MikroTik Cracked With Key Download Free Working. Mikrotik v6.46.6 Crack is advance technology routers system which is developed with specula structures and tools to work with special perspective of Intel and Core pcs. This is a RouterOS system which is broadly used in the world. It automatically starts work and then converts your simple computer into an electronic router which performs the.
CHR do not offer satisfactory setup in our case in term of resources, install process and stability. We already tried.
Also, there is issue if you use 10G interfaces. In x86 mikrotik every 10G card can have ALL Cores of CPU for processing. So if you use CPU with 8 cores, all 8 can process every such 10G NIC. If you use CHR, you have only one core per 10G NIC, which is terrible in performance. That is WHY, 64bit version of mikrotik is needed asap.As 32bit offer possibility to use up to 4GB of RAM it will be maybe satisfactory, but 2GB which mikrotik see with 32bit version is very low, especially if there is more full bgp routing tables.We see some strange memory leak also, but hopefully mikrotik support will handle it.
Possible it is related to quad channel motherboard.Anyway we need x8664bit version asap. They should stop x86 and switch to x8664bit.
CHR do not offer satisfactory setup in our case in term of resources, install process and stability. We already tried. Also, there is issue if you use 10G interfaces. In x86 mikrotik every 10G card can have ALL Cores of CPU for processing.
So if you use CPU with 8 cores, all 8 can process every such 10G NIC. If you use CHR, you have only one core per 10G NIC, which is terrible in performance. That is WHY, 64bit version of mikrotik is needed asap.As 32bit offer possibility to use up to 4GB of RAM it will be maybe satisfactory, but 2GB which mikrotik see with 32bit version is very low, especially if there is more full bgp routing tables.We see some strange memory leak also, but hopefully mikrotik support will handle it. Possible it is related to quad channel motherboard.Anyway we need x8664bit version asap. They should stop x86 and switch to x8664bit.I totally agree with you.Daniele.
I also find myself not using virtualisation either. While i use a routerboard rather than x86, routerOS on TILE is 64 bit being able to address large amounts of RAM.On x86 theres so much hardware that there is the issue of driver support however when it comes to linux, driver support isnt an issue if you could add drivers to it yourself rather than MT's closed policy of not letting anyone mess with the linux bit of routerOS. In the past there were utilities that allowed you to use windows drivers for wifi NICs on linux as well.The main reason why i dont use virtualisation is because the overhead of virtualisation is huge on memory performance, not for CPU performance rather. Since i use my servers for things like game servers, compilations and for massive parallisation virtualisation would hurt performance significantly as memory performance is needed for moving data between other things like GPUs, 10G NICs and such. On routing performance if you look at the CCR1072 it seems that memory performance does effect routing performance as well as shown on the benchmarks. For routing memory performance does matter when you are moving data at speeds significantly proportional to memory bandwidth. This means that if you are using DDR3 dual channel for a 1Gb/s connection, using virtualisation wont hurt but at 10Gb/s it can.on x86, using a 64 bit OS is faster than a 32 bit OS, this is because of the extensions that allow multiple data to be processed at the same time.
So by using 64 bits your instructions will be bigger but will still be fed into the CPU the same rate as if you used a 32 bit OS with 32 bit length data. So using 8 or 16 bit data on 64 bit x86 does not incur any penalty as x86 is a very complex instruction set with extended instructions to accelerate larger data processing focusing on IPCs. On GPUs however using 16 bit integers are slower than using 32 floats with the exception of nvidia's pascal architecture and on AMD before GCN using 16 bit floats, not integers was faster than 32 bit floats.I really think MT needs to look into how the linux distros get their drivers. A lot of vendors will willingly write drivers for routerOS as it uses linux as long as the difference isnt great. I suspect that part of the problem is that a number of drivers require libraries that routerOS will not include in its OS. A 64 bit routerOS for x86 will really be beneficial and people who see routerOS for the first time prefer trying it out on x86 in an office environment rather than buying a routerboard as they have servers laying around. It sounds hard, but I think the only reason is 'they don't want'.
As far as I know there is no technical reason why x86 only accepts 2GB. It just let sell CCRs better.there is a reason why RouterOS on x86 supports only 2GB - the speed of memory addressing.
With high/low setups you would lose 5 to 10% of performance.It is entirely dependent on the memory controller. In the past that may be so when extended memory support wasnt available. From the last decade every PC has extended memory support in the controller itself allowing for more than 4GB of ram to be addressed in 32 bit mode especially if the CPU is a 64 bit architecture as there wont be any performance penalty in that case.Besides even if it isnt true and you get performance penalty, not having enough ram will cause even more performance issues. It sounds hard, but I think the only reason is 'they don't want'. As far as I know there is no technical reason why x86 only accepts 2GB.
