[BENCHMARK] Lmbench 2.5.54-mm2 (impressive improvements)

From: Aniruddha M Marathe (aniruddha.marathe@wipro.com)
Date: Fri Jan 03 2003 - 03:59:04 EST


Here is a comparison of results of 2.5.54 with mm2 and 2.5.54.
There are many improvements with mm2. Please see the short summary below.
The figures in the table below indicate median of 5 repetitions of tests.
This result doesn't have many
Differences than the previous one.

                                                2.5.54mm2 2.5.54
==============================================================================
Processor, Processes - times in microseconds - smaller is better

1. sig handle 3.59 5.38
2. proc fs 1212 1534
3. sh proc 6606 7872
------------------------------------------------------------------------------
Context switching - times in microseconds - smaller is better

1. 2p/0K ctxsw 1.370 1.560
------------------------------------------------------------------------------
*Local* Communication latencies in microseconds - smaller is better
1.AF UNIX 13 19
2.UDP 24 30
3.TCP 58 75
------------------------------------------------------------------------------
File & VM system latencies in microseconds - smaller is better
1. 0K create 90 118
2. 0k delete 28 55
3. 10K create 313 375
4. 10 K delete 79 126
------------------------------------------------------------------------------
*Local* Communication bandwidths in MB/s - bigger is better
1. AF UNIX 277 109
2. TCP 51 25
==============================================================================

*****************************************************************************
                                Lmbench result
                                kernel 2.5.54-mm2
****************************************************************************
L M B E N C H 2 . 0 S U M M A R Y
                 ------------------------------------
                 (Alpha software, do not distribute)

Basic system parameters
----------------------------------------------------
Host OS Description Mhz
                                                    
--------- ------------- ----------------------- ----
benchtest Linux 2.5.54 i686-pc-linux-gnu 790
benchtest Linux 2.5.54 i686-pc-linux-gnu 790
benchtest Linux 2.5.54 i686-pc-linux-gnu 790
benchtest Linux 2.5.54 i686-pc-linux-gnu 790
benchtest Linux 2.5.54 i686-pc-linux-gnu 790

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 TCP inst hndl proc proc proc
--------- ------------- ---- ---- ---- ---- ---- ----- ---- ---- ---- ---- ----
benchtest Linux 2.5.54 790 0.44 0.81 6.28 7.58 1.27 3.63 282 1212 6618
benchtest Linux 2.5.54 790 0.46 0.83 6.42 7.71 33 1.27 3.59 304 1217 6569
benchtest Linux 2.5.54 790 0.46 0.82 6.39 7.68 30 1.24 3.59 337 1211 6609
benchtest Linux 2.5.54 790 0.46 0.82 6.40 7.61 30 1.27 3.63 318 1212 6588
benchtest Linux 2.5.54 790 0.46 0.80 6.41 7.72 32 1.24 3.59 274 1232 6606

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
--------- ------------- ----- ------ ------ ------ ------ ------- -------
benchtest Linux 2.5.54 1.350 4.8700 71 6.8600 176 41 180
benchtest Linux 2.5.54 1.410 4.8200 18 8.3300 180 41 180
benchtest Linux 2.5.54 1.390 4.7400 17 7.8600 179 43 180
benchtest Linux 2.5.54 1.370 4.8200 18 8.0500 178 39 180
benchtest Linux 2.5.54 1.370 4.7900 73 10 178 39 181

*Local* Communication latencies in microseconds - smaller is better
-------------------------------------------------------------------
Host OS 2p/0K Pipe AF UDP RPC/ TCP RPC/ TCP
                        ctxsw UNIX UDP TCP conn
--------- ------------- ----- ----- ---- ----- ----- ----- ----- ----
benchtest Linux 2.5.54 1.350 8.273 13 23 45 32 58 104
benchtest Linux 2.5.54 1.410 8.174 13 24 45 28 58 104
benchtest Linux 2.5.54 1.390 8.221 13 24 45 32 58 104
benchtest Linux 2.5.54 1.370 8.279 13 24 45 32 58 104
benchtest Linux 2.5.54 1.370 8.143 13 21 45 32 58 104

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
--------- ------------- ------ ------ ------ ------ ------- ----- -----
benchtest Linux 2.5.54 90 28 311 72 640 0.962 4.00000
benchtest Linux 2.5.54 91 29 316 79 637 0.963 5.00000
benchtest Linux 2.5.54 90 28 313 79 635 0.971 4.00000
benchtest Linux 2.5.54 90 28 315 77 643 0.969 4.00000
benchtest Linux 2.5.54 90 28 312 79 636 0.965 4.00000

*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
--------- ------------- ---- ---- ---- ------ ------ ------ ------ ---- -----
benchtest Linux 2.5.54 581 230 51 299 355 125 114 354 170
benchtest Linux 2.5.54 595 454 52 294 352 123 112 352 169
benchtest Linux 2.5.54 582 277 51 294 352 123 112 351 168
benchtest Linux 2.5.54 589 238 53 281 352 123 112 351 168
benchtest Linux 2.5.54 566 397 51 294 351 123 112 351 168

Memory latencies in nanoseconds - smaller is better
    (WARNING - may not be correct, check graphs)
---------------------------------------------------
Host OS Mhz L1 $ L2 $ Main mem Guesses
--------- ------------- ---- ----- ------ -------- -------
benchtest Linux 2.5.54 790 3.799 8.8810 175
benchtest Linux 2.5.54 790 3.798 8.8810 176
benchtest Linux 2.5.54 790 3.797 8.8810 176
benchtest Linux 2.5.54 790 3.808 8.8800 176
benchtest Linux 2.5.54 790 3.798 8.8710 176

Aniruddha Marathe
WIPRO Technologies, India
aniruddha.marathe@wipro.com
+91-80-5502001 to 2008 extn 5092
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Tue Jan 07 2003 - 22:00:20 EST