I ran the exact same benchmarks as are in the README, and @Batch no longer holds up to @threads.
I got the following results
serial: 3.031 μs (0 allocations: 0 bytes)
batch: 10.828 ms (273779 allocations: 5.25 MiB)
threads: 5.916 μs (31 allocations: 3.19 KiB)
I am on an M1 MacBook Pro with 6 threads, Julia 1.10.10, Polyester v0.7.18
I have also tested some code on a Linux machine, Julia 1.10.4 with 128 threads, and seen poor performance.
I ran the exact same benchmarks as are in the README, and @Batch no longer holds up to @threads.
I got the following results
serial: 3.031 μs (0 allocations: 0 bytes)
batch: 10.828 ms (273779 allocations: 5.25 MiB)
threads: 5.916 μs (31 allocations: 3.19 KiB)
I am on an M1 MacBook Pro with 6 threads, Julia 1.10.10, Polyester v0.7.18
I have also tested some code on a Linux machine, Julia 1.10.4 with 128 threads, and seen poor performance.