[Fwd: Re: ia64 sched-domains initialisation]
Hi Jesse and Xavier,
Since last February we have implemented a patch for Linux that
considers the number of levels a NUMA computer has. We build the sched domains based on the number of levels, which we extract from the SLIT table. We have tested our code on a HP Superdome computer, and we are willing to share it with someone that would try it. We didn't share it before because we could not test our patch in a bigger machine. The machine we have access to is a 1 NUMA level.
Further, we modelled our strategy using analytical models and
simulation and the results we got are promissing. You can find our patch and copies of two papers that describe our strategy in the following address: http://www.inf.pucrs.br/~peso/loadBalancerEng.htm. We are preparing a paper on the simulation results we got (if you would like to get a copy of the unfinished version of this paper, we can send it directly to you).
It would be nice if you could give us some feedback on our approach.
Monica Correa
Quoted 30 lines · Jesse Barnes
Quoted 5 lines · Xavier Bru
An other alternative is using the
node_distance()that comes from the SLIT to build the sched domains instead of usingSD_NODES_PER_DOMAINon the platform.This option seems like the way to go. The size is far from optimally tuned as well, benchmarking is needed (unless we want boot time heuristics to determine the sched domain layout).
Quoted 5 lines
The following patch sets
SD_NODES_PER_DOMAINas a config/boot parameter, and when the value is 0 uses thenode_distanceto build the sched domains.Maybe you could make that the default behavior?
Jesse
-
To unsubscribe from this list: send the line "unsubscribe linux-ia64" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html