ATTENTION: You are viewing a page formatted for mobile devices; to view the full web page, click HERE.

Main Area and Open Discussion > General Software Discussion

"Do copy acceleration utilities actually lower file transfer speeds?"

<< < (4/4)

f0dder:
For example if the operation only ran for same named files under 100 Megabytes in size then would there really be any problematic slowdown on computer with a newish CPU?-Nod5 (August 13, 2012, 03:50 PM)
--- End quote ---
CPU speed isn't really your main concern - something like Adler32 is pretty damn fast and "probably good enough". MD5 is also pretty fast, and since you're just comparing local files and not trying to be cryptographically safe, it will probably be sufficient.

The problem is that you're doing disk I/O. Instead of "1 read, 1 write", you'll be doing "2xRead + CPU, THEN perhaps 1 read, 1 write". OK, since you propose to use hashing, at least the 2xRead will run at full disk speed, whereas compare-the-bytes would be slower (seeking back and forth on a mechanical drive kills performance). But there's still a lot of overhead in this!

Nod5:
Ok, but doing a SHA1 check on a 150 MB file takes about 0,3 seconds on my system. So a bit over 0,6 seconds for a hash comparison. That is not so bad, especially compared to how Explorer works now, requestion user attention and delaying the file transfer with a popup window with a lot of text.

def:
http://www.freewaregenius.com/do-copy-acceleration-utilities-actually-lower-file-transfer-speeds-our-tests-say-yes/
-paulobrabo (August 10, 2012, 04:12 AM)
--- End quote ---
How can one write an article about copy accelaration utilities and then forget to mention the exact Windows version used?

"Windows version: Windows Home 64 bit."

Very meaningful indeed...

Navigation

[0] Message Index

[*] Previous page

Go to full version