Bigfoot Announces New Xeno NIC

Status
Not open for further replies.

hellwig

Distinguished
May 29, 2008
1,743
0
19,860
26
Don't waste $170 or even $130 buying a NIC that supposedly reduces your latency and increases transfers. These claims are unsupported, although it is good to see the price much lower than the $499 they wanted for the Killer NIC.

Instead, send me only $19.95 (+ 9.95 shipping and handling), and I'll send you my revolutionary new NIC Speed Enhancer. Simply attach the product (looks like a spondgebob sticker, but trust me, there's a lot more than meets the eye) to any IC on any existing NIC product from any company. I guarantee you'll be seeing reduced latency and increased transfers or your money back (minus handling and restocking fees, also, sticker must be removed and returned undamaged, if sticker was scratch-and-sniff, fragrance must still be detectable to qualify for refund). Don't wait, act now, quantities are limited.
 

MoUsE-WiZ

Distinguished
Oct 4, 2008
59
0
18,630
0
Or just get a torrent client that doesn't suck? uTorrent is smart enough with its bandwidth/memory demands that it doesn't effect my latency at all =/
 

lire210

Distinguished
Mar 23, 2009
39
0
18,530
0
wow what a useless product and why would I want to waste a pci-ex1 slot on this my pci was ok but what if I want a sound card cmon xfi or this hmmmmmmmmmmmmmm.
 

H8ff0000

Distinguished
Jun 18, 2008
27
0
18,530
0
I bought a Killer NIC K1 awhile back. Bad idea on my part. It sounded like a cool idea, especially for myself since I torrent like crazy, but the product simply doesn't meet what they state. Look on their forums and you will understand what I'm saying. Lots of reports of problems installing, using, etc. Their drivers are problematic, and much worse, they need bios flashes, that WAY TOO OFTEN go bad. The software that goes along with the card causes a lot of headaches as well. Their support people are nice, which helps, but it didn't stop the card from becoming mere internal case-bling for many users. I'm not saying to NOT buy their products, just make sure to do the research first.
 

curnel_D

Distinguished
Jun 5, 2007
741
0
18,990
1
[citation][nom]MoUsE-WiZ[/nom]Or just get a torrent client that doesn't suck? uTorrent is smart enough with its bandwidth/memory demands that it doesn't effect my latency at all =/[/citation]
That's not true at all, and you know it.
 

dansergiu

Distinguished
Mar 26, 2009
2
0
18,510
0
I don't get it. Either this article is lacking crucial information or it simply states false information. How exactly am I going to reduce lagging using this card while BitTorrents are still running. Every BitTorrent client tries to make as many TCP connections as possible for maximum bandwidth. That's a fact. This card has a separate processing unit and can handle audio. So? does it make sure that I get a dedicated bandwidth for my application (read game)? I don't think so. All it does is take some processing out of the CPU but it also installs a driver in the OS, interrupts and synchronization with the CPU and Northbridge. So you actually reliving less processing from the CPU than you may think. But the BitTorrent client still tries to take as much bandwidth as possible. Nothing stops him from doing that. This card is completely garbage. Don't buy it. This is false marketing and Bigfoot tries to sell it's stupid cards at very high prices just so people would think they are actually worth anything.
 

blackbeastofaaaaagh

Distinguished
Jan 22, 2009
78
0
18,630
0
For such a card to make any sense it has to be able to perform on-board traffic shapping. It should also assign higher priority to outgoing ACK packets over regular TCP packets.
A couple of gateway/router claim to do this (haven't tested them myself). My net-surfing used to slow to a crawl when I had heavy-duty p2p sharing going on. Not anymore. I purchased a software-based traffic shaper and have been very happy with it.

Check out the link. It has a nice crash course tutorial.
http://www.cfos.de/traffic_shaping/traffic_shaping_e.htm
 

bigfootsean

Distinguished
Mar 31, 2009
2
0
18,510
0
[citation][nom]dansergiu[/nom]I don't get it. Either this article is lacking crucial information or it simply states false information. How exactly am I going to reduce lagging using this card while BitTorrents are still running. Every BitTorrent client tries to make as many TCP connections as possible for maximum bandwidth. That's a fact. This card has a separate processing unit and can handle audio. So? does it make sure that I get a dedicated bandwidth for my application (read game)? I don't think so. All it does is take some processing out of the CPU but it also installs a driver in the OS, interrupts and synchronization with the CPU and Northbridge. So you actually reliving less processing from the CPU than you may think. But the BitTorrent client still tries to take as much bandwidth as possible. Nothing stops him from doing that. This card is completely garbage. Don't buy it. This is false marketing and Bigfoot tries to sell it's stupid cards at very high prices just so people would think they are actually worth anything.[/citation]

Dansergui, this is actually the most favorite post I've had to answer today - we're catching up on a week's worth of news here after a very successful GDC. But I digress - you posted the most interesting question of the last week!

