Benchmarking an E6500 against a dual Xeon 3GHz Dell 2650

Lamar Owen <[email protected]> Sun, 4 Sep 2005 23:26:35 -0400
Newsgroups gmane.linux.aurora.devel
Organization Pisgah Astronomical Research Institute
Message-ID <[email protected]>
Doesn't seem fair to do this comparison, but, here it is.

Spacely's results on a Unixbench 4.1 run:  Spacely is, again, 14x400/8M-ecache 
w/ 16GB RAM and an SBUS UW SCSI controller (Fast-20 maybe tops).
==============================================================

  BYTE UNIX Benchmarks (Version 4.1.0)
  System -- Linux spacely.pari.edu 2.6.8-1.571sp1smp #1 SMP Tue Sep 14 
17:55:47 EDT 2004 sparc64 sparc64 sparc64 GNU/Linux
  Start Benchmark Run: Mon Sep  5 02:11:25 EDT 2005
   1 interactive users.
   02:11:25 up 23:13,  1 user,  load average: 0.17, 0.05, 0.01
  lrwxrwxrwx  1 root root 4 Sep  4 00:13 /bin/sh -> bash
  /bin/sh: symbolic link to `bash'
  /dev/sda4             15493524   5456484   9249996  38% /
Dhrystone 2 using register variables     897508.1 lps   (10.0 secs, 10 
samples)
Double-Precision Whetstone                  184.3 MWIPS (10.0 secs, 10 
samples)
System Call Overhead                     182225.1 lps   (10.0 secs, 10 
samples)
Pipe Throughput                           82382.5 lps   (10.0 secs, 10 
samples)
Pipe-based Context Switching              15649.2 lps   (10.0 secs, 10 
samples)
Process Creation                           1236.5 lps   (30.0 secs, 3 samples)
Execl Throughput                            418.4 lps   (29.8 secs, 3 samples)
File Read 1024 bufsize 2000 maxblocks    101291.0 KBps  (30.0 secs, 3 samples)
File Write 1024 bufsize 2000 maxblocks    41396.0 KBps  (30.0 secs, 3 samples)
File Copy 1024 bufsize 2000 maxblocks     28531.0 KBps  (30.0 secs, 3 samples)
File Read 256 bufsize 500 maxblocks       41734.0 KBps  (30.0 secs, 3 samples)
File Write 256 bufsize 500 maxblocks      13349.0 KBps  (30.0 secs, 3 samples)
File Copy 256 bufsize 500 maxblocks        9371.0 KBps  (30.0 secs, 3 samples)
File Read 4096 bufsize 8000 maxblocks    195588.0 KBps  (30.0 secs, 3 samples)
File Write 4096 bufsize 8000 maxblocks   110226.0 KBps  (30.0 secs, 3 samples)
File Copy 4096 bufsize 8000 maxblocks     72245.0 KBps  (30.0 secs, 3 samples)
Shell Scripts (1 concurrent)                710.7 lpm   (60.0 secs, 3 samples)
Shell Scripts (8 concurrent)                377.3 lpm   (60.0 secs, 3 samples)
Shell Scripts (16 concurrent)               244.7 lpm   (60.0 secs, 3 samples)
Arithmetic Test (type = short)            48010.2 lps   (10.0 secs, 3 samples)
Arithmetic Test (type = int)              48967.4 lps   (10.0 secs, 3 samples)
Arithmetic Test (type = long)             48718.9 lps   (10.0 secs, 3 samples)
Arithmetic Test (type = float)           112811.3 lps   (10.0 secs, 3 samples)
Arithmetic Test (type = double)           89922.3 lps   (10.0 secs, 3 samples)
Arithoh                                  1833809.3 lps   (10.0 secs, 3 
samples)
C Compiler Throughput                       189.7 lpm   (60.0 secs, 3 samples)
Dc: sqrt(2) to 99 decimal places          24353.3 lpm   (30.0 secs, 3 samples)
Recursion Test--Tower of Hanoi            15394.4 lps   (20.0 secs, 3 samples)


                     INDEX VALUES
TEST                                        BASELINE     RESULT      INDEX

Dhrystone 2 using register variables        116700.0   897508.1       76.9
Double-Precision Whetstone                      55.0      184.3       33.5
Execl Throughput                                43.0      418.4       97.3
File Copy 1024 bufsize 2000 maxblocks         3960.0    28531.0       72.0
File Copy 256 bufsize 500 maxblocks           1655.0     9371.0       56.6
File Copy 4096 bufsize 8000 maxblocks         5800.0    72245.0      124.6
Pipe-based Context Switching                  4000.0    15649.2       39.1
Process Creation                               126.0     1236.5       98.1
Shell Scripts (8 concurrent)                     6.0      377.3      628.8
System Call Overhead                         15000.0   182225.1      121.5
                                                                 =========
     FINAL SCORE                                                      90.6
[root@spacely unixbench-4.1.0]#

Pulsar: (Dell 2650 dual Xeon 3.0, 1MB L2, 4GB RAM, Dell PERC4e/Di U320 
hardware RAID with RAID1 73GB 10K RPM drives)
==============================================================

