From: Lance G. <ei...@pa...> - 2014-12-18 23:42:08
|
Hello Quinn: So I finally integrated your script into my script, and have some good news and bad news. The good news is this completely fixed the problem with Windows getting an “unknown error,” and the write performance is great at any connection rate. The bad news is I am getting an error at the BIOS level of the client, which reads “Adapter 2 mailbox error 0001”, and the read performance with a 16Gbit connection is still bad. Here’s the performance rates I got – I tested 4Gbit/8Gbit/16Gbit, with Iometer, 8MB transfer lengths with 16 outstanding I/Os: Internal rates on the server W:3.1GB/sec, R:3.3GB/sec Rates from Initiator: 4Gbit W:412MB/sec, R:406MB/sec 8Gbit W:823MB/sec, R:772MB/sec 16Gbit W:1577MB/sec, R:745MB/sec Any ideas on how to fix the 16Gbit read rate? Lance From: Lance Gropper Sent: Monday, December 15, 2014 2:40 PM To: Quinn Tran ; scs...@li... Subject: Re: [Scst-devel] 16Gbit not working Hello Quinn: Also doesn’t look like the –s option is working – after the command is executed with –s, the interrupts still seem to be funneled to CPU 0: [root@QOS ~]# ./split_irq.sh -s before: 34 : ff after : 34 : 11 before: 37 : ff after : 37 : 22 before: 35 : ff after : 35 : 44 before: 43 : ff after : 43 : 88 before: 36 : ff after : 36 : 11 before: 44 : ff after : 44 : 22 [root@QOS ~]# cat /proc/interrupts CPU0 CPU1 CPU2 CPU3 CPU4 CPU5 CPU6 CPU7 0: 127 0 0 0 0 0 0 0 IO-APIC-edge timer 1: 2 0 0 0 0 0 0 0 IO-APIC-edge i8042 3: 2 0 0 0 0 0 0 0 IO-APIC-edge 4: 2 0 0 0 0 0 0 0 IO-APIC-edge 8: 1 0 0 0 0 0 0 0 IO-APIC-edge rtc0 9: 0 0 0 0 0 0 0 0 IO-APIC-fasteoi acpi 10: 2 0 0 0 0 0 0 0 IO-APIC-edge 12: 4 0 0 0 0 0 0 0 IO-APIC-edge i8042 16: 61 0 0 0 0 0 0 0 IO-APIC-fasteoi ehci_hcd:usb1 23: 26 0 0 0 0 0 0 0 IO-APIC-fasteoi ehci_hcd:usb2 24: 5 0 0 0 0 0 0 0 HPET_MSI-edge hpet2 25: 0 0 0 0 0 0 0 0 HPET_MSI-edge hpet3 26: 0 0 0 0 0 0 0 0 HPET_MSI-edge hpet4 27: 0 0 0 0 0 0 0 0 HPET_MSI-edge hpet5 28: 0 0 0 0 0 0 0 0 HPET_MSI-edge hpet6 32: 1367 0 0 0 0 0 0 0 PCI-MSI-edge mpt2sas0-msix0 33: 1178 0 0 0 0 0 0 0 PCI-MSI-edge mpt2sas1-msix0 34: 43 0 0 0 0 0 0 0 PCI-MSI-edge qla2xxx (default) 35: 0 0 0 0 0 0 0 0 PCI-MSI-edge qla2xxx (rsp_q) 36: 0 0 0 0 0 0 0 0 PCI-MSI-edge qla2xxx (atio_q) 37: 43 0 0 0 0 0 0 0 PCI-MSI-edge qla2xxx (default) 38: 5812 0 0 0 0 0 0 0 PCI-MSI-edge ahci 39: 17 0 0 0 0 796 0 0 PCI-MSI-edge eth0 40: 51 0 0 0 0 54 5960 0 PCI-MSI-edge eth1-rx-0 41: 29 0 0 0 0 0 15 898 PCI-MSI-edge eth1-tx-0 42: 2 0 0 0 0 0 0 0 PCI-MSI-edge eth1 43: 0 0 0 0 0 0 0 0 PCI-MSI-edge qla2xxx (rsp_q) 44: 0 0 0 0 0 0 0 0 PCI-MSI-edge qla2xxx (atio_q) NMI: 0 0 0 0 0 0 0 0 Non-maskable interrupts LOC: 27879 15758 8618 7856 17079 17767 19662 12957 Local timer interrupts SPU: 0 0 0 0 0 0 0 0 Spurious interrupts PMI: 0 0 0 0 0 0 0 0 Performance monitoring interrupts IWI: 0 0 0 0 0 0 0 0 IRQ work interrupts RES: 256 254 521 696 408 219 344 382 Rescheduling interrupts CAL: 275 418 417 385 324 414 345 409 Function call interrupts TLB: 142 265 191 142 527 1241 901 604 TLB shootdowns TRM: 0 0 0 0 0 0 0 0 Thermal event interrupts THR: 0 0 0 0 0 0 0 0 Threshold APIC interrupts MCE: 0 0 0 0 0 0 0 0 Machine check exceptions MCP: 7 7 7 7 7 7 7 7 Machine check polls ERR: 0 MIS: 0 Lance From: Quinn Tran Sent: Friday, December 12, 2014 9:43 AM To: Lance Gropper ; scs...@li... Subject: Re: [Scst-devel] 16Gbit not working Lance, The interrupts on the 16G adapter are different than the 8G adapter. There's additional interrupts to handle the higher load. Some OS/kernel does a good job of spreading the Interrupts across different CPUs to spread the work load. Some don't spread it. This looks like a case of where all interrupt lines are funneled to CPU 0. This require fine tuning for specific environment that you're in to get the best perf. Attached is a script I created for this case: a.. -l option: show where the interrupts are routing to by watching highlighted b.. -s option: spread the interrupts across different cpu. Use this option before test start. You're welcome to modify the script to fit your need. The setting does not persist across driver re-load or reboot. Give this a try. Will include this script into the Qlogic git tree in the future. Regards, Quinn Tran From: Lance Gropper <ei...@pa...> Date: Thursday, December 11, 2014 4:42 PM To: "scs...@li..." <scs...@li...> Subject: [Scst-devel] 16Gbit not working Hello All: I got the latest Qlogic GIT for the 16Gbit target drivers – everything builds and loads with no problems. If I attach an 8Gbit Initiator, no problems, and I can saturate the pipes at about 760MB/sec on reads and writes. If, however, I use a 16Gbit initiator, I only get about 277MB/sec on reads and writes. What’s the deal? Lance -------------------------------------------------------------------------------- This message and any attached documents contain information from the sending company or its parent company(s), subsidiaries, divisions or branch offices that may be confidential. If you are not the intended recipient, you may not read, copy, distribute, or use this information. If you have received this transmission in error, please notify the sender immediately by reply e-mail and then delete this message. -------------------------------------------------------------------------------- ------------------------------------------------------------------------------ Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server from Actuate! Instantly Supercharge Your Business Reports and Dashboards with Interactivity, Sharing, Native Excel Exports, App Integration & more Get technology previously reserved for billion-dollar corporations, FREE http://pubads.g.doubleclick.net/gampad/clk?id=164703151&iu=/4140/ostg.clktrk -------------------------------------------------------------------------------- _______________________________________________ Scst-devel mailing list https://lists.sourceforge.net/lists/listinfo/scst-devel |