[BENCHMARK] 2.5.60 Lmbench performance

From: Aniruddha M Marathe (aniruddha.marathe@wipro.com)
Date: Thu Feb 13 2003 - 04:32:12 EST


Hi,
Here are results of Lmbench for 2.5.60. Also look at the summary comparison below.

                                                2.5.60 2.5.59
==============================================================================
Processor, Processes - times in microseconds - smaller is better

1. sh proc 6652 6770
------------------------------------------------------------------------------------------------------------------------------
Context switching - times in microseconds - smaller is better

1. 2p/0K ctxsw 1.430 1.660
------------------------------------------------------------------------------------------------------------------------------
*Local* Communication latencies in microseconds - smaller is better
1. TCP connection 101 110
------------------------------------------------------------------------------------------------------------------------------
File & VM system latencies in microseconds - smaller is better
1. Prot fault 0.562 1.035
-------------------------------------------------------------------------------------------------------------------------------
*Local* Communication bandwidth in MB/s - bigger is better
1. AF unix 411 559
==============================================================================

*********************************************************************************************
                LMBENCH for kernel 2.5.60
*********************************************************************************************

                 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 2.5.60 i686-pc-linux-gnu 790
benchtest Linux 2.5.60 i686-pc-linux-gnu 790
benchtest Linux 2.5.60 i686-pc-linux-gnu 790
benchtest Linux 2.5.60 i686-pc-linux-gnu 790
benchtest Linux 2.5.60 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 2.5.60 790 0.44 0.81 7.13 8.77 1.26 4.01 282 1281 6584
benchtest Linux 2.5.60 790 0.46 0.82 7.13 8.71 34 1.25 3.98 326 1287 6652
benchtest Linux 2.5.60 790 0.46 0.82 7.12 8.68 34 1.26 4.01 311 1303 6627
benchtest Linux 2.5.60 790 0.46 0.82 7.16 8.73 34 1.26 3.98 293 1302 6679
benchtest Linux 2.5.60 790 0.44 0.82 7.30 8.80 34 1.26 3.99 336 1299 6644

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 2.5.60 1.350 4.8400 32 9.8000 175 43 178
benchtest Linux 2.5.60 1.480 4.7900 14 10 178 41 179
benchtest Linux 2.5.60 1.430 4.7900 14 11 178 43 179
benchtest Linux 2.5.60 1.370 4.7400 16 9.2700 178 40 179
benchtest Linux 2.5.60 1.430 4.8300 14 8.6300 178 42 179

*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 2.5.60 1.350 8.616 13 25 47 30 57 101
benchtest Linux 2.5.60 1.480 8.370 13 25 46 30 57 102
benchtest Linux 2.5.60 1.430 8.384 13 24 46 31 57 102
benchtest Linux 2.5.60 1.370 8.498 14 24 46 30 57 103
benchtest Linux 2.5.60 1.430 8.562 14 25 46 30 58 102

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 2.5.60 92 31 362 85 641 0.562 4.00000
benchtest Linux 2.5.60 91 30 324 80 608 0.602 4.00000
benchtest Linux 2.5.60 91 31 325 84 618 0.574 4.00000
benchtest Linux 2.5.60 90 30 325 83 636 0.605 4.00000
benchtest Linux 2.5.60 90 30 325 83 624 0.605 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 2.5.60 344 309 51 295 355 124 113 355 170
benchtest Linux 2.5.60 544 253 53 298 353 123 112 354 169
benchtest Linux 2.5.60 511 411 52 296 353 123 112 354 169
benchtest Linux 2.5.60 575 552 52 272 353 123 112 354 169
benchtest Linux 2.5.60 577 477 53 295 353 123 113 354 170

Memory latencies in nanoseconds - smaller is better
    (WARNING - may not be correct, check graphs)
---------------------------------------------------
Host OS Mhz L1 $ L2 $ Main mem Guesses
--------- ------------- ---- ----- ------ -------- -------
benchtest Linux 2.5.60 790 3.800 56 174
benchtest Linux 2.5.60 790 3.804 8.8710 175
benchtest Linux 2.5.60 790 3.798 8.8810 175
benchtest Linux 2.5.60 790 3.801 8.8810 175
benchtest Linux 2.5.60 790 3.798 10 175

Aniruddha Marathe
WIPRO technologies, India
-
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 : Sat Feb 15 2003 - 22:00:46 EST