This page is to Track the Initial Performance Number for Erasure Code:
Erasure_Code_Performance (Wolf vs Frontera)
Seq | Servers | Clients | IOR Process | Aggregation Disabled | ChunkSize | BlockSize | XferSize | Strip | access | Object Class | Write (MB/Sec) | Read (MB/Sec) | Date | Request | Notes |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1 | 4 | 10 | 200 | 1M | 32M | 1M | ------ | SSF | RP_1G2 | 3430.7 | 6843.38 | ||||
2 | 4 | 10 | 200 | 1M | 32M | 1M | ------ | SSF | RP_3G2 | 3084.85 | 18206.26 | ||||
3 | 4 | 10 | 200 | 32M | 32M | 2M | FULL | SSF | EC_2P1G1 | 3448.7 | 7087.37 | ||||
4 | 4 | 10 | 200 | 32M | 32M | 2M | FULL | SSF | EC_2P2G1 | 3445.28 | 6842.25 | ||||
5 | 4 | 10 | 200 | 32M | 32M | 1M | Partial | SSF | EC_2P1G1 | 3410.76 | 6430.81 | ||||
6 | 4 | 10 | 200 | 32M | 32M | 1M | Partial | SSF | EC_2P2G1 | 1729.42 | 6586.25 | ||||
7 | 4 | 10 | 200 | 1M | 32M | 1M | ------ | FPP | RP_1G2 | 20628.51 | 41909.55 | ||||
8 | 4 | 10 | 200 | 1M | 32M | 1M | ------ | FPP | RP_3G2 | 5079.58 | 43342.75 | ||||
9 | 4 | 10 | 200 | 32M | 32M | 2M | FULL | FPP | EC_2P1G1 | 14021.35 | 40296.71 | Need to add export FI_PSM2_CONN_TIMEOUT=30 at client side to work for most FPP | |||
10 | 4 | 10 | 200 | 32M | 32M | 2M | FULL | FPP | EC_2P2G1 | 9933.62 | 38209.18 | ||||
11 | 4 | 10 | 200 | 32M | 32M | 1M | Partial | FPP | EC_2P1G1 | 669.48 | 28914.79 | ||||
12 | 4 | 10 | 200 | 32M | 32M | 1M | Partial | FPP | EC_2P2G1 | 6423.27 | 27437.02 | ||||
13 | |||||||||||||||
14 | 6 | 10 | 200 | 1M | 32M | 1M | ------ | SSF | RP_3G4 | 6094.69 | 34479.7 | ||||
15 | 6 | 10 | 200 | 32M | 32M | 4M | FULL | SSF | EC_4P1G1 | 6824.54 | 14190.82 | ||||
16 | 6 | 10 | 200 | 32M | 32M | 4M | FULL | SSF | EC_4P2G1 | 6429.27 | 13979.85 | ||||
17 | 6 | 10 | 200 | 32M | 32M | 1M | FULL | SSF | DAOS_OC_EC_K4P2_L32K | 4172.45 | 9294.78 | ||||
18 | 6 | 10 | 200 | 32M | 32M | 128K | FULL | SSF | DAOS_OC_EC_K4P2_L32K | 1403.54 | 2600.15 | ||||
19 | 6 | 10 | 200 | 32M | 32M | 1M | FULL | SSF | DAOS_OC_EC_K4P2_L64K | 4013.9 | 9781.82 | ||||
20 | 6 | 10 | 200 | 32M | 32M | 128K | FULL | SSF | DAOS_OC_EC_K4P2_L64K | 1244.96 | 2415.02 | ||||
21 | 6 | 10 | 200 | 1M | 32M | 1M | ------ | FPP | RP_3G4 | 7188.65 | 60362.96 | ||||
22 | 6 | 10 | 200 | 32M | 32M | 4M | FULL | FPP | EC_4P1G1 | 25080.65 | 61916.86 | ||||
23 | 6 | 10 | 200 | 32M | 32M | 4M | FULL | FPP | EC_4P2G1 | 20152.52 | 61087.66 | ||||
