[Benchmarks] Cleanup fp_reduction_latency benchmarks

[mratsim]

The fp_reduction_latency benchmarks were the very first benchmark, optimization and primitive code tested in Laser.

Unfortunately it is currently very confusing.

It should be reorganized:

• 1. Multiple accumulators: https://github.com/numforge/laser/blob/af191c086b4a98c49049ecf18f5519dc6856cc77/benchmarks/fp_reduction_latency/reduction_bench.nim and https://github.com/numforge/laser/blob/af191c086b4a98c49049ecf18f5519dc6856cc77/benchmarks/fp_reduction_latency/reduction_packed_accum.nim
• 2. raw vector intrinsics measurements: https://github.com/numforge/laser/blob/af191c086b4a98c49049ecf18f5519dc6856cc77/benchmarks/fp_reduction_latency/reduction_packed_sse.nim and https://github.com/numforge/laser/blob/af191c086b4a98c49049ecf18f5519dc6856cc77/benchmarks/fp_reduction_latency/reduction_sse_bench.nim
• 3. Measuring max/min implementation: https://github.com/numforge/laser/blob/af191c086b4a98c49049ecf18f5519dc6856cc77/benchmarks/fp_reduction_latency/reduction_max_bench.nim

This reorg should take into account nim-lang/Nim#9514