On Sunday, 9 August 2015 at 00:50:16 UTC, Ali Çehreli wrote:
Ali
Now benchmarks write and read separately: https://github.com/nordlow/justd/blob/0d746b2c1800a82a61a6cb7edcabfd9664066b2c/tests/t_rawio.d Couldn't the chunk logic be deduced aswell? Something like: void rawWriteInAutoChunks(R)(File f, R r) {const count = preferred_disk_write_size / sizeof(ElementType!R);
return r.chunks(count).each!(a => f.rawWrite(a.array)); } What would a suitable value for `preferred_disk_write_size` be?