Footy

Kleines Update (18.01.2016) : 
System Info-----------Processor : Intel(R) Xeon(R) CPU E5-2630 v3 @ 2.40GHzCPU Cores : 2Frequency : 2397.115 MHzMemory : 2048 MBSwap : 1024 MBUptime : 13 days, 2:33,OS : Ubuntu 15.04Arch : x86_64 (64 Bit)Kernel : 2.6.32-042stab108.8Hostname : 13214Speedtest (IPv4 only)---------------------Your public IPv4 is xxxLocation Provider SpeedCDN Cachefly 24.1MB/sAtlanta, GA, US Coloat 941KB/s Dallas, TX, US Softlayer 5.45MB/s Seattle, WA, US Softlayer 6.44MB/s San Jose, CA, US Softlayer 6.34MB/s Washington, DC, US Softlayer 10.4MB/s Tokyo, Japan Linode 3.80MB/s Singapore Softlayer 4.51MB/s Rotterdam, Netherlands id3.net 17.5MB/sHaarlem, Netherlands Leaseweb 25.8MB/s Disk Speed----------I/O (1st run) : 145 MB/sI/O (2nd run) : 86.1 MB/sI/O (3rd run) : 85.2 MB/sAverage I/O : 105.433 MB/s Update (21.01.2016) :
 
System Info-----------Processor       : Intel(R) Xeon(R) CPU E5-2630 v3 @ 2.40GHzCPU Cores       : 2Frequency       : 2397.115 MHzMemory          : 2048 MBSwap            : 1024 MBUptime          : 16 days, 17:16,OS              : Ubuntu 15.04Arch            : x86_64 (64 Bit)Kernel          : 2.6.32-042stab108.8Hostname        : 13214Speedtest (IPv4 only)---------------------Your public IPv4 is xxxLocation                Provider        SpeedCDN                     Cachefly        32.8MB/sAtlanta, GA, US         Coloat          890KB/sDallas, TX, US          Softlayer       8.73MB/sSeattle, WA, US         Softlayer       6.94MB/sSan Jose, CA, US        Softlayer       7.26MB/sWashington, DC, US      Softlayer       4.50MB/sTokyo, Japan            Linode          2.81MB/sSingapore               Softlayer       4.04MB/sRotterdam, Netherlands  id3.net         21.2MB/sHaarlem, Netherlands    Leaseweb        29.4MB/sDisk Speed----------I/O (1st run)   : 53.3 MB/sI/O (2nd run)   : 42.2 MB/sI/O (3rd run)   : 42.1 MB/sAverage I/O     : 45.8667 MB/s                    Second run:
 
  System Info-----------Processor       : Intel(R) Xeon(R) CPU E5-2630 v3 @ 2.40GHzCPU Cores       : 2Frequency       : 2397.115 MHzMemory          : 2048 MBSwap            : 1024 MBUptime          : 16 days, 17:27,OS              : Ubuntu 15.04Arch            : x86_64 (64 Bit)Kernel          : 2.6.32-042stab108.8Hostname        : 13214Speedtest (IPv4 only)---------------------Your public IPv4 is xxxLocation                Provider        SpeedCDN                     Cachefly        23.4MB/sAtlanta, GA, US         Coloat          1.17MB/sDallas, TX, US          Softlayer       8.48MB/sSeattle, WA, US         Softlayer       6.08MB/sSan Jose, CA, US        Softlayer       3.55MB/sWashington, DC, US      Softlayer       7.49MB/sTokyo, Japan            Linode          2.40MB/sSingapore               Softlayer       4.15MB/sRotterdam, Netherlands  id3.net         20.4MB/sHaarlem, Netherlands    Leaseweb        32.3MB/sDisk Speed----------I/O (1st run)   : 50.5 MB/sI/O (2nd run)   : 42.4 MB/sI/O (3rd run)   : 42.2 MB/sAverage I/O     : 45.0333 MB/s

LAN communications will not be done by the "router" portion of the U-Verse gateway. It will use the internal switch inside.
 
The benefit of linking both switches directly to the Router, is that you'll have less traffic going through a single 1Gbit connection. In the original diagram, all traffic from both switches will be going through a single cable from Switch to Router (Of course, this is assuming traffic will be going to the router in this scenario).
 
As I said, it might not make any difference, but there is a potential bottleneck there. In reality, it will make probably no difference in day to day usage.

The next big thing is going to be HDDs and Ram being combined...
http://newsroom.intel.com/community/intel_newsroom/blog/2015/07/28/intel-and-micron-produce-breakthrough-memory-technology

Virgin Media.....
 
