Monday, July 16, 2012

CPU performace comparison

For comparison purposes we tested some modern CPU (previous post) .
Main aim is to evaluate one core.

So, we wrote two bash scripts :


 time for ((i=0;i<10000000;i++)); do :; done

((i=0)); time while (( i<10000000 )); do ((i+=1)); done


Because both test give similar results I show only one column (timings in seconds):

---------------- -----------------------------------
                 for((i=0;i<10000000;i++));do :;done
---------------- -----------------------------------
SuperMicro                         52.38-53.85     
Intel Xeon X5675 3.07GHz          sys 3.17-3.30s   
                                                   
ODA                                58.22-60.07     
Intel Xeon X5675 3.07GHz          sys 3.32-3.51s   
                                                   
Exadata                            57.80-59.20     
Intel Xeon X5675 3.07GHz          sys 3.32-3.85s   
                                                   
T4 max-ipc                          138-149.11     
2824MHz                           sys 0.005-0.012s 
                                                   
T4 max-throughput                   144.78-145.30  
2824MHz                           sys 0.004s       
                                                   
Sun SPARC Enterprise M4000          161.78         
System clock frequency: 1012 MHz  sys 0.008s       
Run MHz 2150                                       
                                                   
SunFire 240                         290.25         
SUNW,UltraSPARC-IIIi              sys 0.70s        
1503 MHz                                           
                                                   
IBM JS22 POWER6 4 ГГц               227.30         
IBM,                              sys 6.672s       
frequency 4005000000 Processor Speed               
-----------------------------------------------------

Full table here.

                                                   
Conclusions:
- we cannot see the difference between T4 max-ipc and max-throughput modes.
- T4 slowly than Intel Xeon 5675 approximately 2.5-3 times


No comments:

Post a Comment

Note: Only a member of this blog may post a comment.

Does DEALLOCATE UNUSED or SHRINK SPACE will free space occupied by LOB segment?

Lets check how it works. My env is DB 19.20@Linux-x64 1) I created the table with 4 LOB columns of 4 different LOB types: BASICFILE BLOB, BA...