24 | 6 | 10 | 200 | YES | 32M | 32M | 1M | FULL | FPP | DAOS_OC_EC_K4P2_L32K | 7326.18 | 23294.16 | |||
25 | 6 | 10 | 200 | YES | 32M | 32M | 1M | FULL | FPP | DAOS_OC_EC_K4P2_L64K | 12545.25 | 27339.02 | |||
26 | 6 | 10 | 200 | 8M | 8M | 1M | FULL | FPP | DAOS_OC_EC_K4P2_L32K | 11291.99 | 37411.33 | ||||
27 | |||||||||||||||
28 | 10 | 10 | 200 | 1M | 32M | 8M | ------ | SSF | RP_1G8 | 13845.6 | 28329.19 | ||||
29 | 10 | 10 | 200 | 32M | 32M | 8M | FULL | SSF | EC_8P2G1 | 12262.07 | 26527.38 | ||||
30 | 10 | 10 | 200 | 32M | 32M | 1M | Partial | SSF | EC_8P2G1 | 1698.5 | 4526.24 | ||||
31 | 10 | 10 | 200 | 32M | 32M | 1M | FULL | SSF | DAOS_OC_EC_K8P2_L32K | 7305.94 | 3880.9 | ||||
32 | 10 | 10 | 200 | 32M | 32M | 256K | FULL | SSF | DAOS_OC_EC_K8P2_L32K | 2438.23 | 3189.18 | ||||
33 | 10 | 10 | 200 | 32M | 32M | 8M | FULL | SSF | DAOS_OC_EC_K8P2_L32K | 5283.49 | 18123.56 | ||||
34 | 10 | 10 | 200 | 32M | 32M | 1M | FULL | SSF | DAOS_OC_EC_K8P2_L64K | 5869.3 | 13040.1 | ||||
35 | 10 | 10 | 200 | 32M | 32M | 256K | FULL | SSF | DAOS_OC_EC_K8P2_L64K | 1012.43 | 520.82 | ||||
36 | 10 | 10 | 200 | 32M | 32M | 8M | FULL | SSF | DAOS_OC_EC_K8P2_L64K | 10032.96 | 21768.8 | ||||
37 | 10 | 10 | 200 | 1M | 32M | 8M | ------ | FPP | RP_1G8 | 41363.64 | 60083.43 | ||||
38 | 10 | 10 | 200 | 32M | 32M | 8M | FULL | FPP | EC_8P2G1 | 35246.91 | 78044.88 | ||||
39 | 10 | 10 | 200 | 16M | 16M | 1M | Partial | FPP | EC_8P2G1 | 12679.74 | 75909.24 | ||||
40 | 10 | 10 | 200 | YES | 32M | 32M | 1M | FULL | FPP | DAOS_OC_EC_K8P2_L32K | 7706.1 | 23577.2 | |||
41 | 10 | 10 | 200 | YES | 32M | 32M | 8M | FULL | FPP | DAOS_OC_EC_K8P2_L32K | 16883.87 | 68451.2 | |||
42 | 10 | 10 | 200 | YES | 32M | 32M | 1M | FULL | FPP | DAOS_OC_EC_K8P2_L64K | 15362.67 | 29499.67 | |||
43 | 10 | 10 | 200 | YES | 32M | 32M | 8M | FULL | FPP | DAOS_OC_EC_K8P2_L64K | 13249.46 | 61664.51 | |||
44 | 10 | 10 | 200 | 8M | 8M | 1M | FULL | FPP | DAOS_OC_EC_K8P2_L32K | 23407.6 | 63203.37 | ||||
45 | |||||||||||||||
46 | 18 | 10 | 200 | 1M | 32M | 1M | ------ | SSF | RP_1G16 | 27322.27 | 54265.83 | 10/19/2020 | Reran on latest master 89f02db3700005079d5e76411f336e82ac50b35a | ||
47 | 18 | 10 | 200 | 1M | 32M | 1M | ------ | SSF | RP_1G16 | 26523.52 | 51431.10 | 10/20/2020 | On 10/19/2020 From Di to run with Debug Log | Reran on master 89f02db3700005079d5e76411f336e82ac50b35a with debug log and logs and IOR console output copied on Wolf /scratch/samirrav/Defect_logs/ior-2020-10-21_00-16-24-RP_1G16/ | |
