DBT2-MySQL Benchmark. "hugemem" vs "smp" kernels

[Peter Zaitsev]

Hi,

Some half a year ago I wrote about my experience of performance of
"hugemem" vs "sm" kernels for 2.4.21-9.EL kernel from RH AS 3.0

The results I got previously were 

                 smp          hugemem 
IO BOUND:      1450TPM        1250TPM
CPU BOUND:     7000TPM        4500TPM


Results are in Transactions per Minute.

This was significant regression and it was some kind of expected.

The results with newer kernel 2.4.21-15.EL  are very strange however:


                 smp          hugemem 
IO BOUND:      1150TPM        1650TPM
CPU BOUND:     6300TPM        5000TPM


Well I'm pretty happy to see "hugemem" kernel improved a lot, but 
what happened with SMP kernel ? Why it is slow now for IO bound case 
and why it became slower than 2.4.21-9.EL ? 




I'm using 4*Pentium Xeon 2.0Ghz w HT,  4GB RAM,  3WARE RAID10, EXT3


-- 
Peter Zaitsev, Senior Support Engineer
MySQL AB, www.mysql.com



-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/