Benchmark : Speedup

General CULA Dense (LAPACK & BLAS) support and troubleshooting. Use this forum if you are having a general problem or have encountered a bug.

Benchmark : Speedup

Postby stage » Fri May 21, 2010 7:10 am

I run the benchmark (benchmark_.exe) and I obtain that :

Code: Select all
Initializing CULA...
Initializing MKL...

Benchmarking the following functions:
-------------------------------------
             SGEQRF
             SGETRF
             SGELS
             SGGLSE
             SGESV
             SGESVD
-------------------------------------


     -- SGEQRF Benchmark  --

Size   CULA (s)    MKL (s)   Speedup
------ ---------- ---------- ---------
4096       4.47       1.43    0.3209
5120       7.75       2.46    0.3176
6144      13.29       4.21    0.3166
7168      20.97       6.50    0.3101
8192      19.97      10.40    0.5206

     -- SGETRF Benchmark  --

Size   CULA (s)    MKL (s)   Speedup
------ ---------- ---------- ---------
4096       1.78       0.79    0.4462
5120       3.17       1.37    0.4317
6144       5.29       2.42    0.4582
7168       8.17       3.45    0.4226
8192      12.00       5.47    0.4557

     -- SGELS Benchmark  --

Size   CULA (s)    MKL (s)   Speedup
------ ---------- ---------- ---------
4096       4.51       1.98    0.4387
5120       8.49       2.72    0.3198
6144      14.42       4.75    0.3297
7168      22.43       7.59    0.3383
8192      21.74      10.83    0.4981

     -- SGGLSE Benchmark  --

Size   CULA (s)    MKL (s)   Speedup
------ ---------- ---------- ---------
4096       4.87       3.77    0.7734
5120       9.05       6.60    0.7296
6144      14.77       9.97    0.6753
7168      23.22      14.37    0.6187
8192      22.66      20.48    0.9039

     -- SGESV Benchmark  --

Size   CULA (s)    MKL (s)   Speedup
------ ---------- ---------- ---------
4096       1.75       0.74    0.4227
5120       3.22       1.33    0.4128
6144       5.35       2.31    0.4322
7168       8.27       3.44    0.4162
8192      12.14       5.61    0.4616

     -- SGESVD Benchmark  --

Size   CULA (s)    MKL (s)   Speedup
------ ---------- ---------- ---------
4096     211.94     114.19    0.5388
5120     407.30     202.67    0.4976
6144
CULA Error: Insufficient memory to complete this operation


Is it normal ?

I use CULAtools 1.3a with Nvidia Quadro FX 580 and Intel Core i7 860 on Windows 7 Professionnal (64 bit).

Thank you ;)

(Sorry for my english, i'm french)
stage
 
Posts: 2
Joined: Thu Apr 22, 2010 1:10 am

Re:Benchmark : Speedup

Postby john » Fri May 21, 2010 7:20 am

This is to be expected given your hardware setup. The Quadro 580 part isn't meant for CUDA performance - it has 32 cores vs the 240 for the Quadro 5800. And now the Fermi parts have 448. I'm afraid you need a bigger GPU.
john
Administrator
 
Posts: 587
Joined: Thu Jul 23, 2009 2:31 pm


Return to CULA Dense Support

Who is online

Users browsing this forum: No registered users and 1 guest

cron