Sinister Posted March 14, 2020 Share Posted March 14, 2020 Hello all i recently purchased 2 Dell Broadcom 57810 Dual Port 10Gb PCIe SFP+ for faster transfer speeds between my Windows 10 PC and my unRAID server running 6.8.2. I watched multiple guides and have googled an epic ton and believe i have setup everything properly but i am only able to get a speed of 400+ MB/s capping at 500. I have setup a ramdisk on the windows pc and tried transferring between a Unassigned Devices SSD, and have tried transferring between an SSD on the windows PC and an SSD on the unRAID server. i have also setup iperf3 and run the necessary test's and am only getting about maybe 3.89 to 4 GB bandwith speed. I have tweaked the settings of the MTU/jumbo frames and it absolutely makes no difference, what is it that i am missing ? Quote Link to comment
halfelite Posted March 14, 2020 Share Posted March 14, 2020 Couple things to try how many threads in iperf3? As you dont give much hardware specs are you sure the cards are in a full 8x slot. And not an 8x that shares pci lanes making it a 4x on bandwidth Quote Link to comment
Sinister Posted March 14, 2020 Author Share Posted March 14, 2020 13 minutes ago, halfelite said: Couple things to try how many threads in iperf3? As you dont give much hardware specs are you sure the cards are in a full 8x slot. And not an 8x that shares pci lanes making it a 4x on bandwidth For the windows PC the card is in a PCIE x16 slot and the same for the unRAID server the windows PC is the P8Z77-WS with 32 gigs of ram I7 3770k processor. unRAID server is supermicro X9DRI-LN4F+ with 64 GB of ECC ram and 2x INTEL XEON E5-2680V2 10 CORE 2.80GHz Quote Link to comment
JorgeB Posted March 14, 2020 Share Posted March 14, 2020 4 hours ago, halfelite said: Couple things to try how many threads in iperf3? Did you do this? What's the iperf max bandwidth with a single transfer? Quote Link to comment
Sinister Posted March 14, 2020 Author Share Posted March 14, 2020 49 minutes ago, johnnie.black said: Did you do this? What's the iperf max bandwidth with a single transfer? Im sorry i dont understand im not sure i understand how to run that test i followed the latest spaceinvaderone video tutorial regarding iperf i dont think he did anything in particular to get it working for him. I followed along and definitelydid not get the same result Quote Link to comment
JorgeB Posted March 14, 2020 Share Posted March 14, 2020 I didn't see the video, but iperf is very simple to use, if you can't follow the video try google. Quote Link to comment
bastl Posted March 14, 2020 Share Posted March 14, 2020 10 hours ago, Sinister said: 400+ MB/s capping at 500 10 hours ago, Sinister said: transferring between an SSD on the windows PC and an SSD on the unRAID server You're hitting the max speeds of the SSDs. If you need more than 500MB/s you need NVME drives for example on both sides or storage that is able to handle more than that. Quote Link to comment
Sinister Posted March 14, 2020 Author Share Posted March 14, 2020 8 hours ago, bastl said: You're hitting the max speeds of the SSDs. If you need more than 500MB/s you need NVME drives for example on both sides or storage that is able to handle more than that. that makes sense which is why spaceinvaderone used 2 ramdisks in his tutorial. but then my question is why do i get these speeds using iperf3 which ignores hard drives. my PC on the left unRAID server on the right Quote Link to comment