It just let sell CCRs better.there is a reason why RouterOS on x86 supports only 2GB - the speed of memory addressing. With high/low setups you would lose 5 to 10% of performance.5-10% it's not a problem at all. I doubt this will be the case but as we have 1-20% load on x86 it will not be an issue. On the other hand, 2GB of RAM which ROS see, is a big PROBLEM.As SystemErrorMessage said, nowadays memory controllers are very fast, so we will not lose any of performance. 5-10% penalty from memory is if you use registered/buffered ram for those using large amount of ram.
So unless you plan to have more than 80GB of ram this doesnt matter. The memory controller will work at the same speed relatively regardless of the amount of ram for the memory controller on the CPU. Since it is on the CPU it is so fast with very fast bus to the CPU that the amount of memory, pointer sizes and such dont matter.Look at the x86 architecture now, the memory controller will be just as fast regardless of memory size as for a 64 bit CPU, the pointer size will also be 64 bits.
This isnt trying to use extended memory on a 32 bit system where extra processing is added to the controller.you say 5-10% less memory performance. Lets take a typical dual channel memory system, you have 2x1GB vs 2x4GB, both have the same frequency, timings, etc. Will addressing all the 8GB really make it slower?
I mean if the ram bandwidth is 32GB/s does it matter how much memory because im pretty sure even with extra load that memory controller is still gonna work at 32GB/s limited by the ram itself and not the controller.As i said not having enough RAM is a bigger performance killer than an insignificant difference in ram speed from having more ram if such even exists. Nowadays memory controllers are very fast, so we will not lose any of performance.You will, in fact. And that is a well known fact. Nowadays memory controllers are very fast, so we will not lose any of performance.You will, in fact. And that is a well known fact. I got an idea Normis.How about this.Eventually deprecate the x86 installation.
Then go to something like The Yocto Project and build a very lean Linux VM hypervisor but only for a singular VM, the CHR VM. Make it a completely automated installation and have it setup as a baseline Linux/KVM with one CHR host.That way there's always a stable, sustainable, and really fast hypervisor/CHR host. As well as a completely separated and uninterrupted RouterOS installation that can be tuned without having to constantly add tons and tons of drivers. Then one can let the Linux kernel to add drivers and said drivers for said devices can be imported to the CHR through the VM hypervisor as generic interfaces.It almost seems like a triple win.
Really lean VM host + non-changing RouterOS/CHR environment + abstraction of all device drivers as generic interfaces that can be further optimized.The.ONLY. thing that would need to happen is that there.needs. to be better 10Gb/sec optimizations between CHR and the hardware so that one can actually achieve those data rates. It seems that krzysztof manually changed architecture-name: x86 to architecture-name: x8664 only in post.he also 'changed' free- and total-memorynow just look at 'Extra packages' links in Download page.
This.zip contains everything including 'system' package (which contains the kernel, I believe )X86 link:CHR link:now find the differenceso seems like 'x8664' is just a flag somewhere in the file systemby the way, I have old good RouterOS virtual machine which was installed when no CHR existed. Code: admin@TestPlace system resource printuptime: 1w1d2h51m41sversion: 6.39rc38 (testing)build-time: Feb/24/2017 08:46:35free-memory: 21.7MiBtotal-memory: 88.4MiBcpu: Intel(R)cpu-count: 1cpu-frequency: 2933MHzcpu-load: 0%free-hdd-space: 58.8MiBtotal-hdd-space: 99.2MiBwrite-sect-since-reboot: 394778write-sect-total: 394778architecture-name: x8664board-name: x86platform: MikroTikin some early v6 versions there was a checkbox under 'System - Resources - Hardware' for enabling 64-bit mode. Looks like they removed it and created the CHR as a separate system. Code: admin@MikroTik sys resource printuptime: 55sversion: 6.31build-time: Aug/14/2015 15:42:51free-memory: 233.8MiBtotal-memory: 250.0MiBcpu: Intel(R)cpu-count: 1cpu-frequency: 2933MHzcpu-load: 0%free-hdd-space: 65.6MiBtotal-hdd-space: 99.1MiBwrite-sect-since-reboot: 614write-sect-total: 614architecture-name: x86board-name: x86platform: MikroTik(look at 'cpu-count: 1' - probably, that's the reason why download links for that version don't work)go to System - Resource - Hardware - check 'Allow x86-64' (in WinBox. Can't find it in CLI). Rebootand voila!
