Bigfoot Networks Killer 2100 reviewed, 'completely destroys' on-board NICs
By Tim Stevens 
posted May 13th 2010 9:18AM
posted May 13th 2010 9:18AM
We're still trying to wrap our heads around the idea that an aftermarket NIC can really improve your gaming experience, but time and time again Bigfoot Networks Killer line of network interface controllers comes through reviews with flying colors, and its latest, the Killer 2100, is no exception. Boot Daily took one for a spin and found that it "completely destroys the typical on-board LAN port in all the gaming tests." Latency and UDP throughput were clearly improved, and the new software was found to be comprehensive and useful. We're still not seeing this as a product for casual gamers, but if you've ever blamed lag for your inability to rack up the headshots then this is $129 well-spent -- assuming you'll be able to come up with some other excuse.
Bring on the pain!!!
@mad0maxx Their super-fast-loading server completely destroys to see the review itself this week. Must be KILLER NIC...
@mad0maxx Don't read that bullshit.
Summary:
they tested LAN. Somehow made 2 PCs in lan lag whole 6-10ms while KILLER had 0.1 ms lag.
The rest of results show no difference between KILLER NIC and a build-in solution for $0.
They did not:
run speedtest.net test with KILLER and build-in NIC
run pingtest.net test
no online MMO games
no online FPS games
no nothing.
Yet they say "its so cool I'm going to piss in the pants!!!" Yeah, because LAN no longer lags. Thats just soooooooo cool.
I wish I had extra $130 - I'd buy it just to make a proper review of it. Always wanted to since the thing came out few years ago but economy crisis got me faster than the card crossed the border of my country.
BIG_FART xD
@Shinigami internet connectivity has too many variables, a LAN test is the most accurate, controlled and stable way to show performance versus an on-board module. Start testing against the internet, you're going to hit limits and performance decreases when you start jumping through external devices, like cable or DSL modems and any external network infrastructure.
The problem with your statement is that you are expecting magical things to happen with your internet connection when in reality your internet ping and bandwidth results are ultimately decided by your provider hardware.
This is a network interface card, not a modem.
NO CARRIER
@mad0maxx
This is why this card is BS
if you have a GB lan at home, ping another computer. what do you get? Sub 1ms latency.
Enough said.
@lawrencealan
And that is exactly the problem with the Bigfoot NIC. Why plunk down $100+ for a NIC when 99% of your network problems are coming from the servers or the transit? The Bigfoot makes the 1% that you can control even faster!
Gamers may be their target market, but I have a hard time seeing it gain any traction at that price point, especially considering all the other hardware in a system vying for upgrade dollars, and in light of the first 2 versions of the Bigfoot making close to no difference whatsoever vs a $0 onboard NIC in real-world online gaming scenarios.
@cinnabuns "I have a hard time seeing it gain any traction at that price point"
That's what reviews like the one linked in this article are for.
@ALL
You all are fools! I never even read the article because the product is a complete waste of money! You all assumed I read the article and support the product because of my silly comment! haha ;-)
@Shinigami
One small correction, the first benchmark was a latency *variation* test, which means the Killer could have still produced higher pings but little variation between them.
Not very helpful test!
This card is really just for rich kids and nut jobs.
@Shinigami
My guess is that the drivers for the NIC are designed to report a "different" latency to the system than the true latency.
If two computers are connected in a LAN and you're testing pings by sending requests to the other computer, then the ping will always be as slow as the slowest computer, it doesn't matter that one computer has a different NIC installed. This test is entirely falacious. The fact that they report a "lower" latency for the computer with this hardware NIC installed on a two computer LAN system just goes to show the card is misreporting latency information. The manufacturer's information even says its drivers 'replace the Windows LAN stack' or something to that effect.
Everyone is making interesting points.Why so harsh & skeptic..?
The facts are in the benchmarks and testing. You can clearly see the performance results of this gaming network card improving latency in the GaNE tool and boost in speed in the Netperf & file transfer tests, etc
These tests show the gaming network card performing. Its in the undisputed #s. Overall, if your a competative gamer would you not simply want the fastest network card on the planet regardless? Even if its just a few ms differece. Once that first game packet leaves your wire from the PC its that much more ahead of the curve on all the rest of the competing packets along the way...
To get the fastest NIC that can help with your latency/speed performance plus the software in which helps you completely manage your whole network with visual controls for $100 seems reasonable especially in light of the $300-$400 graphic cards. What have you got to lose? We all know you can buy this card easily at an etailer and if your not happy return it. If you dont believe the benchmarks and testing showing this card performs, just go try it for yourself. Ultimately its the experience you feel in the game (smooth & faster game play) that matters anyways ... I see pro gamers using this product..
@tech3399: Nice, you created an account just to make this one post. No way you could be a BigFoot employee or anything.
