encoding isn't supported on the architecture, I've reverted to a...

The following code would fix the crash and also ensure both lines are properly aligned no matter the length of the benchmark name. It can probably be cleaned up further,...