Messages in this thread | | | Date | Sat, 8 May 1999 12:22:46 +0200 (CEST) | From | Mikulas Patocka <> | Subject | RE: Mindcraft III |
| |
> One novice question. How are "Well" configured NT's in comparison with > "Well" tuned Linux boxes ?In the best possible case, we should beat NT like > a drum( as tested by zdnet.com ). The question is whether this can be > replicated in the Mindcraft.
There's one more bad thing in Mindcraft tests. They're testing Linux and NT server with 95 clients and SMB protocol. SMB is the native protocol of NT, I don't wonder that it's faster (NT has SMB wired in kernel; in Linux it's just userspace program). The same way NFS is surely faster under Linux (knfsd) than NT. Somebody setting up linux network certainly won't run samba on linux server and smb client on linux workstations. Mindcraft is doing like SMB is the only file protocol and 95 is the only client! Grrr... They are still under control of Micro$hit.
They should do tests with all 4 combinations of clients and servers (nt - 95, nt - linux, linux - 95, linux - linux) and protocols (smb, ncp, nfs, coda). Such results would be much more helpful, that just saying "nt is 3 times faster than linux".
If you want some results, it's easy to pick up one of these 16 configurations, do tests and get results you wanted. If you test NT versus Linux server with Linux clients over NFS or CODA, you can surely prove, that linux is 10 times faster... Or - if you test it on a server with 16M ram, you get that Linux is 100000 times faster :-)
The fact that Mindcraft ignores all other clients and protocols should have been criticized more than bad tuned Linux.
Mikulas Patocka
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |