Beta
×

Welcome to the Slashdot Beta site -- learn more here. Use the link in the footer or click here to return to the Classic version of Slashdot.

Thank you!

Before you choose to head back to the Classic look of the site, we'd appreciate it if you share your thoughts on the Beta; your feedback is what drives our ongoing development.

Beta is different and we value you taking the time to try it out. Please take a look at the changes we've made in Beta and  learn more about it. Thanks for reading, and for making the site better!

'Killer' Network Card Actually Reduces Latency

Zonk posted more than 7 years ago | from the i'll-be-a-monkey's-uncle dept.

292

fatduck writes "HardOCP has published a review of the KillerNIC network card from Bigfoot Networks. The piece examines benchmarks of the product in online gaming and a number of user experiences. The product features a 'Network Processing Unit' or NPU, among other acronyms, which promise to drastically reduce latency in online games. Too good to be true? The card also sports a hefty price tag of $250." From the article: "The Killer NIC does exactly what it is advertised to do. It will lower your pings and very likely give you marginally better framerates in real world gaming scenarios. The Killer NIC is not for everyone as it is extremely expensive in this day and age of "free" onboard NICs. There are very likely other upgrades you can make to your computer for the same investment that will give you more in return. Some gamers will see a benefit while others do not. Hardcore deathmatchers are likely to feel the Killer NIC advantages while the middle-of-the road player will not be fine tuned enough to benefit from the experience. Certainly though, the hardcore online gamer is exactly who this product is targeted at."

cancel ×

292 comments

Sorry! There are no comments related to the filter you selected.

I waited long enough (1, Funny)

Anonymous Coward | more than 7 years ago | (#17178300)

First Post

Re:I waited long enough (3, Funny)

tempest69 (572798) | more than 7 years ago | (#17178786)

Strange, that FP is actually ontopic.... AC's and their bigfoot cards.

correct me if I'm wrong... (3, Interesting)

zappepcs (820751) | more than 7 years ago | (#17178310)

But the only real concern in making a killer NIC is keeping all the processing off of the CPU and bus. If the CPU/MB can shuffle packets at and from the NIC at the speed of the data bus, then it can't get much faster unless you want to offload protocols to the NIC etc.

A killer NIC? LOL what a phrase... Aren't there several of these Nicolas guys in jail already? right next to the killer Bobs and killer Joes.... sheesh

Re:correct me if I'm wrong... (1)

fatduck (961824) | more than 7 years ago | (#17178374)

I can't see how this is going to help the average gamer that the company is clearly targeting:

"Whether you have an hour to play at night or you are a dedicated pro gamer: the Killer NIC is for anyone that cherishes their online gaming time."
I'm a little skeptical about this proprietary Lag and Latency Reduction (LLR(TM)) Technology. Lag and Latency reduction seems a little redundant?

Re:correct me if I'm wrong... (2, Insightful)

porkThreeWays (895269) | more than 7 years ago | (#17178964)

It sounds like bullshit, but it actually works. The guys who made this aren't schmucks, but experienced designers. I know 2 people who bought one and they both confirmed slightly lower latency in games. Look at it this way, a 500 dollar pair of running shoes really isn't going to help the average person much compared to a 50 dollar pair. However, a professional runner is going to benefit. This card is designed for the professional gamer (oddly enough, they do exist today).

Re:correct me if I'm wrong... (4, Insightful)

BJH (11355) | more than 7 years ago | (#17179382)

This is an "emperor with no clothes" thing - if you can't tell the difference, you must not be an experienced gamer. Since I'm an experienced gamer, I can tell the difference. HORSE PUCKY, boy!

Re:correct me if I'm wrong... (5, Funny)

KillerBob (217953) | more than 7 years ago | (#17178562)

25 years and counting, they still haven't caught me. :-)

Re:correct me if I'm wrong... (2, Funny)

slashbob22 (918040) | more than 7 years ago | (#17178646)

Call me Mint Jelly; cause I'm still on the lamb. We Bob's gotta stick together.

Re:correct me if I'm wrong... (3, Funny)

Stormwatch (703920) | more than 7 years ago | (#17178728)

We Bob's gotta stick together.
FROG BLAST THE VENT CORE!

Re:correct me if I'm wrong... (2, Funny)

slashbob22 (918040) | more than 7 years ago | (#17178804)

I'm on the Blue Team and "They're Everywhere!"

Re:correct me if I'm wrong... (4, Informative)

jd (1658) | more than 7 years ago | (#17179080)

The "traditional" network stack shuffles data from the application from memory into the CPU then into kernel memory for running through the protocol stack then into the CPU then into kernel memory for running into the actual NIC driver then into the CPU then onto the bus. This assumes that the data doesn't fall out of cache during any of the processing of the packet.


A better design would be to have networked data in well-defined regions which the card can DMA directly out of and into. The PCI bus can handle a 4K transfer as an atomic operation on a single channel, so a 4 channel PCI card can simultaneously send and receive streams of jumbo packets without requiring any CPU intervention. The driver would merely need to be passed two lists of physical page pointers - one for inputs, one for outputs - for each of the open connections and to pass back signals from the board that an entire packet and/or message had been uploaded for a given connection.