Oops, 'architecture-name: x86' no worry, just upgrade to the latest version. X8664 - This is super secret.RouterOS X8664 with BGP works around 200-300% faster than CCR1036 and CCR1072.It is important to turn off Hyper-threading!X8664 works just fine with the platforms where intel is on the motherboard is very important!For example, the X8664 works well with the HP DL180G6 plus 2 x Intel® Xeon® Processor X5675 (12M Cache, 6 x 3.06 GHz, 6.40 GT / s)CCR1036 and CCR1072 for BGP do not differ in performance.In RouterOS for BGP uses only one core! That's why X8664 is so powerful.Normis please do not liquidate X8664.Without X8664 our computer networks will not work well.RegardsKrzysztof Pawluk. Guys is there any real.benefit. from running CHR in x8664 native environment?I know VM puts some overhead, but c'mon, that's not 2006, It's 2017 when most of hardware, including CPUs and NICs have hardware virtualization support.Overhead is small nowadays, I am running two BGP servers and they are just work perfect in CHR.Plus, you can also monitor your CPU/mem usage, and add some additional, low-priority tasks like DDOS monitoring or anything you like, and mirorr all traffic inside VM kernel, no need to occupy switch space.Plus you just dump your VM image for backup/restore schemas. I think VM is.correct.
future, unless you forward really huge amount on traffic, but then you already need specialised routing or MPLS asic for that.Also you just use vmxnet3 driver (if not passthroughing), and don't care about underlaying hardware in most of cases.I just think that there's just more advantages than disadvantages of this way.my 2 cents. Hello to all,I was ready this thread very carefully and also 'between the lines' and must now please enter that discussion about the x8664Bit, because there are some thingsabout that must be spoken if you are talking about MikroTik and x8664Bit. (Only in my opinion)I company rules, from who given, dictated or written, all employees of a company should and must follow! And this is surely not on at MikroTik in Latvia so,because we all have wishes, demands and/or needs that could be 'easily' solved out by a x8664Bit router software, it means that this must be not seenon both ends of that discussion. And yes, I respect the standing point of MikroTik Latvia and perhaps their partners or suppliers, but for sure it will be fineto get in deeper zone of this forum thread. At first I want to say thanks to MikroTik to give us a forum and podium to discuss this out and that freely and onthe other side because I really know many of us are in business contact with MikroTik and their complete product portfolio and being not only home users.But then money comes into the game play and the company MikroTik must also see what they will get from their clients and customers, for sure it soundsnice to get hands on a MikroTik RouterOS x8664Bit version, but is it worth to play then with the whole company security? Who gives them a promise thatonly one copy will be used or if someone will be installing one system and copy it over and over and over only with one License code?
What is then? Whocares then on their business and an their income or plain on their money? For sure getting hands on a copy will be fine but how many users will then tradeover P2P networks this registered copy then?I personally think x8964Bit would be a really hammer smashing into the whole market and many router and firewall vendors must find then a way to answerthat! This could be a game changer, a cash cow and many things more for Mikrotik and for sure also us. But please also think on the other side of the business!At this moment you will be able to see many changes in the market, AES-NI support here and there, Intel´s QuickAssist (QAT) is out and drivers are there forLinux, DPDK (enabled software) could be transporting more then 3x the amount of IP packets then normal, and this using the same platform for sure. Multi-CPUcore usage and HT technology will be then doing the rest to get closer and come to a really routing champion device. HotLava is producing multi Port NICs withoriginal Intel Chipsets and that will lets gain up the entire port density of many routers for sure.As for the memory lag and speed, this must be given both, the speed and amount must be matching right to that units, but with 64Bit no problem the most of us arethinking.
I mean really if there is not enough amount of RAM you can get surely into pain, but on the other side if the RAM is to low clocked the entire router systemcan be saturated and you will be owning a freezing or lacking router then! Getting or producing huge packet losses or packet drops will be the result of that behaviour.At this time we will see many existing and new platforms on the market for any kind of suitable cases and segments, all four areas will be served and sorted by only oneplatform, for home playground, over SOHO market, entering the Pro area and ending up the enterprise zone. This could be a game changer for many of us. Who gives them a promise that only one copy will be used or if someone will be installing one system and copy it over and over and over only with one License code? What is then? Who cares then on their business and an their income or plain on their money?
For sure getting hands on a copy will be fine but how many users will then trade over P2P networks this registered copy then?Well, License fees are very small MikroTik's income part compared to RouterBoards sales So actually nothing will happen to the business =). Hi boysi have some doubts regarding the x86-64 architecture.but first things first, is it just me or there seems to have a bug on all versions of mikrotik x86 that cannot read RAM memory available correctly?
