| Message ID | 20200318150257.198402-1-maxg@mellanox.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | nvmet-rdma/srpt: SRQ per completion vector | expand |
On 2020-03-18 08:02, Max Gurtovoy wrote: > I used RoCE link layer. For SRP, I used 1 server with RoCE loopback connection > (results are not mentioned below) for testing. Hopefully I'll get a tested-by > signature and feedback from Laurence and Rupesh on the SRP part during the review > process. Hi Max, The MAD code in ib_srpt is not triggered when using RoCE. Please also test SRP over IB. Additionally, how does this patch series affect SRP performance? Thanks, Bart.
This set is a renewed version of the feature for NVMEoF/RDMA target. In this series I've decided to implement it also for SRP target that had similar implementatiom (SRQ per HCA) after previous requests from the community. The logic is intended to save resource allocation (by sharing them) and utilize the locality of completions to get the best performance with Shared Receive Queues (SRQs). We'll create a SRQ per completion vector (and not per device) using a new API (basic SRQ pool, added to this patchset too) and associate each created QP/CQ/channel with an appropriate SRQ. This will also reduce the lock contention on the single SRQ per device (today's solution). For NVMEoF, my testing environment included 4 initiators (CX5, CX5, CX4, CX3) that were connected to 4 subsystems (1 ns per sub) throw 2 ports (each initiator connected to unique subsystem backed in a different bull_blk device) using a switch to the NVMEoF target (CX5). I used RoCE link layer. For SRP, I used 1 server with RoCE loopback connection (results are not mentioned below) for testing. Hopefully I'll get a tested-by signature and feedback from Laurence and Rupesh on the SRP part during the review process. The below results were made a while ago using NVMEoF. Configuration: - Irqbalancer stopped on each server - set_irq_affinity.sh on each interface - 2 initiators run traffic throw port 1 - 2 initiators run traffic throw port 2 - On initiator set register_always=N - Fio with 12 jobs, iodepth 128 Memory consumption calculation for recv buffers (target): - Multiple SRQ: SRQ_size * comp_num * ib_devs_num * inline_buffer_size - Single SRQ: SRQ_size * 1 * ib_devs_num * inline_buffer_size - MQ: RQ_size * CPU_num * ctrl_num * inline_buffer_size Cases: 1. Multiple SRQ with 1024 entries: - Mem = 1024 * 24 * 2 * 4k = 192MiB (Constant number - not depend on initiators number) 2. Multiple SRQ with 256 entries: - Mem = 256 * 24 * 2 * 4k = 48MiB (Constant number - not depend on initiators number) 3. MQ: - Mem = 256 * 24 * 8 * 4k = 192MiB (Mem grows for every new created ctrl) 4. Single SRQ (current SRQ implementation): - Mem = 4096 * 1 * 2 * 4k = 32MiB (Constant number - not depend on initiators number) results: BS 1.read (target CPU) 2.read (target CPU) 3.read (target CPU) 4.read (target CPU) --- --------------------- --------------------- --------------------- ---------------------- 1k 5.88M (80%) 5.45M (72%) 6.77M (91%) 2.2M (72%) 2k 3.56M (65%) 3.45M (59%) 3.72M (64%) 2.12M (59%) 4k 1.8M (33%) 1.87M (32%) 1.88M (32%) 1.59M (34%) BS 1.write (target CPU) 2.write (target CPU) 3.write (target CPU) 4.write (target CPU) --- --------------------- --------------------- --------------------- ---------------------- 1k 5.42M (63%) 5.14M (55%) 7.75M (82%) 2.14M (74%) 2k 4.15M (56%) 4.14M (51%) 4.16M (52%) 2.08M (73%) 4k 2.17M (28%) 2.17M (27%) 2.16M (28%) 1.62M (24%) We can see the perf improvement between Case 2 and Case 4 (same order of resource). We can see the benefit in resource consumption (mem and CPU) with a small perf loss between cases 2 and 3. There is still an open question between the perf differance for 1k between Case 1 and Case 3, but I guess we can investigate and improve it incrementaly. Thanks to Idan Burstein and Oren Duer for suggesting this nice feature. Changes from v1: - rename srq_set to srq_pool (Leon) - changed srpt to use ib_alloc_cq (patch 4/5) - removed caching of comp_vector in ib_cq - minor fixes got from Leon's review Max Gurtovoy (5): IB/core: add a simple SRQ pool per PD nvmet-rdma: add srq pointer to rdma_cmd nvmet-rdma: use SRQ per completion vector RDMA/srpt: use ib_alloc_cq instead of ib_alloc_cq_any RDMA/srpt: use SRQ per completion vector drivers/infiniband/core/Makefile | 2 +- drivers/infiniband/core/srq_pool.c | 75 +++++++++++++ drivers/infiniband/core/verbs.c | 3 + drivers/infiniband/ulp/srpt/ib_srpt.c | 187 +++++++++++++++++++++++-------- drivers/infiniband/ulp/srpt/ib_srpt.h | 28 ++++- drivers/nvme/target/rdma.c | 203 ++++++++++++++++++++++++++-------- include/rdma/ib_verbs.h | 4 + include/rdma/srq_pool.h | 18 +++ 8 files changed, 419 insertions(+), 101 deletions(-) create mode 100644 drivers/infiniband/core/srq_pool.c create mode 100644 include/rdma/srq_pool.h