Ronciak, John writes:

 > So we still need to see a case where performance is hurt by the
 > prefetching.  We have some data coming from another group here at Intel
 > next week which we'll share once we have it which also shows the
 > performance gains with prefetching.

Hello!

Well here is another small test with prefetching in 6.2.15. Workload is 
still forwarding but now forcing both hash spinning, FIB lookups and GC 
to run as 32k concurrent flows at 10 packets each are injected. Routing 
table 123 kroutes 64 byte pkts. Dual Opteron 1.6 GHz with dual 82546GB  
at 100 resp. 133 MHz. Aggergated performance from both CPU's.

0 prefetch 756 kpps 
1 prefetch 805 kpps (first prefetch only)
2 prefetch 821 kpps (first two)
5 prefetch 803 kpps (all)

As seen this indicates an optimum with the two first prefetches in ths
setup. Seems I should have tested 3 and 4 as well...

Cheers.
                                        --ro



Input 32k concurrent flows @ 10 pkts each.
658 + 653 kpps. e1000 6.2.15 no prefetch == 756 kpps

Kernel Interface table
Iface   MTU Met  RX-OK RX-ERR RX-DRP RX-OVR  TX-OK TX-ERR TX-DRP TX-OVR Flags
eth0   1500   0 5366094      0 4633906      0      5      0      0      0 BRU
eth1   1500   0      1      0      0      0 5365737      0      0      0 BRU
eth2   1500   0 6189729      0 3810271      0      3      0      0      0 BRU
eth3   1500   0      2      0      0      0 6189374      0      0      0 BRU

0051e14f 00000000 000045e8 00000000 00000000 00000000 00000000 00000000 00000000
005e72a3 00000000 0000509e 00000000 00000000 00000000 00000000 00000000 00000000

661 + 659 kpps. e1000 6.2.15  first  prefetch == 805 kpps

Kernel Interface table
Iface   MTU Met  RX-OK RX-ERR RX-DRP RX-OVR  TX-OK TX-ERR TX-DRP TX-OVR Flags
eth0   1500   0 5736811      0 4263189      0      5      0      0      0 BRU
eth1   1500   0      1      0      0      0 5736454      0      0      0 BRU
eth2   1500   0 6476815      0 3523185      0      3      0      0      0 BRU
eth3   1500   0      1      0      0      0 6476394      0      0      0 BRU

0057896c 00000000 00004ab9 00000000 00000000 00000000 00000000 00000000 00000000
0062d410 00000000 0000545a 00000000 00000000 00000000 00000000 00000000 00000000


661 + 655 kpps. e1000 6.2.15  two first  prefetch == 821 kpps

Kernel Interface table
Iface   MTU Met  RX-OK RX-ERR RX-DRP RX-OVR  TX-OK TX-ERR TX-DRP TX-OVR Flags
eth0   1500   0 5994571      0 4005429      0      5      0      0      0 BRU
eth1   1500   0      1      0      0      0 5994214      0      0      0 BRU
eth2   1500   0 6507839      0 3492161      0      5      0      0      0 BRU
eth3   1500   0      1      0      0      0 6507246      0      0      0 BRU

005b784c 00000000 00004e13 00000000 00000000 00000000 00000000 00000000 00000000
00634d40 00000000 000054c2 00000000 00000000 00000000 00000000 00000000 00000000

655+ 661 kpps. e1000 6.2.15  all five  prefetch == 803 kpps


Kernel Interface table
Iface   MTU Met  RX-OK RX-ERR RX-DRP RX-OVR  TX-OK TX-ERR TX-DRP TX-OVR Flags
eth0   1500   0 6081924      0 3918076      0      5      0      0      0 BRU
eth1   1500   0      1      0      0      0 6081331      0      0      0 BRU
eth2   1500   0 6149019      0 3850981      0      3      0      0      0 BRU
eth3   1500   0      2      0      0      0 6148663      0      0      0 BRU

005ccd85 00000000 00004f35 00000000 00000000 00000000 00000000 00000000 00000000
005dd39d 00000000 00005017 00000000 00000000 00000000 00000000 00000000 00000000
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to