They way we do this is straightforward. (I almost said simple, but the engineers would kill me.) By assigning each process a separate priority in hardware bandwidth control, you can ensure that your game (or your voice chat, or your streaming music) gets the highest priority in your network.

Our founder and CTO uses this to literally strangle Windows Update, iTunes, Adobe updater, pretty much any non-gaming process, while prioritizing whatever voice package he's using and giving utmost priority to a game.

Last month I used this feature on the Killer NIC (our first-gen card) to download Dawn of War II via Steam while playing about 5 uninterrupted, awesome, lag-free hours of Day of Defeat. I assigned Priority 1 to hl2.exe and lowest priority to both Steam.exe and svchost.exe. I also leave my gaming machine in the DMZ on my router and use the hardware firewall in the card, but that's personal preference.

If I was a totally hardcore h4xx0r, I'd find a way to make Steam downloads go to the USB storage device that you can attach to the back of the card. As it is, that's where I put FTP and BitTorrent downloads. This way, the download never even touches the system bus, let alone my hard drive.

So there are two answers to this - yes, you can leave your torrent client hungry, and let him try to open as many connections as he likes. However, by assigning him a lower priority than your game, you can ensure that your system plays your game first and downloads Ubuntu second.
Additionally, you can allocate bandwidth in our bandwidth control application, so that your torrent client only eats up a chunk of the available bandwidth.

Myself, I feel like if I'm paying for the 10Mbps down, I might as well use the whole 10Mbps - especially if I'm only using 128Kb down/up for the game that I'm playing, why not use the other stuff alongside it?

Thanks again for the question, and make sure to check out the website in the upcoming weeks for more information.
 

dansergiu

Distinguished
Mar 26, 2009
2
0
18,510
0
Dear bigfootsean,

First of all thank you for your answer and please accept my apologies for the flame post there. I got a little carried away. But you have to admit that the article is misleading.

I now understand what your card does and more important the role of the card's driver. But mostly, it's innovation. And I am also an engineer and I know that achieving this type of separation is no easy task :)

I think that Tom's Hardware should do a more technical article about this cards, some benchmarking and actually measure CPU load with and without the card in similar download conditions ( it's not hard if you think about it. All you need are 2 PC's - one server and one client the NIC on the client and simply download a 30GB file while running some 3d benchmarks ). Also a similar test can be set up to measure laging in online games (only that you need to use 3 PC's and a server that would simulate an official gaming server). But I don't think is too hard to achieve. This would be, in my opinion, a lot more interesting to read and would actually provide the reader with vital information in his decision of buying the card.

I must say your answer made look more into this cards because until now I felt that they were actually only marketing and nothing really innovative. So thank you again :)
 

bigfootsean

Distinguished
Mar 31, 2009
2
0
18,510
0
[citation][nom]dansergiu[/nom]I think that Tom's Hardware should do a more technical article about this cards, some benchmarking and actually measure CPU load with and without the card in similar download conditions ( it's not hard if you think about it. All you need are 2 PC's - one server and one client the NIC on the client and simply download a 30GB file while running some 3d benchmarks ).[/citation]


Well, Tom's Guide did a great review of the Killer M1 NIC (Our first-gen PCI cards) last year:
http://www.tomsguide.com/us/killer-m1-nic,review-1083.html

As for the download test, we actually improve latency, while the Windows Network Stack is best tuned for file transfers. A better, more appropriate test would be to download a file while gaming, or download a file to the USB device while gauging CPU, system bus and HDD load. While gaming even. :)

And the best gaming tests are real-world tests. Two computers and a third server as the game server are fine for lab purposes when we test functionality, but it doesn't properly portray what a person feels or sees when they're playing from home. Two computers side by side connected to a crowded, offsite server in a busy game is a better test.
For our performance testing, and what we recommend to reviewers, we get two identical computers (1 Xeno, 1 no-Xeno) and connect them to an external server. CS:S is a good start, but we test a lot of games. Two testers play, while a third tester has a stopwatch and a score sheet. Every minute or so, the scorer says "time" and the testers read off their framerate and ping. Repeat as necessary.

Since gaming and networking are essentially random, noisy exercises, we gather a LOT of data. This could be 30 data points over 60 minutes or so, in order to smooth out the occurrence of random events (you could be dead, you could be looking at a wall) and get good results that will feed back into real world predictability.

Thanks again!

Sean
 
Status
Not open for further replies.

ASK THE COMMUNITY