JorgeB Posted March 15, 2020 Share Posted March 15, 2020 Iperf tests the actual LAN bandwidth, you're unlikely to get significantly more speed during a single transfer than that result, it's usually the hardware used limiting that, NIC, cable, switch, even board/CPU combo. Quote Link to comment
Sinister Posted March 15, 2020 Author Share Posted March 15, 2020 10 minutes ago, johnnie.black said: Iperf tests the actual LAN bandwidth, you're unlikely to get significantly more speed during a single transfer than that result, it's usually the hardware used limiting that, NIC, cable, switch, even board/CPU combo. at this point im betting heavily on the cards not meeting expectations due to age do you think i should switch Quote Link to comment
Sinister Posted March 19, 2020 Author Share Posted March 19, 2020 So i have replaced the broadcom cards with Mellanox MCX311A-XCAT CX311A ConnectX-3 EN 10G Ethernet 10GbE SFP+ x2 I have created a ramdisk in windows 10 and also unraid. I have 64GB of ddr3 ecc memory in unraid server 32 GB of ram in windows machine. i could be wrong but wouldnt this test ignore any hardware limitations and yield better results ? Quote Link to comment
Vr2Io Posted March 20, 2020 Share Posted March 20, 2020 (edited) 3 hours ago, Sinister said: this test ignore any hardware limitations and yield better results ? iperf test shouldn't affect by how many memory. BTW, suggest ref below info. and make some more test, especially NUMA node and RSS setting and confirm no other NIC in same network domain. http://www.darrylvanderpeijl.com/windows-server-2016-networking-optimizing-network-settings/ Below are ConnectX-3 in Unraid and Intel X520 in Windows ( jumbo frame disable and both side single CPU system ) Connecting to host 192.168.9.181, port 5201 [ 4] local 192.168.9.182 port 37938 connected to 192.168.9.181 port 5201 [ ID] Interval Transfer Bandwidth Retr Cwnd [ 4] 0.00-1.00 sec 951 MBytes 7.97 Gbits/sec 0 325 KBytes [ 4] 1.00-2.00 sec 951 MBytes 7.98 Gbits/sec 0 279 KBytes [ 4] 2.00-3.00 sec 970 MBytes 8.14 Gbits/sec 0 277 KBytes [ 4] 3.00-4.00 sec 957 MBytes 8.03 Gbits/sec 0 277 KBytes [ 4] 4.00-5.00 sec 974 MBytes 8.17 Gbits/sec 0 277 KBytes [ 4] 5.00-6.00 sec 975 MBytes 8.18 Gbits/sec 0 277 KBytes [ 4] 6.00-7.00 sec 971 MBytes 8.14 Gbits/sec 0 277 KBytes [ 4] 7.00-8.00 sec 977 MBytes 8.19 Gbits/sec 0 277 KBytes [ 4] 8.00-9.00 sec 966 MBytes 8.10 Gbits/sec 0 277 KBytes [ 4] 9.00-10.00 sec 972 MBytes 8.15 Gbits/sec 0 277 KBytes - - - - - - - - - - - - - - - - - - - - - - - - - [ ID] Interval Transfer Bandwidth Retr [ 4] 0.00-10.00 sec 9.44 GBytes 8.11 Gbits/sec 0 sender [ 4] 0.00-10.00 sec 9.43 GBytes 8.10 Gbits/sec receiver Name : 乙太網路 2 InterfaceDescription : Intel(R) Ethernet Server Adapter X520-1 Enabled : True NumberOfReceiveQueues : 2 Profile : NUMAStatic BaseProcessor: [Group:Number] : 0:0 MaxProcessor: [Group:Number] : 0:6 MaxProcessors : 4 RssProcessorArray: [Group:Number/NUMA Distance] : 0:0/0 0:2/0 0:4/0 0:6/0 IndirectionTable: [Group:Number] : 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 0:0 0:2 Edited March 20, 2020 by Benson Quote Link to comment