48 | 18 | 10 | 200 | 64K | 32M | 64k | ------ | SSF | RP_1G16 | 14044.89 | 23006 | ||||
49 | 18 | 10 | 200 | 64K | 32M | 1M | ------ | SSF | RP_1G16 | 14180.71 | 25624.46 | ||||
50 | 18 | 10 | 200 | 32M | 32M | 16M | FULL | SSF | EC_16P2G1 | 25379.17 | 25194.69 | 10/19/2020 | Reran on latest master 89f02db3700005079d5e76411f336e82ac50b35a | ||
51 | 18 | 10 | 200 | 32M | 32M | 16M | FULL | SSF | EC_16P2G1 | 24766.26 | 46143.7 | 10/19/2020 | With xuezhao’s patch https://github.com/daos-stack/daos/pull/3648 and https://github.com/daos-stack/daos/pull/3690 | ||
52 | 18 | 10 | 200 | 32M | 32M | 16M | FULL | SSF | EC_16P2G1 | 25332.99 | 51189.53 | 10/20/2020 | On 10/19/2020 From Di to run with Debug Log | With xuezhao’s patch https://github.com/daos-stack/daos/pull/3648 and https://github.com/daos-stack/daos/pull/3690 Logs and IOR console output copied on Wolf /scratch/samirrav/Defect_logs/ior-EC_16P2G1-2020-10-21_00-48-17/ | |
53 | 18 | 10 | 200 | 32M | 32M | 16M | FULL | SSF | EC_16P2G1 | 24982.03 | 29195.24 | For Debug EC_16P2 fetch performance is not scaling on Wolf /scratch/samirrav/Defect_logs/RP_1G16_For_Debug/ | |||
54 | 18 | 10 | 200 | 32M | 32M | 1M | Partial | SSF | EC_16P2G1 | 4820.76 | 2444.39 | ||||
55 | 18 | 10 | 200 | 32M | 32M | 1M | FULL | SSF | DAOS_OC_EC_K16P2_L32K | 6374.05 | 2390.31 | 10/19/2020 | Reran all on latest master 89f02db3700005079d5e76411f336e82ac50b35a | ||
56 | 18 | 10 | 200 | 32M | 32M | 1M | FULL | SSF | DAOS_OC_EC_K16P2_L64K | 6242.84 | 2153.63 | 10/19/2020 | |||
57 | 18 | 10 | 200 | 32M | 32M | 16M | FULL | SSF | DAOS_OC_EC_K16P2_L64K | 18174.64 | 39528.03 | 10/19/2020 | |||
58 | 18 | 10 | 200 | 32M | 32M | 16M | FULL | SSF | DAOS_OC_EC_K16P2_L128K | 22452.41 | 40896.44 | 10/19/2020 | 89f02db3700005079d5e76411f336e82ac50b35a | ||
59 | 18 | 10 | 200 | 32M | 32M | 16M | FULL | SSF | DAOS_OC_EC_K16P2_L128K | 23092.06 | 40052.01 | 10/20/2020 | On 10/19/2020 From Di to run with Debug Log | 89f02db3700005079d5e76411f336e82ac50b35a Ran with debug log and logs and IOR console output copied on wolf /scratch/samirrav/Defect_logs/ior-DAOS_OC_EC_K16P2_L128K-2020-10-21_00-33-53 | |
60 | 18 | 10 | 200 | 32M | 32M | 16M | FULL | SSF | DAOS_OC_EC_K16P2_L256K | 21340.56 | 25030.99 | 10/19/2020 | 89f02db3700005079d5e76411f336e82ac50b35a | ||
61 | 18 | 10 | 200 | 32M | 128M | 16M | FULL | SSF | DAOS_OC_EC_K16P2_L256K | 23737.38 | 51349.88 | 10/21/2020 | With higher Block size IOR_log.txt | ||
62 | 18 | 10 | 200 | 32M | 32M | 16M | FULL | SSF | DAOS_OC_EC_K16P2_L512K | 24840.69 | 26299.4 | 10/19/2020 | 89f02db3700005079d5e76411f336e82ac50b35a | ||
63 | 18 | 10 | 200 | 32M | 128M | 16M | FULL | SSF | DAOS_OC_EC_K16P2_L512K | 22949.58 | 52070.04 | 10/21/2020 | With higher Block size IOR_log.txt | ||
64 | 18 | 10 | 200 | 32M | 32M | 16M | FULL | SSF | DAOS_OC_EC_K16P2_L512K | 23946.71 | 33672.49 | 10/20/2020 | On 10/19/2020 From Di to run with Debug Log | 89f02db3700005079d5e76411f336e82ac50b35a Ran with debug log and logs and IOR console output on wolf /scratch/samirrav/Defect_logs/ior-DAOS_OC_EC_K16P2_L512K-2020-10-21_00-37-08/ | |
65 | 18 | 10 | 200 | 32M | 32M | 1M | FULL | SSF | DAOS_OC_EC_K16P2_L64K | 9502.52 | 6734.47 | With Patch from Di to disable //obj_auxi->flags |= ORF_DTX_SYNC; | |||
66 | 18 | 10 | 200 | 1M | 32M | 1M | ------ | FPP | RP_1G16 | 74121.37 | 98367.39 | 10/19/2020 | Reran all on latest master 89f02db3700005079d5e76411f336e82ac50b35a | ||
67 | 18 | 10 | 200 | 1M | 32M | 1M | ------ | FPP | RP_1G16 | 78331.87 | 117012.32 | 11/3/2020 | Latest master d73374cb6cef61b830bd030a2b5d85791342d2d0 IOR_Console.txt | ||
68 | 18 | 10 | 200 | YES | 1M | 32M | 16M | ------ | FPP | RP_1G16 | 223.66 | 68781.28 | |||
69 | 18 | 10 | 200 | 64K | 32M | 64k | ------ | FPP | RP_1G16 | 50569.41 | 84687.21 | ||||
70 | 18 | 10 | 200 | 64K | 32M | 1M | ------ | FPP | RP_1G16 | 62811.66 | 89913.44 | ||||
71 | 18 | 10 | 200 | 32M | 32M | 16M | FULL | FPP | EC_16P2G1 | 16746.81 | 73468.09 | 10/19/2020 | Reran on latest master 89f02db3700005079d5e76411f336e82ac50b35a Error rendering macro 'jira' : Unable to locate Jira server for this macro. It may be due to Application Link configuration. | ||
72 | 18 | 10 | 200 | 32M | 32M | 16M | FULL | FPP | EC_16P2G1 | 16329.4 | 80395.84 | 10/19/2020 | With xuezhao’s patch https://github.com/daos-stack/daos/pull/3648 and https://github.com/daos-stack/daos/pull/3690 | ||
73 | 18 | 10 | 200 | 32M | 32M | 16M | FULL | FPP | EC_16P2G1 | 17129.21 | 123063.79 | 11/3/2020 | Latest master d73374cb6cef61b830bd030a2b5d85791342d2d0 Write is same compare to RP_1G16 so not going to open defect | ||
74 | 18 | 10 | 200 | 32M | 32M | 16M | FULL | FPP | EC_16P2G1 | With echo mode (Add this env in your server yml file DAOS_IO_BYPASS=target) dfs ERR src/client/dfs/dfs.c:968 open_sb() SB does not exist | |||||
