On 08/27/2010 03:41 PM, Fabio Marconi wrote:
> Hello
> Is this problem present with the latest updated Karmic's or Lucid's packages?
> Thanks in advance
> Fabio
>
> ** Changed in: ubuntu
> Status: New => Incomplete
>
The problem appears to be gone or almost gone in Lucid. I ran benchmarks
and got the following results. The performance of 4 threads on 4 CPUs
(Intel Core i7) is very uniform for the CPU benchmarks and almost
uniform for memory benchmarks. The disparity in the memory benchmarks is
small but still significant. The test times below were 5 seconds for
each test (report row).
CPU Performance, MOPs/sec
loop time (nanosecs) 3.42
integer add/subtract 917 916 917 917
integer multiply/divide 314 314 314 314
float add/subtract 531 531 531 531
float multiply/divide 231 231 231 231
sqrt() function 88.2 88.2 88.1 88.2
pow() function 16.4 16.4 16.4 16.4
sin() function 41.2 41.2 41.2 41.2
asin() function 35.5 35.5 35.5 35.5
ready
CPU Performance, MOPs/sec
loop time (nanosecs) 3.41
integer add/subtract 916 916 916 916
integer multiply/divide 314 314 314 314
float add/subtract 531 531 531 531
float multiply/divide 231 231 231 231
sqrt() function 88.2 88.2 88.1 88.2
pow() function 16.4 16.4 16.4 16.4
sin() function 41.2 41.2 41.2 41.2
asin() function 35.5 35.5 35.4 35.5
ready
C-Lib Memory Move Performance
memory range megabytes/sec
10 KB 18778 17337 19508 18898
50 KB 12026 12047 10540 11981
200 KB 5714 4471 4896 5037
1000 KB 2306 2277 2305 2396
5000 KB 1422 1423 1421 1427
ready
C-Lib Memory Move Performance
memory range megabytes/sec
10 KB 18761 17341 19517 18911
50 KB 12119 12052 11895 12046
200 KB 4317 4204 4605 4481
1000 KB 2429 2426 2391 2389
5000 KB 1419 1423 1426 1418
ready
CPU Performance, MOPs/sec
loop time (nanosecs) 3.42
integer add/subtract 917 917 917 917
integer multiply/divide 314 314 314 314
float add/subtract 531 531 531 531
float multiply/divide 231 231 231 231
sqrt() function 88.2 88.2 88.2 88.2
pow() function 16.4 16.4 16.4 16.4
sin() function 41.2 41.2 41.2 41.2
asin() function 35.5 35.5 35.5 35.5
ready
CPU Performance, MOPs/sec
loop time (nanosecs) 3.41
integer add/subtract 916 916 916 916
integer multiply/divide 314 314 314 314
float add/subtract 531 531 531 531
float multiply/divide 231 231 231 231
sqrt() function 88.2 88.2 88.2 88.1
pow() function 16.4 16.4 16.4 16.3
sin() function 41.2 41.2 41.2 41.1
asin() function 35.5 35.5 35.4 35.4
ready
C-Lib Memory Move Performance
memory range megabytes/sec
10 KB 18281 18730 19500 18541
50 KB 12087 12128 12083 12063
200 KB 4603 4341 4429 4459
1000 KB 2412 2469 2474 2481
5000 KB 1423 1422 1427 1426
ready
C-Lib Memory Move Performance
memory range megabytes/sec
10 KB 18608 19503 17351 18837
50 KB 12154 12080 12120 12152
200 KB 5110 4353 4360 4429
1000 KB 2474 2474 2474 2476
5000 KB 1418 1420 1423 1426
ready
--
erratic thread scheduling, kernel 2.6.31
https://bugs.launchpad.net/bugs/459628
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
--
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