JorgeB Posted March 20, 2020 Share Posted March 20, 2020 I'm also using Mellanox ConnectX-2 and 3, Microtik CRS309-1G-8S+IN switch, usually get around 8 and 9Gb/s: D:\temp\iperf>iperf3 -c 10.0.0.7 Connecting to host 10.0.0.7, port 5201 [ 4] local 10.0.0.50 port 52218 connected to 10.0.0.7 port 5201 [ ID] Interval Transfer Bandwidth [ 4] 0.00-1.00 sec 1.01 GBytes 8.67 Gbits/sec [ 4] 1.00-2.00 sec 896 MBytes 7.51 Gbits/sec [ 4] 2.00-3.00 sec 964 MBytes 8.09 Gbits/sec [ 4] 3.00-4.00 sec 1017 MBytes 8.53 Gbits/sec [ 4] 4.00-5.00 sec 1012 MBytes 8.49 Gbits/sec [ 4] 5.00-6.00 sec 1.00 GBytes 8.62 Gbits/sec [ 4] 6.00-7.00 sec 1023 MBytes 8.58 Gbits/sec [ 4] 7.00-8.00 sec 1.03 GBytes 8.82 Gbits/sec [ 4] 8.00-9.00 sec 984 MBytes 8.25 Gbits/sec [ 4] 9.00-10.00 sec 983 MBytes 8.24 Gbits/sec - - - - - - - - - - - - - - - - - - - - - - - - - [ ID] Interval Transfer Bandwidth [ 4] 0.00-10.00 sec 9.76 GBytes 8.38 Gbits/sec sender [ 4] 0.00-10.00 sec 9.76 GBytes 8.38 Gbits/sec receiver Quote Link to comment
Vr2Io Posted March 20, 2020 Share Posted March 20, 2020 (edited) SMB transfer 1.1GB/s max ( a 23GB file ), so it higher then iperf result. Edited March 20, 2020 by Benson Quote Link to comment
Sinister Posted March 21, 2020 Author Share Posted March 21, 2020 (edited) I have tried everything in the posted article i was able to understand and still i was unable to get anywhere near that speed C:\iperf3>iperf3 -c 10.10.20.199 Connecting to host 10.10.20.199, port 5201 [ 4] local 10.10.20.196 port 63119 connected to 10.10.20.199 port 5201 [ ID] Interval Transfer Bandwidth [ 4] 0.00-1.00 sec 426 MBytes 3.57 Gbits/sec [ 4] 1.00-2.00 sec 429 MBytes 3.60 Gbits/sec [ 4] 2.00-3.00 sec 444 MBytes 3.72 Gbits/sec [ 4] 3.00-4.00 sec 433 MBytes 3.63 Gbits/sec [ 4] 4.00-5.00 sec 450 MBytes 3.77 Gbits/sec [ 4] 5.00-6.00 sec 445 MBytes 3.74 Gbits/sec [ 4] 6.00-7.00 sec 454 MBytes 3.80 Gbits/sec [ 4] 7.00-8.00 sec 445 MBytes 3.74 Gbits/sec [ 4] 8.00-9.00 sec 446 MBytes 3.74 Gbits/sec [ 4] 9.00-10.00 sec 443 MBytes 3.72 Gbits/sec - - - - - - - - - - - - - - - - - - - - - - - - - [ ID] Interval Transfer Bandwidth [ 4] 0.00-10.00 sec 4.31 GBytes 3.70 Gbits/sec sender [ 4] 0.00-10.00 sec 4.31 GBytes 3.70 Gbits/sec receiver not sure what im doing wrong here when im using iperf to test and frames and every other setting has been checked. Cards have been changed and even transferring from ram disk to ram disk isnt working. id say only 2 times ever when transferring to a unassigned devices SSD did i see it go well over 1.09GB/s according to this post iperf is a ram only test without involving any kind of disk Edited March 21, 2020 by Sinister more info Quote Link to comment
Sinister Posted March 25, 2020 Author Share Posted March 25, 2020 So after some times i have made numerous changes and seem to still be having an issue, i have changed the mellanox card in the server to Mellanox MCX312A-XCBT CX312A ConnectX-3 EN 10GbE SFP+ Dual-Port PCI-E NIC. I have also purchased 2x Samsung 970 EVO Plus SSD 250GB - M.2 NVMe Interface Internal Solid State Drive with V-NAND Technology one in the desktop one in the desktop one in the server as an unassigned device. MTU settings 9014 on both cards. on the windows side both send and receive buffers 4096, interruption moderation disabled, receive side scaling enabled. Quote Link to comment
Vr2Io Posted March 26, 2020 Share Posted March 26, 2020 (edited) Could you make test on single CPU or other platform in both side, even Windows or Unraid. I make some test on Threadripper platform and setting with CPU pinning, the result different was small. [ ID] Interval Transfer Bandwidth Retr Cwnd [ 4] 0.00-1.00 sec 866 MBytes 7.27 Gbits/sec 0 399 KBytes [ 4] 1.00-2.00 sec 856 MBytes 7.18 Gbits/sec 0 365 KBytes [ 4] 2.00-3.00 sec 881 MBytes 7.39 Gbits/sec 2 405 KBytes [ 4] 3.00-4.00 sec 880 MBytes 7.38 Gbits/sec 0 376 KBytes [ 4] 4.00-5.00 sec 876 MBytes 7.35 Gbits/sec 2 422 KBytes [ 4] 