Steve
December 12th 03, 05:43 PM
Is there any way to improve the performance of windows
for sequential access to large files? When copying a
large file, for example, my measurements suggest that
windows only reads 16-64KB from the source file before
doing a seek and writing the destination file. As a
result the throughput is only a couple megabytes per
second which is a small fraction of the hardware
capabilities. I would expect windows to fetch numerous
clusters from the source file -- several megabytes --
buffer them in its cache and only occasionally do seeks
between the source and destination files.
for sequential access to large files? When copying a
large file, for example, my measurements suggest that
windows only reads 16-64KB from the source file before
doing a seek and writing the destination file. As a
result the throughput is only a couple megabytes per
second which is a small fraction of the hardware
capabilities. I would expect windows to fetch numerous
clusters from the source file -- several megabytes --
buffer them in its cache and only occasionally do seeks
between the source and destination files.