Or is it just some versions or Memory RAM modules that does not display it correctly?i installed mine 6.39.3 bugfix with 4GB ram hyperfury ddr3 1333mhz, and on memory only detected 1886mb memory.then i uploaded the all package 6.31 x86 to folder and hit downgrade on system packages and installed 6.31 version.went on system resources hardware and enabled multi-cpu and also hit tab x86-64 architecture. And my RAM memory reading on the mikrotik went up from 1886Mb to 3663MB total memory. Ok that seems a more realistic number of the memory insterted so i shutdown the PC, and plugged in another 4GB memory module with total 8GB ram, and rebooted the machine but still only detects 3663MB ram.upgraded it back up to 6.39.3 bugfix version. And readings on RAM now still 3663MB prior to the 1886MB intially.aldo on system resources print i now get x86-64 architecture showing on the CPu. Code: bat@cloudtik /system logging /sys reso printuptime: 3d5h38sversion: 6.41.1 (stable)build-time: Jan/30/2018 10:26:14free-memory: 22.0GiBtotal-memory: 23.4GiBcpu: Intel(R)cpu-count: 16cpu-frequency: 2266MHzcpu-load: 0%free-hdd-space: 3663.5MiBtotal-hdd-space: 3766.4MiBwrite-sect-since-reboot: 484656write-sect-total: 484656architecture-name: x8664board-name: x86platform: MikroTiknow i have some more, but they just get all stopped suddenly, and i don't know why this happens.
The VMs have 512MB-1024GB RAM each, and some negligible amount of hard disk image hosted on HDDs. So far i wasn't able to correlate this issue with any event, and i'm curious whether anyone may faced similar issues.
The VMs use virtIO networking, each of them has its dedicated virtual ethernet that is routed (e.g. Not part of any bridge).
There is a reason why RouterOS on x86 supports only 2GB - the speed of memory addressing. With high/low setups you would lose 5 to 10% of performance.The cost of PAE is way lower on modern x86 processors, since they have such large TLBs (as are optimised for much larger datasets with 4KB pages), and all the processors are highly-optimised for the larger pointer size used in PAE anyway. I suspect Mikrotik don't enable x86-64 kernel by default as they are using an older kernel with vendor drivers that have some build/integration issues, so PAE may be a good lifeline until RouterOS v7.So, just enable PAE for now. It's a big win for many cases, not only since the kernel can optimally size many network related hashtables, but also because there is less pressure (therefore page reclaimation) on various pools.Finally, going CHR means lower uptime, as there is more configuration, performance overhead and patching for security issues due to the VM host. For maximum single-stream throughput on critical infrastructure, I cannot wait to recommend native x86-64 RouterOS. Looking under the wireless package file it looks like the file structure is the same for 32 bit and 64 bit drivers as far as i can tell and theres even a 64 bit driver file so im confused as to why wireless is disabled for 64 bit router os.
I mean i can go back to 32 bit, but with ddr3 being so cheap secondhand theres really no reason not to use 8 gigs of ram. I use this in kind of a homelab setting to test changes before they ever get implemented so i know im not the typical user, it just seems strange to have 90% of the functionality already implemented and hidden behind a simple flag and not use it when even atom boards have been 64 bit for the last 8 years or so. It seems that krzysztof manually changed architecture-name: x86 to architecture-name: x8664 only in post.he also 'changed' free- and total-memorynow just look at 'Extra packages' links in Download page.
This.zip contains everything including 'system' package (which contains the kernel, I believe )X86 link:CHR link:now find the differenceso seems like 'x8664' is just a flag somewhere in the file systemby the way, I have old good RouterOS virtual machine which was installed when no CHR existed. Code: admin@TestPlace system resource printuptime: 1w1d2h51m41sversion: 6.39rc38 (testing)build-time: Feb/24/2017 08:46:35free-memory: 21.7MiBtotal-memory: 88.4MiBcpu: Intel(R)cpu-count: 1cpu-frequency: 2933MHzcpu-load: 0%free-hdd-space: 58.8MiBtotal-hdd-space: 99.2MiBwrite-sect-since-reboot: 394778write-sect-total: 394778architecture-name: x8664board-name: x86platform: MikroTikin some early v6 versions there was a checkbox under 'System - Resources - Hardware' for enabling 64-bit mode. Looks like they removed it and created the CHR as a separate systemHello,ROS X86 6.38.3 is supported 64 bit? Its not 'me or many' who wants not to use routerboards, it's the limitations of VM about 10Gbps, 40Gbps fiber modules, not needed usage overhead and the extra costs of HyVs. Why I need to use a VM if I can use a PC by my own to do the job? Why to depend in other OS to have the intended OS running?
Undesired OS updates ruins the uptime, and also aggregates the extra security measures to avoid supporting OS become vulnerable to attacks.We use routerboards, we love it, but in the Core we need the upgradeability and brute force from PC's.