Read KIT data in cache-sized blocks

The reader asked for up to 100 MB per block, then cast the whole block to
float64 and scaled it in place, so the working set was several hundred MB and
none of it stayed in cache.

  raw.get_data(), interleaved cross-process medians:

    136 MB file (161ch x 425000)   258.5 -> 145.6 ms   1.78x

A sweep over 32/16/8/4/2/1 MiB puts the optimum at 2 MiB (32 MiB is only 1.14x,
8 MiB 1.38x, 4 MiB 1.66x, 1 MiB 1.60x).

No shipped fixture regresses; two of them improve because they were already
being split differently:

    ArtificalSignalData_Yokogawa_1khz.con   2.87 -> 2.26 ms   1.27x
    ArtificalSignalData_RICOH_1khz.con      3.06 -> 2.77 ms   1.10x
    Example_PQA160C_1001-export.con                           0.99x
    010409_Motor_task_coregist-export.con                     0.98x

Output is bit-identical on every readable KIT fixture. The large file was
synthesised by tiling the raw-data section of data_berlin.con -- it is the last
section in the file -- and patching n_samples in the acquisition header.

Same treatment as #14241 and #14246; the constant differs because the optimum
tracks time points per block, so it moves with the channel count.

Author: bruAristimunha

mne/io/kit/kit.py

@@ -82,6 +82,11 @@ def __init__(self, sqd_version, *args, **kwargs):
         ValueError.__init__(self, *args, **kwargs)
 
 
+# read in cache-sized blocks rather than one huge one (1.8x on a 136 MB file);
+# each block is cast to float64 and scaled, so a smaller one stays in cache
+_BLOCK_BYTES = 2 * 1024**2
+
+
 @fill_doc
 class RawKIT(BaseRaw):
     r"""Raw object from KIT SQD file.
@@ -209,8 +214,7 @@ def _read_segment_file(self, data, idx, fi, start, stop, cals, mult):
 
         n_bytes = sqd["dtype"].itemsize
         assert n_bytes in (2, 4)
-        # Read up to 100 MB of data at a time.
-        blk_size = min(data_left, (100000000 // n_bytes // nchan) * nchan)
+        blk_size = min(data_left, (_BLOCK_BYTES // n_bytes // nchan) * nchan)
         with open(self.filenames[fi], "rb", buffering=0) as fid:
             # extract data
             pointer = start * nchan * n_bytes