Why is it so difficult to find kernel code for Infiniband server/client? Can someone direct me to any sample infiniband server/client code? Secondly, what is the best source of help to write kernel space infiniband server/client code?
Hi,
though i might not give you an answer to this, i have a quick tip: register to Mellanox Academy site. (Take the associate free membership) They have a free RDMA programing course (and few other good courses) for folks at all levels.
Hope it helps ;-)
Thanks! It worked.
Hi,
I am wondering if there is any existing benchmark suite to measure IOPS. perftest provides benchmarks on bandwidth and latency, but does not include anything on IOPS.
The newest ConnectX-4 NIC claims to have 137M ops/s. Is there any program I can use to get such performance numbers?
Thank you,
Sam
Hi
We have the 10Gbe cards installed on our Dell Blades, we are using VMware EXSi 5.5. our current problem is that VMWare is only seeing the one adapter but these cards are duel. I have been troubleshooting with them for the past 2 weeks with no luck. They are advising that the driver and firmware need to match and be compatible with this version of ESXi. The version they are recommending is (driver 2.3.3.1 & firmware 2.30.51.00). I have now installed that driver version but cannot find the correct firmware version and how to even install it. the current firmware version installed is 2.30.51.60. I need urgent assistance on this if anyone can help me.
Thank you
Stan
Hi all,
I have a customer using this Mellanox cable:
Mellanox Breakout Cables 40GbE to 4x10GbE usage and configuration
and they need to provide temporary metallic loop-back, at the PCB connector into which the 40 GbE end of the above connector normally inserts, for debugging purposes.
Does anyone know of a loop-back plug, made by Mellanox or anyone else, that can provide this?
Thanks!
Hi, thanks a lot for your response.
This is my xml file:
<domain type='kvm'>
<name><VM_LABEL></name>
<uuid><VM_UUID></uuid>
<description>Timos VM</description>
<memory>62914560</memory>
<currentMemory>62914560</currentMemory>
<cpu mode='host-passthrough'/>
<memoryBacking>
<hugepages/>
</memoryBacking>
<numatune>
<memory mode='strict' nodeset='1'/>
</numatune>
<vcpu current='11'>11</vcpu>
<cputune>
<vcpupin vcpu='0' cpuset='13'/>
<vcpupin vcpu='1' cpuset='14'/>
<vcpupin vcpu='2' cpuset='15'/>
<vcpupin vcpu='3' cpuset='16'/>
<vcpupin vcpu='4' cpuset='17'/>
<vcpupin vcpu='5' cpuset='18'/>
<vcpupin vcpu='6' cpuset='19'/>
<vcpupin vcpu='7' cpuset='20'/>
<vcpupin vcpu='8' cpuset='21'/>
<vcpupin vcpu='9' cpuset='22'/>
<vcpupin vcpu='10' cpuset='23'/>
<emulatorpin cpuset='12'/>
</cputune>
<os>
<type arch='x86_64' machine='pc'>hvm</type>
<smbios mode='emulate'/>
</os>
<clock offset='utc'>
<timer name='pit' tickpolicy='delay'/>
<timer name='rtc' tickpolicy='delay'/>
</clock>
<devices>
<emulator>/usr/libexec/qemu-kvm</emulator>
<disk type='file' device='disk'>
<driver name='qemu' type='qcow2' cache='none'/>
<source file='/<TESTBED>/images/flexibed/<VM_NAME>/boot_cf3.qcow2'/>
<target dev='hda' bus='virtio'/>
</disk>
<interface type='bridge'>
<mac address='<VM_MAC_ETH0>'/>
<source bridge='breth0'/>
<model type='virtio'/>
</interface>
<interface type='bridge'>
<mac address='<VM_MAC_ETH1>'/>
<source bridge='breth1'/>
<model type='virtio'/>
</interface>
<!--SRIOVMODE -->
<interface type='hostdev' managed='yes'>
<mac address='<VM_MAC_ETH4>'/>
<source>
<address type='pci' domain='0x0000' bus='0x82' slot='0x0' function='0x1'/>
</source>
</interface>
<console type='tcp'>
<source mode='bind' host='0.0.0.0' service='<VM_CONSOLE>'/>
<protocol type='telnet'/>
<target type='virtio' port='0'/>
</console>
</devices>
<seclabel type='none'/>
</domain>
Regards,
Kavita
Hi,
I have several computers with ConnectX-3 cards in a 10Gb LAN. Sometimes they send packets to each other. These packets are small, about 100-200 bytes. My goal is to deliver and process them as quickly as it's possible.