5.00-6.00 sec 872 MBytes 7.32 Gbits/sec 0 362 KBytes [ 4] 6.00-7.00 sec 874 MBytes 7.33 Gbits/sec 0 405 KBytes [ 4] 7.00-8.00 sec 871 MBytes 7.31 Gbits/sec 0 385 KBytes [ 4] 8.00-9.00 sec 870 MBytes 7.30 Gbits/sec 2 436 KBytes [ 4] 9.00-10.00 sec 878 MBytes 7.36 Gbits/sec 4 368 KBytes - - - - - - - - - - - - - - - - - - - - - - - - - [ ID] Interval Transfer Bandwidth Retr [ 4] 0.00-10.00 sec 8.52 GBytes 7.32 Gbits/sec 10 sender [ 4] 0.00-10.00 sec 8.52 GBytes 7.32 Gbits/sec receiver But I notice some problem happen ( not relate pinning ), you can found some "retry" between test and this also indicate on switch side. Anyway I need fix it later. What iperf3 version you are running ? I notice you haven't "Retr" column. Edited March 26, 2020 by Benson Quote Link to comment
Vr2Io Posted March 26, 2020 Share Posted March 26, 2020 (edited) I boot Threadripper into baremetel Windows10, the iperf3 result really poor and unstable ( no any hardware change ), some help by NIC parameter tuning. Quote Connecting to host 192.168.9.181, port 5201 [ 4] local 192.168.11.188 port 50231 connected to 192.168.9.181 port 5201 [ ID] Interval Transfer Bandwidth [ 4] 0.00-1.00 sec 227 MBytes 1.90 Gbits/sec [ 4] 1.00-2.00 sec 219 MBytes 1.84 Gbits/sec [ 4] 2.00-3.00 sec 231 MBytes 1.93 Gbits/sec [ 4] 3.00-4.00 sec 211 MBytes 1.77 Gbits/sec [ 4] 4.00-5.00 sec 215 MBytes 1.81 Gbits/sec [ 4] 5.00-6.00 sec 206 MBytes 1.73 Gbits/sec [ 4] 6.00-7.00 sec 212 MBytes 1.78 Gbits/sec [ 4] 7.00-8.01 sec 213 MBytes 1.78 Gbits/sec [ 4] 8.01-9.00 sec 233 MBytes 1.96 Gbits/sec [ 4] 9.00-10.00 sec 217 MBytes 1.82 Gbits/sec - - - - - - - - - - - - - - - - - - - - - - - - - [ ID] Interval Transfer Bandwidth [ 4] 0.00-10.00 sec 2.13 GBytes 1.83 Gbits/sec sender [ 4] 0.00-10.00 sec 2.13 GBytes 1.83 Gbits/sec receiver Connecting to host 192.168.9.181, port 5201 [ 4] local 192.168.11.188 port 50233 connected to 192.168.9.181 port 5201 [ ID] Interval Transfer Bandwidth [ 4] 0.00-1.00 sec 320 MBytes 2.68 Gbits/sec [ 4] 1.00-2.00 sec 364 MBytes 3.05 Gbits/sec [ 4] 2.00-3.00 sec 340 MBytes 2.85 Gbits/sec [ 4] 3.00-4.00 sec 291 MBytes 2.44 Gbits/sec [ 4] 4.00-5.00 sec 318 MBytes 2.67 Gbits/sec [ 4] 5.00-6.00 sec 335 MBytes 2.80 Gbits/sec [ 4] 6.00-7.00 sec 346 MBytes 2.91 Gbits/sec [ 4] 7.00-8.00 sec 363 MBytes 3.04 Gbits/sec [ 4] 8.00-9.00 sec 324 MBytes 2.72 Gbits/sec [ 4] 9.00-10.00 sec 360 MBytes 3.02 Gbits/sec - - - - - - - - - - - - - - - - - - - - - - - - - [ ID] Interval Transfer Bandwidth [ 4] 0.00-10.00 sec 3.28 GBytes 2.82 Gbits/sec sender [ 4] 0.00-10.00 sec 3.28 GBytes 2.82 Gbits/sec receiver Finally, my NIC ( X520) haven't NUMA node assignment support, so I set RSS from default 8 -> 4 -> 1 or interruption moderation disabled then got a good result, but still slightly unstable. Quote Connecting to host 192.168.9.181, port 5201 [ 4] local 192.168.11.188 port 50414 connected to 192.168.9.181 port 5201 [ ID] Interval Transfer Bandwidth [ 4] 0.00-1.00 sec 958 MBytes 8.04 Gbits/sec [ 4] 1.00-2.00 sec 983 MBytes 8.25 Gbits/sec [ 4] 2.00-3.00 sec 986 MBytes 8.27 Gbits/sec [ 4] 3.00-4.00 sec 987 MBytes 8.28 Gbits/sec [ 4] 4.00-5.00 sec 984 MBytes 8.26 Gbits/sec [ 4] 5.00-6.00 sec 979 MBytes 8.21 Gbits/sec [ 4] 6.00-7.00 sec 983 MBytes 8.24 Gbits/sec [ 4] 7.00-8.00 sec 984 MBytes 8.25 Gbits/sec [ 4] 8.00-9.00 sec 980 MBytes 8.22 Gbits/sec [ 4] 9.00-10.00 sec 983 MBytes 8.25 Gbits/sec - - - - - - - - - - - - - - - - - - - - - - - - - [ ID] Interval Transfer Bandwidth [ 4] 0.00-10.00 sec 9.58 GBytes 8.23 