@Shinigami
This Maximum PC review of last year's card give a better idea how this tech works? http://www.maximumpc.com/article/reviews/evga_killer_xeno_pro
@tech3399
But the numbers ARE disputed, because the methodology of the tests are unclear and potentially flawed. The above sourced review states they created a local network and transfered files between two computers, presumably one with the external NIC and one with the onboard NIC. Regardless of which computer initiated the transfer, it should still be as slow as the slowest computer and the ping should be as slow as the slowest computer. Perhaps the review article needs to include more information on the methods they used, but if they connected two computers, one using on onboard NIC and one an external hardware NIC, If the computer with the hardware NIC displayed a significantly lower latency, it's because its driver is intentionally deceiving the operating system.
If the test was using two computers both with the external NIC, then it's incredibly deceiving, because in what situations will you have two people with desktop computers both with the same external NIC hooked up together, unless you buy TWO $130 (!!) cards to make yourself a two player LAN party at home. If anyone else joins you LAN, the fact that you have a faster connection is meaningless and this is becomes an entirely moot point. When you're playing games over the internet, or over a LAN with one slow computer, it doesn't matter if your ping is ZERO, if you've got a player jittering around with a ping of 300, it's impossible to predict their location. Having a lower ping does nothing to improve your own performance, it only decreases the difference of your own perceived location and your actual location for OTHER gamers.
In fact, who would even bother with this when you could just play Quake Live Instagib, for instance, and have the server effectively make everyone have a ping of zero.
@tech3399
Also, according to the review website, the hardware NIC manages an INCREDIBLE 900+ MB/sec over a gigabit ethernet router, which is only capable of 150-200 MB/sec ABSOLUTE MAXIMUM.
Either this website has no journalistic credibility, or the hardware NIC is producing false numbers to the benchmarking software. Either way, this should not encourage people to blow $130.
Here's an idea, $130 will give a starving kid in Somalia food for a several years. That sounds like a better deal to me.
$129 for NIC is absolutly insane.
@ChillyCat
I agree. This is like monster cables, how can a dedicated NIC improve latency ffs?
@roflcopter
"how can a dedicated NIC improve latency" ??
It can't !!
@ChillyCat
Indeed, especially seeing that most latency can't be blamed on your adapter, its much more likely to be your broadband connection latency, not to mention the number of hops to and from the game server and general internet traffic lag caused by other peoples traffic.
@roflcopter you should read up on the card and then you'll understand how. It's just like buying a dedicated video card for gaming. The computer itself no longer handles traffic. The card itself will handle the network traffic.
@roflcopter
Your OS creates some of the latency. Not, much, just a few milliseconds.
Really, the main benefit from these cards is FPS improvements in games with high network traffic like MMOs in crowded cities in raids. When I got my M1 years back, I did testing in WoW-IF and the Vanguard Elf City, and saw about a 3-5% framerate increase on average (some 'intense' testing around the IF bank-auction area saw up to 8%), and fewer/less severe framerate spikes.
As far as latency improvements, I never really noticed much. About all I ever see is when my brother comes over and plays the same games, my ping is typically a few ms lower...nothing I would consider worthwhile.
In any case, it's definitely not the first (or second, third, or fourth) thing to spend your money on when building a gaming PC, but you can see some respectable improvements in the right situations. Not some magic miracle card, but far from the snake oil some claim it to be.
@ChillyCat
But the amount of processing your CPU does to interpret the data from the onboard NIC is inconsequential, far less than 1% of your CPU cycles. Especially considering you generally have about 20 other servers or devices your connection hops through. 1% of 5%, at most conservative scenarios, is not significant. Say you have an average cable internet connection, this back-of-the-napkin calculation indicates that around 0.025 ms is lost due to your onboard NIC.
If you're a "professional" gamer (*chortle*) and play LAN parties for money and have sponsors (*chortle*) this might be something you'd have in your setup, just like a football player might put a pink rabbit's foot in their helmet to improve their performance on the field.
@techi25
I'm lowering your status to techi19. You lose 6 points for even
suggesting that reading the Mfg's literature will explain how it
works.
"Paper don't refuse ink" .....I can put anthing I want on paper. fact or fiction
How does one find out if their connection sucks due to onboard network cards, or because their service provider sucks? And are the benefits of something like this card destroyed if you're also using a router that didn't cost $1k?
@Grither
It all depends on your level of cowbell
@djessemoody +1000000000
@Grither If you're using a router make sure you're computer with the killer card is set to the DMZ. It will have less restrictions to the outside world.
Would it not be more appropriate to say that it "kills" on-board NICs?
@xdreamer it would not.
Meh. Might get this for my home server. But that price. For a NIC? Way too high. Maybe $80 or so I would be willing to pay. $130? GTFO.
