Re: benchmark results of page granurality patch
In article <[🔎] Pine.LNX.4.21.0007140438120.4722-100000@spawn.hockeyfiend.com>
chris@debian.org writes:
>>
>> On Fri, 14 Jul 2000, Naohiko Shimizu wrote:
>>
>> > The kernel patch http://shimizu-lab.et.u-tokai.ac.jp/~nshimizu makes
>> > HPC applications faster. I ran some benchmark with pached kernel.
>>
>> The benchmarks are impressive! What were you using to benchmark it?
>>
The matrix transpose and multiply benchmarks are on the web, and
I used FFT from Ron Mayer's code and modified to repeat 10 times
between the dtime() functions. I am sure that many HPC applications
(but not written by FORTRAN) will be faster. I will look into the g77
code to support writable initial allocation. But it is difficult for
COMPAQ's Fortran, and I will look for some solution from kernel side.
>> > BTW, why 2.4.0-test? kernel crashes on the mount?
>> > Is there any solution on the mount problem?
>>
>> I'm looking into it. I think it's that potato has an older version of
....
>>
It works with normal HDD in spite of the warnings, but NFS root crashes.
I am not sure NFS root problem is related to the warnings. But it is
inconvenient to whom debuging a kernel (long fsck or corrupt files when
the underway kernel crashes).
Thank you,
Naohiko Shimizu
Dept. Communication Engr./Univ. TOKAI
1117 Kitakaname Hiratsuka 259-12 Japan
TEL.+81-463-58-1211(ext. 4084) FAX.+81-463-58-8320
Reply to: