On August 15, 2001 04:10 pm, Marcelo Tosatti wrote:
> I've compared 2.4.8-ac4 performance with 2.4.9pre3 using postmark. I used
> this specific benchmark because it allows me to set read/write ratio of IO
> transactions.
>
> [...]
>
> The tests have shown that the 2.4.9pre is a bit slower than
> 2.4.8ac4+newzone.
I reaarranged the results to clarify...
> bias read 10 (approx 10/1 read/write ratio)
>
> 2.4.8-ac4-newzoned
> 45000.00 megabytes read (19.87 megabytes per second)
> 25230.00 megabytes written (11.14 megabytes per second)
>
> 2.9.4-pre3
> 45000.00 megabytes read (19.88 megabytes per second)
> 25230.00 megabytes written (11.14 megabytes per second)
The read-heavy results are virtually identical.
> bias read 1 (approx 1/10 read/write ratio)
>
> 2.4.8-ac4-newzoned
> 4380.00 megabytes read (6.75 megabytes per second)
> 25230.00 megabytes written (38.88 megabytes per second)
>
> 2.9.4-pre3
> 4380.00 megabytes read (6.35 megabytes per second)
> 25230.00 megabytes written (36.57 megabytes per second)
For even read/write balancing ac4-newzoned is 6% faster. It's still
anybody's race ;-)
-- Daniel - 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 : Wed Aug 15 2001 - 21:00:59 EST