Thats your problem.
 
SpeedTest.net is allowed through at highest priority
Most other things are subject to their bandwidth policies...

just force it
 
https://www.microsoft.com/en-us/software-download/windows10

Yup I got tri channel

I would say 8 is fine for most people
But I always over load RAM it's cheap and unused ram goes to buffers and cache to increase read performance on your HDD
Just keep in mind that windows automatically increases the size of you swap file based on your RAM amount assuming you haven't changed those settings
I got 24gb

Im not an expert or anything but...
When people overclock and hit stability issues generally they increase the voltages a little to see if its a power issue or not, if that works then they push the OC more while keeping an eye on temps.
So surely if you run a test to max your gpu and slowly reduce the voltages till it crashes then up them a little you should have your stable point...

Haha. Ohh I know about most of that already. I'm also not totally new, just trying to see what kind of jokes the community has regarding IT.

The i75960X only supports 64GB
 
http://ark.intel.com/products/82930?_ga=1.228092573.1085659344.1436309170

First off, why are you complaining? Nobody was telling you to run errands or do what you don't want to do. Make it clear first, the 'taking it to HA', is it what I want or what you want? Because I never said or implied anything even remotely close to 'taking it to HA'. Now if it's what you want, why are you complaining to me?
 
The plot is very simple, I said yes, you said no. I said I'm ready and willing to give the test subjects, testing methods, and procedures for you to recreate. You're free to make your choice. You want to challenge my test data or methods or results, sure, why not. I like to be challenged, as that's one of many ways to learn. If I'm proven wrong, then I'll say I'm wrong. You want/need to ask somebody else (in this case, people in HA), that's your privilege, not your liability. Hell even if you wanna take this all the way up to MIT, then go ahead. Just don't put it on me, because I never told you to do anything you don't want to.
 
The way the wording is, it sounds as if what you're saying is 'You're wrong. I don't know what or how, but you're still wrong. To know what or how, I need to ask more knowledgeable people, but I don't want to, so stop asking me to prove you're wrong and just accept that you're wrong'

If not taking into account the scale of differences, I actually agrees with him, and in audible spectrum of course. The (audible) differences are there, just in tiny scale. I'd say it's like wearing 100% cotton shirt vs 90% cotton + 10% polyester (or something) shirt. Is it different? Yes. Can you feel it? On and off for me, but maybe someone somewhere can. Is it worth the extra money? Not for me, dunno about others. 

I have noticed you are sighing a lot, perhaps you should go see a doctor, depression is a real and dangerous thing

I fully understand your point of view
However this is part of the provider acquisition process, I personally ask all providers for the minimum garunteed bandwidth
In this country UK, Virgin media quoted me 50:1 on a 153/12mb line and BT quoted me at 1:1 on a up to 76/19mb vdsl with expected speed due to distance of 65/18mb.
I went with BT because the minimum garuntee is more important that spike bandwidth
Some users want high max speed but don't care about stable throughput
You have to remember the bandwidth you buy for 5$/month is shared because it costs your provider more than that to buy peering bandwidth
This is why business lines cost so much more so the provider can garuntee bandwidth, stability and include SLAs
So yes you are right, it's annoying
And yes you should complain to your isp
However essentially I have to take the view that you get what you pay for....

Thanks,
I think Ill go take a look at Synology and QNAP.
Personally I come from the EMC, 3PAR storage world, used to work with netapps a few years back but stopped when I found the write cache bug.

Short answer No... Maybe... Virgin has a long way to go before it could be considered "fastest" they have a mediocre service provided over old and crumbling infrastructure which they are ill equipped to maintain or upgrade going forwards. Their network backbone is a joke (not even a good one) and their habit of snooping, throttling, threatening and rate limiting their customers is a serious put off.
 
The quality of routers depends entirely on which one they provide with their "service". They have a few different models they ship which are all pretty bad. So long as you buy something with cable compatibility from a respectable vendor you should be able to increase your local performance for a modest sum.
 
The fastest consumer/residential broadband provider in the UK is provided by Hyperoptic... it's a real fiber optic cable run directly to the property... and they're really quite awesome...
 

 
The only other large ISP preparing to bring these speeds to residential customers is Kcom... and that's only available if you live in Hull.