Gbits/sec sender [ 4] 0.00-10.00 sec 9.58 GBytes 8.23 Gbits/sec receiver Connecting to host 192.168.9.181, port 5201 [ 4] local 192.168.11.188 port 50416 connected to 192.168.9.181 port 5201 [ ID] Interval Transfer Bandwidth [ 4] 0.00-1.00 sec 874 MBytes 7.33 Gbits/sec [ 4] 1.00-2.00 sec 880 MBytes 7.38 Gbits/sec [ 4] 2.00-3.00 sec 821 MBytes 6.89 Gbits/sec [ 4] 3.00-4.00 sec 813 MBytes 6.82 Gbits/sec [ 4] 4.00-5.00 sec 816 MBytes 6.84 Gbits/sec [ 4] 5.00-6.00 sec 803 MBytes 6.74 Gbits/sec [ 4] 6.00-7.00 sec 816 MBytes 6.85 Gbits/sec [ 4] 7.00-8.00 sec 810 MBytes 6.80 Gbits/sec [ 4] 8.00-9.00 sec 807 MBytes 6.77 Gbits/sec [ 4] 9.00-10.00 sec 805 MBytes 6.75 Gbits/sec - - - - - - - - - - - - - - - - - - - - - - - - - [ ID] Interval Transfer Bandwidth [ 4] 0.00-10.00 sec 8.05 GBytes 6.92 Gbits/sec sender [ 4] 0.00-10.00 sec 8.05 GBytes 6.92 Gbits/sec receiver Edited March 26, 2020 by Benson Quote Link to comment
Sinister Posted March 26, 2020 Author Share Posted March 26, 2020 2 hours ago, Benson said: Could you make test on single CPU or other platform in both side, even Windows or Unraid. I make some test on Threadripper platform and setting with CPU pinning, the result different was small. [ ID] Interval Transfer Bandwidth Retr Cwnd [ 4] 0.00-1.00 sec 866 MBytes 7.27 Gbits/sec 0 399 KBytes [ 4] 1.00-2.00 sec 856 MBytes 7.18 Gbits/sec 0 365 KBytes [ 4] 2.00-3.00 sec 881 MBytes 7.39 Gbits/sec 2 405 KBytes [ 4] 3.00-4.00 sec 880 MBytes 7.38 Gbits/sec 0 376 KBytes [ 4] 4.00-5.00 sec 876 MBytes 7.35 Gbits/sec 2 422 KBytes [ 4] 5.00-6.00 sec 872 MBytes 7.32 Gbits/sec 0 362 KBytes [ 4] 6.00-7.00 sec 874 MBytes 7.33 Gbits/sec 0 405 KBytes [ 4] 7.00-8.00 sec 871 MBytes 7.31 Gbits/sec 0 385 KBytes [ 4] 8.00-9.00 sec 870 MBytes 7.30 Gbits/sec 2 436 KBytes [ 4] 9.00-10.00 sec 878 MBytes 7.36 Gbits/sec 4 368 KBytes - - - - - - - - - - - - - - - - - - - - - - - - - [ ID] Interval Transfer Bandwidth Retr [ 4] 0.00-10.00 sec 8.52 GBytes 7.32 Gbits/sec 10 sender [ 4] 0.00-10.00 sec 8.52 GBytes 7.32 Gbits/sec receiver But I notice some problem happen ( not relate pinning ), you can found some "retry" between test and this also indicate on switch side. Anyway I need fix it later. What iperf3 version you are running ? I notice you haven't "Retr" column. I am using iperf version 3.1.3 on bare metal windows 10 pro Quote Link to comment
Sinister Posted March 28, 2020 Author Share Posted March 28, 2020 So an update. to further try and resolve this i ended up transferring the card into a shuttle PC which is currently running windows 7 and when doing speed tests the cards gave me 1GB going both directions between the server and client. and then thought to myself something must be wrong with my windows 10 install, I made a partition backup and re-installed the OS and tested the speed and speed was great this time between the original machine and server. I began to install programs one by one testing transfer speed immediately afterwards, when i finally got to my antivirus (avast) I found that when installed it slows down my LAN transfers to about 300MB/s and 200 going back to client PC even trying to add the network shares to be excluded from the program entirely it made no difference im not sure whats up with their firewall but i have googled and found multiple posts regarding this. and have not found what the solution could be 1 Quote Link to comment
JorgeB Posted March 29, 2020 Share Posted March 29, 2020 Thanks for reporting back, AV software can have a noticeable performance impact in various areas. Quote Link to comment