I've also never understood how these add-ons can perform so much better than the onboard NICs. I have to assume the southbridge chips, since they don't just handle network I/O, give the onboard NICs a much lower priority than they deserve, and probably have to be more power conservative as well. Which makes me wonder why we don't hear much about improved latency from all the motherboards out there designed especially for gamers. If this standalone product goes for $129, having that kind of performance onboard shouldn't add a significant amount to such a board, which normally runs in the $300 range anyway.
Fail on so many levels.... The sad thing is that a LOT of people will buy this snake oil thinking it's going to help. New flash folks, your ISP and the interwebs are still going to determine 99% of your latency/packet loss/jitter.
@ccampb6
I'd agree, and I don't know why, but for some reason that review really doesn't sit well with me.
- At the beginning when they are describing the changes from the old model, they say that the upgrade "really ratchets up performance across the board over the competition". How would they know? They haven't even started testing it yet!
- They use a "Game network efficiency" test to show better performance. A quick Google suggests that nobody else has ever used this test. Was this test invented just for this card?
- They completely fail to describe the other test that shows better performance: a UDP throughput test. How was it done? Why was it done? What does this all mean?!
I think it's a bit odd too that Boot Daily is the only place that has got hold of one for review instead of one of the bigger tech sites like Anandtech or Bit-Tech and yet aren't gloating about an exclusive.
Definitely not the best review I've seen, I'd want a second and third opinion before believing it.
@Art
I agree, the "review" is definitely written like a promotional piece. They didn't make any attempt to test it in real-world gaming scenarios, nor did they compare the *actual* ping times in games with the two card solutions. My feeling is that your local PC contributes so little to your latency that even if you were able to entirely eliminate it, your results would still fall barely outside the margin of error.
@Art
Anandtech is reviewing one next week.
This is probably the most ridiculous thing I've heard in a long time. My concentration while getting my CS degrees was in networking, and I can honestly tell you that the majority of contributing factors to your "latency" or round trip time of packets DOES NOT come from your computer, but from outside issues like queuing delays from routers, propagation delays that you simply cannot bypass, and the transmission delay on your link to the ISP.
How do you know this? Ping the loopback on your computer. It will be less than 1 ms. Ping a game server or google, and it will probably be between 30-100 ms depending on your ISP. The ONLY portion of this latency that the killer NIC can affect is the part internal to your system, NOT the part attributed with your router, or ISP. If that portion is roughly 1 ms, and your router and ISP are attributed for roughly 50 ms, you are talking about affecting 2% at MOST of your latency with any speedup. Even if it doubles the rate at which your machine is able to create and transmit packets (which is unlikely), the speedup will be almost unnoticeable, if not immeasurable.
This is a perfect example of a product which has no real benefit at all.
@defendertx
you can bypass to a certain degree by tunneling to/via a server closer to your destination
@defendertx
Loopback is not a useful measure of on-desktop latency. Please see:
http://en.wikipedia.org/wiki/Loopback#Virtual_network_interface
You have to use the $100 gold plated ethernet cable to get the most from this card.
Or even better this:
http://geekstersparadise.com/wp-content/uploads/2008/06/denon_ethernet_cable.jpg
@bjb
Yeah, but if your source isn't using those cables either, your line degradation is still gonna be a factor.
Just cause you put chrome on your exhaust, doesn't mean your emissions are going to be magically transformed (but they might sound different :D )
@xler8r You thought he was being serious?
@xler8r Can I transform my emissions into gold unicorns?
Awesome gamer kid´s rip-off ... I really like this! Great approach. I´m waiting for the platium/gold plated "low ping" network cables. Bigfoot Networks can learn a lot from the esoteric high-end hifi business.
MPP
http://paepcke.de
Well, in my opinion the test is not that good. They testet a cheap realtek onboard chip against a 129€ card ... at least some GOOD onboard intel chip or some other (maybe not onboard) good card should have been testet aswell
thats something like whats faster, a porsche or a old beetle ...
Maybe Bigfoot Networks Killer 2100 should install this device in their webserver....it's SLOOOOOWWWWWWWW.
Why am I not really impressed...
Plus I wonder how it holds up against my Intel pro/1000 PT
@airbag888
Exactly. The Intel NICs have been around for years -- they've been tested and are rock solid...I have a dual port PRO\1000 MT and it's the shiznit...it has two gige ports, and costs less than this.
@DoctarPeppar
That's exactly what I have. salvaged from a dead server so I can't say I made an educated purchase. There's a lot of knobs to turn on this one, I'd be glad to find a walkthrough of the settings to be used somewhere to optimise for gaming and which driver to use (MS 7 default or Intel's X64 one)
id rather buy a couple of beer for $129. fuck latency.