I wonder if anyone know if there is any actual performance difference between VMA and DPDK in terms of latency?
Say, if I can sacrifice a core just for busy-polling, would I benefit from switching VMA to DPDK?
Could show me a exact model number?
You can find that on your back side sticker of your adapter.
MT27500
Hi Serge,
You can review the following tuning guide by Mellanox to help you with further tuning server and applications. In there you can also find instruction for how to "pin" a certain app to a certain core. also, how to tune specifically to a specific CPU type, OS flavor, etc..
Hope it helps... good luck
for example,
MCX354-FCBT for infiniband HCA or
MCX314A-BCCT for ethernet adapter...:)
Hello Yairi,
I have certainly red this guide, thanks. My question is not about the performance itself rather than the possible performance difference between VMA and DPDK. Are these two just an interface and do the same exactly thing with the underlying layer, or one of them is a bit closer to hardware thus providing an opportunity to improve the latency? Is there any benefit from switching to DPDK in my case?
Hi Philip,
Did you install on client and server Mellanox OFED version 3.3-1.0.4.0?
What is the kernel revision running on client and server?
Could you confirm the FW on the HCA's (both side client & server).
Putting aside NFS, are you able to validate an RDMA connection between client and server?
Regards,
Sophie.
Hello Sandeep,
Please use the link below to repository for missing packages for Step 7.
GitHub - SoftRoCE/librxe-dev: Development repository for RXE user space code.
Cheers,
.R
Is there any new driver for vSphere 6.0?
include iSER udpated driver, too!
I've got three ConnectX-3 cards and we're setting up a 40GbE network using these and an IBM Switch. Before our network guy comes in to help us configure it next week, I'm trying to get the cards installed in the three Windows 7 machines that will be running them.
Installation of WinOF 5.22 went smoothly, but when I check the Device Manager after rebooting, under "System Devices" I'm getting a yellow warning for the Mellanox ConnectX-3 VPI Adapter: The driver properties window reports:
"Windows cannot verify the digital signature for the drivers required for this device. A recent hardware or software change might have installed a file that is signed incorrectly or damaged, or that might be malicious software from an unknown source. (Code 52)"
Short of disabling signature checking altogether (which we could do but I'd rather not), is there a way to fix this?
Thanks!
We released today NEO 1.6.1 for general availability. This is an update release to main NEO 1.6 version from May, 2016.
Highlights
The software and full documentation is available under the NEO product page at MyMellanox site:
Hello Steve,
I bought this switch from Paul on Ebay to build a small private renderfarm for VR
and photogrammetry (for students and IT enthusiasts).
I can ping the MGMT Port but the browser (MS-IE11) can‘t open the WebUI page.
The software revision is:
U-Boot 2009.01 SX_PPC_M460EX SX_3.2.0330-82 ppc (Dec 20 2012 - 17:53:54)
Default image: 'SX_PPC_M460EX SX_3.3.5006 2014-05-20 12:19:44 ppc'
Is there a possibillity to get an MLNX-OS update for this switch?
Thank you!
Daniel
Anyone? I've tried disabling signature enforcement using:
bcdedit -set loadoptions DISABLE_INTEGRITY_CHECKS
bcdedit -set TESTSIGNING ON
But it makes no difference. The driver still doesn't appear to be loading, and I'm still getting the driver signature warning in the Device Manager.