75 | 18 | 10 | 200 | 8M | 8M | 1M | Partial | FPP | EC_16P2G1 | 21759.33 | 82566.56 | ||||
76 | 18 | 10 | 200 | YES | 32M | 32M | 1M | FULL | FPP | DAOS_OC_EC_K16P2_L32K | 19547.06 | 22717.03 | 10/19/2020 | Reran all on latest master 89f02db3700005079d5e76411f336e82ac50b35a | |
77 | 18 | 10 | 200 | YES | 32M | 32M | 1M | FULL | FPP | DAOS_OC_EC_K16P2_L64K | 18283.14 | 22995.8 | 10/19/2020 | ||
78 | 18 | 10 | 200 | YES | 32M | 32M | 16M | FULL | FPP | DAOS_OC_EC_K16P2_L64K | 14874.25 | 83421.25 | 10/19/2020 | ||
79 | 18 | 10 | 200 | YES | 32M | 32M | 16M | FULL | FPP | DAOS_OC_EC_K16P2_L128K | 16220.05 | 123331.59 | 10/20/2020 | IOR_Log.txt | |
80 | 18 | 10 | 200 | YES | 32M | 32M | 16M | FULL | FPP | DAOS_OC_EC_K16P2_L256K | |||||
81 | 18 | 10 | 200 | YES | 32M | 128M | 16M | FULL | FPP | DAOS_OC_EC_K16P2_L256K | Open new issue just to be sure it's not some thing in DAOS or CART side | ||||
82 | 18 | 10 | 200 | YES | 32M | 32M | 16M | FULL | FPP | DAOS_OC_EC_K16P2_L512K | 16520.71 | 73037.21 | 10/19/2020 | On latest master 89f02db3700005079d5e76411f336e82ac50b35a | |
83 | 18 | 10 | 200 | YES | 32M | 128M | 16M | FULL | FPP | DAOS_OC_EC_K16P2_L512K | 40705.93 | 127393.64 | 10/21/2020 | With higher 128M Blocksize IOR_log.txt | |
84 | This are SX object ran to verify the Network and system BW | ||||||||||||||
85 | 4 | 10 | 200 | 32M | 32M | 2M | SSF | SX | 20194.87 | 40302.62 | |||||
86 | 4 | 10 | 200 | 32M | 32M | 2M | FPP | SX | 21357.07 | 42227.33 | |||||
87 | 6 | 10 | 200 | 32M | 32M | 4M | SSF | SX | 25087.01 | 48330.26 | |||||
88 | 6 | 10 | 200 | 32M | 32M | 4M | FPP | SX | 28762.59 | 56743.31 | |||||
89 | 10 | 10 | 200 | 32M | 32M | 8M | SSF | SX | 39641.49 | 76674.21 | |||||
90 | 18 | 10 | 400 | 32M | 32M | 2M | SSF | SX | 71998.83 | 116292.9 | |||||
91 | 18 | 10 | 200 | 1M | 32M | 1M | FPP | SX | 79356.44 | 122079.5 | 10/2/2020 | 70b49b97ca40d596a0c98f28684378b159fdd66a as of 10/2/2020 | |||
92 | 18 | 10 | 200 | 1M | 32M | 1M | FPP | SX | 74514.32 | 127545.89 | 10/19/2020 | 70b49b97ca40d596a0c98f28684378b159fdd66a as of 10/19/2020 | |||
93 | 18 | 10 | 200 | 1M | 32M | 1M | FPP | SX | 77373.74 | 130906.87 | 10/19/2020 | 89f02db3700005079d5e76411f336e82ac50b35a as of 10/19/2020 | |||
94 | 18 | 10 | 200 | 1M | 32M | 1M | FPP | SX | 80206.02 | 110426.12 | 11/3/2020 | d73374cb6cef61b830bd030a2b5d85791342d2d0 on 11/3/2020 IOR_Console.txt | |||
95 | 16 | 10 | 200 | 1M | 32M | 1M | FPP | SX | 42897.06 | 129865.19 | 1/5/2021 | d7548abeaa0d1a94d6ed67373a894fed04e80a1c with verbs provider IOR_Console.txt |