Sinister Posted March 29, 2020 Author Share Posted March 29, 2020 9 minutes ago, johnnie.black said: Thanks for reporting back, AV software can have a noticeable performance impact in various areas. But I would not think that it would have a strict firewall for a locally mapped drive and even in the settings no discernable way to resolve it. Even when you specify to.leave specific locations alone if I cant figure it out I'll certainly be moving on to something else that doesn't perform this way, I'm open to any suggestions for a new AV software Quote Link to comment
SuberSeb Posted September 27, 2020 Share Posted September 27, 2020 On 3/29/2020 at 12:42 PM, Sinister said: But I would not think that it would have a strict firewall for a locally mapped drive and even in the settings no discernable way to resolve it. Even when you specify to.leave specific locations alone if I cant figure it out I'll certainly be moving on to something else that doesn't perform this way, I'm open to any suggestions for a new AV software Do you solve that issue? I'm also have 3.5Gbps speeds instead of 10Gbps. 1 Quote Link to comment
Sinister Posted September 27, 2020 Author Share Posted September 27, 2020 8 hours ago, SuberSeb said: Do you solve that issue? I'm also have 3.5Gbps speeds instead of 10Gbps. Its not something i can fix used AVAST as an antivirus and firewall the only way to get around this issue was to remove the firewall component of the software completely. Not even disabling it was good enough Quote Link to comment
Bertel Posted November 9, 2020 Share Posted November 9, 2020 (edited) I am hunting down a similar perplexing situation. I have a powerful (16core Threadripper, 128G of RAM) Windows 10 machine connected to a hi-power (32core Threadripper, 256G of RAM) Ubuntu Linux machine via 10GBE network. A while ago, network speed between both machines became very slow. I booted the Windows box with Ubuntu, and iperf3 speed between the two machines was at 9.41 Gbit/sec. That ruled out hardware problems. I did a completely new install to Windows 10 Pro, Version 2004, OS build 19041.610 . That brought the iperf3 speed to around 2 Gbit/sec. Booting in Safe mode and shutting down all services that could be shut down did not improve matters until Zonealarm was shut down. Speed now at around 2.4 Gbit. After I completely removed Zonealarm and replaced it with Kaspersky, iperf3 speed was at around 3 Gbit/s, but that was as fast as I could tweak it. HOWEVER, when I booted the machine from an emergency a Windows 10 Pro, Ver 1909 build 18363.1139 install, the speed jumped to 6.14 Gigabit/second. All possible tweaks (jumbo frames, large buffers, etc.) applied with near zero impact. Changed and updated drivers, even swapped 10 GBE NICs on the Windows machine (Asus/Aquantia XG-C100C and Intel X540-T2, both in a x8 slot of the Windows machine). The fact that the same hardware runs at close to wire speed with Ubuntu on both sides, and that Ubuntu to Windows 10 Pro, Ver 1909 runs twice as fast as Ubuntu to Windows 10 Pro, Version 2004, tells me that something is wrong with the latest Windows. Reported to Microsoft, did not hear back. P.S.: The Zonealarm impact is perplexing. The 10GBE NICs and network segments were excluded from Zonealarm. Also note the speed difference between TURNING OFF Zonealarm and REMOVING Zonealarm. Edited November 9, 2020 by Bertel Quote Link to comment
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.