  BYTE UNIX Benchmarks (Version 4.1.0)
  System -- Linux pulsar.pari.edu 2.6.9-11.106.unsupportedsmp #1 SMP Wed Jun 8 
22:05:04 CDT 2005 i686 i686 i386 GNU/Linux
  Start Benchmark Run: Sun Sep  4 22:15:09 EDT 2005
   4 interactive users.
   22:15:09 up 3 days,  4:18,  4 users,  load average: 0.52, 0.45, 0.27
  lrwxrwxrwx  1 root root 4 Aug 24 14:40 /bin/sh -> bash
  /bin/sh: symbolic link to `bash'
                        66086592  51933204  10796308  83% /
Dhrystone 2 using register variables     4531227.3 lps   (10.0 secs, 10 
samples)
Double-Precision Whetstone                  960.6 MWIPS (10.6 secs, 10 
samples)
System Call Overhead                     576356.0 lps   (10.0 secs, 10 
samples)
Pipe Throughput                          509622.1 lps   (10.0 secs, 10 
samples)
Pipe-based Context Switching              37867.6 lps   (10.0 secs, 10 
samples)
Process Creation                           6755.8 lps   (30.0 secs, 3 samples)
Execl Throughput                           2243.9 lps   (29.8 secs, 3 samples)
File Read 1024 bufsize 2000 maxblocks    573427.0 KBps  (30.0 secs, 3 samples)
File Write 1024 bufsize 2000 maxblocks   322954.0 KBps  (30.0 secs, 3 samples)
File Copy 1024 bufsize 2000 maxblocks    193524.0 KBps  (30.0 secs, 3 samples)
File Read 256 bufsize 500 maxblocks      163903.0 KBps  (30.0 secs, 3 samples)
File Write 256 bufsize 500 maxblocks      88538.0 KBps  (30.0 secs, 3 samples)
File Copy 256 bufsize 500 maxblocks       57000.0 KBps  (30.0 secs, 3 samples)
File Read 4096 bufsize 8000 maxblocks    1497407.0 KBps  (30.0 secs, 3 
samples)
File Write 4096 bufsize 8000 maxblocks   750400.0 KBps  (30.0 secs, 3 samples)
File Copy 4096 bufsize 8000 maxblocks    490659.0 KBps  (30.0 secs, 3 samples)
Shell Scripts (1 concurrent)               3298.0 lpm   (60.0 secs, 3 samples)
Shell Scripts (8 concurrent)                712.5 lpm   (60.0 secs, 3 samples)
Shell Scripts (16 concurrent)               366.0 lpm   (60.0 secs, 3 samples)
Arithmetic Test (type = short)           727961.2 lps   (10.0 secs, 3 samples)
Arithmetic Test (type = int)             795201.4 lps   (10.0 secs, 3 samples)
Arithmetic Test (type = long)            795517.8 lps   (10.0 secs, 3 samples)
Arithmetic Test (type = float)           630136.3 lps   (10.0 secs, 3 samples)
Arithmetic Test (type = double)          629729.0 lps   (10.0 secs, 3 samples)
Arithoh                                  15237826.2 lps   (10.0 secs, 3 
samples)
C Compiler Throughput                      1008.9 lpm   (60.0 secs, 3 samples)
Dc: sqrt(2) to 99 decimal places          85192.0 lpm   (30.0 secs, 3 samples)
Recursion Test--Tower of Hanoi           111295.1 lps   (20.0 secs, 3 samples)


                     INDEX VALUES
TEST                                        BASELINE     RESULT      INDEX

Dhrystone 2 using register variables        116700.0  4531227.3      388.3
Double-Precision Whetstone                      55.0      960.6      174.7
Execl Throughput                                43.0     2243.9      521.8
File Copy 1024 bufsize 2000 maxblocks         3960.0   193524.0      488.7
File Copy 256 bufsize 500 maxblocks           1655.0    57000.0      344.4
File Copy 4096 bufsize 8000 maxblocks         5800.0   490659.0      846.0
Pipe Throughput                              12440.0   509622.1      409.7
Process Creation                               126.0     6755.8      536.2
Shell Scripts (8 concurrent)                     6.0      712.5     1187.5
System Call Overhead                         15000.0   576356.0      384.2
                                                                 =========
     FINAL SCORE                                                     468.1
[root@pulsar unixbench-4.1.0]# 

Yes, the E6500 takes a pretty good beating, against the state of the art 
midrange server (not a bad showing for a 5 year old midrange server that only 
takes three times as much power (the 2650 sucks a lot of power!)).  But 
Unixbench isn't the best multiprocessor benchmark out there.  The 16 
concurrent shell test tells an interesting story, with the E6500 at 244lpm 
and the 2650 at 366.  The E6500 holds its own in the extreme concurrency 
point.

On disk read throughput:
E6500:
[root@spacely unixbench-4.1.0]# hdparm -t /dev/sda

/dev/sda:
 Timing buffered disk reads:   78 MB in  3.03 seconds =  25.75 MB/sec
[root@spacely unixbench-4.1.0]#  

Dell 2650:
[root@pulsar unixbench-4.1.0]# hdparm -t /dev/sda

/dev/sda:
 Timing buffered disk reads:  154 MB in  3.01 seconds =  51.14 MB/sec
[root@pulsar unixbench-4.1.0]# 

Spacely's not shabby on disk, considering:
[root@localhost ~]# hdparm -t /dev/hda

/dev/hda:
 Timing buffered disk reads:   70 MB in  3.02 seconds =  23.16 MB/sec
[root@localhost ~]#

What is localhost?  My Dell Inspiron 600m laptop with an ATA100 Toshiba 60GB 
drive, 4200 RPM.

-- 
Lamar Owen
Director of Information Technology
Pisgah Astronomical Research Institute
1 PARI Drive
Rosman, NC  28772
(828)862-5554
www.pari.edu
_______________________________________________
Aurora-sparc-devel mailing list
[email protected]
http://lists.auroralinux.org/mailman/listinfo/aurora-sparc-devel
Aurora FAQ: http://www.ecs.soton.ac.uk/~mas01r/aurorafaq.html