Target Comparison on d73374cb6cef61b830bd030a2b5d85791342d2d0:
1 | 10 | 10 | 200 | 1M | 32M | 8M | ------ | FPP | RP_1G8 | 43421.06 | 96042.05 | 11/3/2020 | With targets: 16, nr_xs_helpers: 0 IOR_Console.txt | ||
2 | 10 | 10 | 200 | 1M | 32M | 8M | ------ | FPP | RP_1G8 | 33252.07 | 79346.13 | 11/3/2020 | With targets: 8, nr_xs_helpers: 8 IOR_Console.txt | ||
3 | 10 | 10 | 200 | 1M | 32M | 8M | ------ | FPP | RP_1G8 | 40803.24 | 62832.65 | 11/4/2020 | With targets: 16, nr_xs_helpers: 16 IOR_Console.txt | ||
4 | 10 | 10 | 200 | YES | 1M | 32M | 8M | ------ | FPP | RP_1G8 | 38090.32 | 54018.09 | 11/4/2020 | With targets: 16, nr_xs_helpers: 16 IOR_Console.txt + Reverting commit of "86330a6fc7dff4f78d8a625975d8a8eae900f2cb" | |
5 | 10 | 10 | 200 | 32M | 32M | 8M | FULL | FPP | EC_8P2G1 | 23595.63 | 91343.78 | 11/3/2020 | With targets: 16, nr_xs_helpers: 0 IOR_Console.txt | ||
6 | 10 | 10 | 200 | 32M | 32M | 8M | FULL | FPP | EC_8P2G1 | 26437.26 | 84216.77 | 11/3/2020 | With targets: 8, nr_xs_helpers: 8 IOR_Console.txt | ||
7 | 10 | 10 | 200 | 32M | 32M | 8M | FULL | FPP | EC_8P2G1 | 18385.84 | 55319.19 | 11/4/2020 | With targets: 16, nr_xs_helpers: 16 IOR_Console.txt | ||
8 | 10 | 10 | 200 | YES | 32M | 32M | 8M | FULL | FPP | EC_8P2G1 | 21954.21 | 63754.15 | 11/4/2020 | With targets: 16, nr_xs_helpers: 16 IOR_Console.txt + Reverting commit of "86330a6fc7dff4f78d8a625975d8a8eae900f2cb" | |
9 | 18 | 10 | 200 | 32M | 32M | 16M | FULL | FPP | EC_16P2G1 | 17129.21 | 123063.79 | 11/3/2020 | With targets: 16, nr_xs_helpers: 0 IOR_Console.txt | ||
10 | 18 | 10 | 200 | 32M | 32M | 16M | FULL | FPP | EC_16P2G1 | 14765.76 | 107413.18 | 11/3/2020 | With targets: 8, nr_xs_helpers: 8 IOR_Console.txt | ||
11 | 18 | 10 | 200 | 32M | 32M | 16M | FULL | FPP | EC_16P2G1 | 25576.86 | 105462.85 | 11/4/2020 | With targets: 16, nr_xs_helpers: 16 IOR_Console.txt | ||
12 | 18 | 10 | 200 | YES | 32M | 32M | 16M | FULL | FPP | EC_16P2G1 | 11/4/2020 | With targets: 16, nr_xs_helpers: 16 + Reverting commit of "86330a6fc7dff4f78d8a625975d8a8eae900f2cb" Hitting the same issue as |
Summary:
- Severs side fan-out has big performance impact
- Write is forwarded by EC group leader
- 16+2 write performance is not scaling well
- Read does not have the same issue
- Client has the same RPC fan-out as server, why?
- Fragmented RDMA
- Performance of 64K cell size is not good enough