Thread: [User-mode-linux-user] lmbench numbers?

SourceForge Headquarters 1320 Columbia Street Suite 310 San Diego, CA 92101 +1 (858) 422-6466

Has anyone tried lmbench in uml?

Here are some results.  The benchmark crashed the kernel before finishing
the context switch and memory latency measurements.

"i686-linu" is Linux 2.2.5 running on hardware, and "um-linuxo" is
user-mode Linux 2.3.51 on the same machine.  I have edited the results
slightly for clarity.

                 L M B E N C H  1 . 9   S U M M A R Y
                 ------------------------------------
                 (Alpha software, do not distribute)

Processor, Processes - times in microseconds - smaller is better
----------------------------------------------------------------
Host                 OS  Mhz null null      open selct sig  sig  fork exec sh  
                             call  I/O stat clos       inst hndl proc proc proc
--------- ------------- ---- ---- ---- ---- ---- ----- ---- ---- ---- ---- ----
i686-linu Linux 2.2.5-2  499  0.6  0.9    4    5 0.04K  1.9    3 0.3K   1K   6K
um-linuxo Linux 2.3.51-  499 95.1  159  126  230 0.14K 98.5   31 9.2K  19K  56K

Context switching - times in microseconds - smaller is better
-------------------------------------------------------------
Host                 OS 2p/0K 2p/16K 2p/64K 8p/16K 8p/64K 16p/16K 16p/64K
                        ctxsw  ctxsw  ctxsw ctxsw  ctxsw   ctxsw   ctxsw
--------- ------------- ----- ------ ------ ------ ------ ------- -------
i686-linu Linux 2.2.5-2    1     14     43    15    128      19     157
um-linuxo Linux 2.3.51-    ( no results due to crash )

*Local* Communication latencies in microseconds - smaller is better
-------------------------------------------------------------------
Host                 OS 2p/0K  Pipe AF     UDP  RPC/   TCP  RPC/ TCP
                        ctxsw       UNIX         UDP         TCP conn
--------- ------------- ----- ----- ---- ----- ----- ----- ----- ----
i686-linu Linux 2.2.5-2     1     6   13   ( no results because I
um-linuxo Linux 2.3.51-         988 1136     didn't measure networking )

File & VM system latencies in microseconds - smaller is better
--------------------------------------------------------------
Host                 OS   0K File      10K File      Mmap    Prot    Page
                        Create Delete Create Delete  Latency Fault   Fault
--------- ------------- ------ ------ ------ ------  ------- -----   ----- 
i686-linu Linux 2.2.5-2     14      1     26      2    22126     1    0.6K
um-linuxo Linux 2.3.51-     48     15     60     18     6635          0.2K

*Local* Communication bandwidths in MB/s - bigger is better
-----------------------------------------------------------
Host                OS  Pipe AF    TCP  File   Mmap  Bcopy  Bcopy  Mem   Mem
                             UNIX      reread reread (libc) (hand) read write
--------- ------------- ---- ---- ---- ------ ------ ------ ------ ---- -----
i686-linu Linux 2.2.5-2  356  121   -1    176    354    150    146  355   194
um-linuxo Linux 2.3.51-   -1        -1     -1     -1     -1     -1   -1    -1
                                        ( these are completely bogous )

Memory latencies in nanoseconds - smaller is better
    (WARNING - may not be correct, check graphs)
---------------------------------------------------
Host                 OS   Mhz  L1 $   L2 $    Main mem    Guesses
--------- -------------   ---  ----   ----    --------    -------
i686-linu Linux 2.2.5-2   499     6     46         151
um-linuxo Linux 2.3.51-   499    ( no results due to crash )

lars brinkhoff <la...@no...> writes:
> Processor, Processes - times in microseconds - smaller is better
> ----------------------------------------------------------------
> Host               OS  Mhz null null      open selct sig  sig  fork exec sh  
>                            call  I/O stat clos       inst hndl proc proc proc
> --------- ----------- ---- ---- ---- ---- ---- ----- ---- ---- ---- ---- ----
> i686-linu Linux 2.2.5  499  0.6  0.9    4    5 0.04K  1.9    3 0.3K   1K   6K
> um-linuxo Linux 2.3.51 499 95.1  159  126  230 0.14K 98.5   31 9.2K  19K  56K

User-mode Linux seems to be about 100 times slower than Linux on bare
hardware on fast operations.  Slower operations like fork() and exec()
does better.

> File & VM system latencies in microseconds - smaller is better
> --------------------------------------------------------------
> Host                 OS   0K File      10K File      Mmap    Prot    Page
>                         Create Delete Create Delete  Latency Fault   Fault
> --------- ------------- ------ ------ ------ ------  ------- -----   ----- 
> i686-linu Linux 2.2.5-2     14      1     26      2    22126     1    0.6K
> um-linuxo Linux 2.3.51-     48     15     60     18     6635          0.2K

Surprisingly, user-mode Linux seems to be do mmap() and page faults
faster than Linux on hardware.  I wonder why?

lars brinkhoff <la...@no...> writes:
> Here are some results.  The benchmark crashed the kernel before finishing
> the context switch and memory latency measurements.

When I try to run the context switch benchmark, uml forks off lots
of subprocesses and I get a lot of
        fork: Resource temporarily unavailable
messages from the hosting kernel.

Thread: [User-mode-linux-user] lmbench numbers?

user-mode-linux-user