I don't think Virgin is the fastest in the UK (http://www.speedtest.net/my-result/4411744243) I pay for 1/1 gb and as you can see at peak time on friday evening I get "terrible" speeds =D
+ I have had massive problems with Virgin in the past where I was paying for their fastest service and couldn't get more than 1mb except to the virgin test server.

So if all your doing is acting as a relay/translation server then you can think a bit more simply.
 
If you say a cpu can handle 20 concurrent users (115mhz each) you can build a 1U server with 2x18 core cpus for about 10K USD (handling 720 users)
10000/720 = 13.8  (so 14 servers 140K USD)
 
Lets say each user uses 20KB/s (160kbit)
720 = 112.5 mbit
10000 = 1.56 gbit
So you need 2gb of external bandwidth and a switch that can handle 10GB uplinks and 1GB downlinks
lets say a 3048 - ~7K USD
 
Firewall (2GB spec)  5525-x  ~6-7k USD
Load balancer 2000s 5gb spec ~12k USD
Internet 2Gbit in a DC ~1k/month
 
Rack, power and cooling ~1k/month
so....
140+7+7+12 = 166k capex and 2k/month
allocate 10k for wires and other unexpected things like isp install fees

you often get things like this when managed switches and connected with unmanaged ones, or a typically fun one is ip phones.
generally you will want to enable a few things on the edge of your managed network;
BPDU Guard (often easy to integrate with portfast)
storm-control (use it to shut down ports if broadcast traffic exceeds your limits)
 
Another reason I see this happen is because of excessive switch chaining. spanning tree can only calculate loops on default settings over a network with a diameter of 7, so splitting up your L2 network with routers can help or even just connecting your switches in a few simple star configurations and linking the core of each star can help.

Depends on what else you have in your PC
295X seems to pull no more than 500W
i7 5960X  140W  (640)
RAM CMD64GX4M8A2400C14 190W  (830W)
HDD 0F23267 115W (945W)
 
sound card
network card
more hdd's
bluray drive
case fans
cpu fans
led lighting
 
there are many situations that an 850W PSU will drive a whole PC and a 295x2, however you should do your sums before you buy the product.
 
Don't just look at benchmark stats, look at the games you play and at the resolution you play them at. (whats the point of looking at 1440 or 4k if your running 1080 right?)
There are some really interesting issues with some games where the 295x2 can actually perform worse than a single 290x.
The 980 is really good with some games at some resolutions even beating the 295x2 in a few, but on average the 295x2 will beat a single 980 add a 2nd later and the story changes a lot, however with a 295x2 it is very difficult to add another, there is the additional radiator you have to mount and the 1500w PSU you need to get.
 
Make sure your decision is based on what is right for you! not on how loud fan boys from either camp shout.

need a L3 switch with ACL support

If I were you I would buy a 960, by the time you need to upgrade you will probably want a new mobo anyway (http://www.nvidia.com/object/nvlink.html)

The technology exists to have fast speeds everywhere, however its all about infrastructure.
Speeds in the UK vary from 2mbit to 1gbit for home users. 
In about 2005 BBB in Sweden released 1gbit home Internet however they didn't have enough peering bandwidth to cope, so anything off the BBB network was slow.
 
For you specifically in Indonesia its an infrastructure investment problem ( http://www.submarinecablemap.com/ ) however even when more bandwidth becomes available you need to drive competition in order to get it to the people.
If you have 1 ISP where is the pressure to provide more for less.
One of the big drivers for connectivity comes from datacentres which subsequently provides the scope to give better bandwidth to residents.
 
But essentially infrastructure needs $$$, so more $$$ means more mbits

Abe,
 
Glad it made some sort of sense.
I haven't used this particular mainboard before.
 
The obvious things to try would be;
 * Disable the console redirection.   
 * Change the cables over in the back of the server ONLY between NICs 1 & 2
 
There are 3 reasons for a connection being 100mb, 1 the server nic, 2 the cable, 3 the switch port
The cable change will be useful because;
  100mb needs 2 pairs (4 cables)
  1000mb needs 4 pairs (8 cables)
so doing the switch will take a working 1000mbit cable from a working switch port into your server, if the switch port still runs at 100mb then its a server side setting
If it runs at 1000mb its either a cable or switch port issue.
Do this and let me know the result ^^.
 
I don't know how console redirection works on this board.
However if you have 2 NICs showing to your OS (Windows/Linux) and they are connected to the same network (subnet) then it will only SEND traffic via one of them due to how routing works.
you may find if this is the case putting them on different subnets may help your cause. however it does add some complexity with routing if you wanted to connect to both the NICs from external sources.