(Jumbo packets can go up to 8K, and you can do 8K over two PCI channels, so that's a good unit to be working with.)


The next improvement that can be made to such a system is to improve the buffering. The network will be slower than the computer, almost always, so being able to queue up multiple packets for sending is a Good Thing. Filtering out packets that have not been sent but are no longer worth sending can be done entirely in parallel and does not require anything extra at the end of the pipeline. Not all packets are of equal value, though. You want to deliver the information in the order that will give the best possible benefit - which may not be the order in which the program generates the traffic. Hierarchical Fair Service Curve, Class Based Queueing, and a bunch of other similar techniques, have been developed to fix exactly that sort of problem.


If the ISPs would stop being so bloody stupid, you could also enable protocols such as your basic multicasting (for your UDP stuff) and Scalable Reliable Multicast (for the stuff that needs to reliably get through). That hacks, slashes, butchers and roasts (with just a hint of parsley) problems associated with sending identical state information to multiple end-points.


Bear in mind that Myrinet, Dolphinics, and a bunch of other vendors, use essentially the above mechanisms already and are achieving latencies in the region of 2.5 - 3 microseconds. I say essentially, because I'm not convinced they've optimized quite to the degree I'm suggesting - reliable, scalable multicast RDMA isn't something you'll see a lot of even at a supercomputer fair. True, you're not getting that kind of latency over the Internet whatever you do, but if you can achieve a hard real-time guarantee of 3 microsecond delivery in a LAN party, you WILL notice a difference. At the very least, in the door price, which will now be expressed in exponential notation to fit on the door.

How ... (1)

lintux (125434) | more than 7 years ago | (#17178322)

How can a NIC decrease the latency in any noticable way? Especially when playing over the Internet? Does it process ICMP echo reply packets in hardware so that the ping values will just look more l33t? :-P

Re:How ... (4, Insightful)

Arker (91948) | more than 7 years ago | (#17178390)

I suspect it will produce a noticeable improvement in situations where your computer is running heavily loaded. If you're playing a game that keeps your cpu pegged or near most of the time, your latency will be noticeably higher because of that (using the typical network card, which is a bit 'winmodemish' in that it's relying on the cpu to do much of its work.) So having a card that does all the network processing itself, without relying on the CPU, would avoid that slowdown.

Re:How ... (1, Interesting)

Anonymous Coward | more than 7 years ago | (#17178552)

Most network chips these days have checksum and TCP layer offloading.

This article is pure BS. If anything, this card probably increases the latency because of the additional layer of software involved on the card itself.

Re:How ... (1)

Schraegstrichpunkt (931443) | more than 7 years ago | (#17178750)

Most network chips these days have checksum and TCP layer offloading.

Can you list a few examples, preferably with datasheets? I'm not aware of most consumer-level chips being much different than the ones we've bee using for the last 5-10 years.

Re:How ... (1)

JebusIsLord (566856) | more than 7 years ago | (#17178830)

The nforce chipset does it... can't be bothered to dig up data sheets though.

Re:How ... (1)

profplump (309017) | more than 7 years ago | (#17179044)

TCP offloading in consumer chips, not so much. Mostly becuase the drivers are harder to write and while throughput goes up latency doesn't go down (at least not much), and consumer-level users are rarely NIC-bandwidth limited.

But unless you've got like a 3c509 or something, it's got on-board checksumming. It's been an extremely common feature even in cheap cards for several years at least -- even things like the RealTek 8139C+ have buffer-based transfers and on-board checksumming.

Re:How ... (2, Interesting)

jpop32 (596022) | more than 7 years ago | (#17179052)

Can you list a few examples, preferably with datasheets?

Here are two datapoints. A $10 PCI NIC, and a $100 mobo I bought lately (with an integrated NIC) feature checksum offloading. They are both GBit, so I guess you get that for free on any GBit NIC nowadays.

Other than that, I really don't see how a NIC can decrease latencies. The latency of that first hop off your computer is below 1ms anyways.

Re:How ... (1)

Kiaser Wilhelm II (902309) | more than 7 years ago | (#17179064)

Nvidia, 3Com, Intel, Broadcom, Realtek to name a few. Consumer grade parts, too. They all do offloading. The works. Perhaps they are missing something that a good server grade chip has, but probably not much.

Its also hard to believe that, given the load that you'd be putting thru them, that you would even detect a difference. With multi-ghz CPUs out, the benefits of offloading are slim to none and not noticed (except, maybe on server grade equipment that pushes out multi gigabits constantly, off many interfaces)

I don't have the datasheets. Do your own research.

Its been like this for years. The only time I haven't seen offloading is when I bought those cheap NE2000 compatible ethernet cards.

Probably not... (2, Interesting)

eklitzke (873155) | more than 7 years ago | (#17178842)

It's not like a computer sends you some data and the network card is immediately able to reply. To formulate a response, it probably needs data from the CPU, e.g. about your position, your health, or whatever it is that you need to transfer back and forth in a game.

An ICMP echo reply is totally different though. Unless you have a weird firewall setup going on, it's pretty much just safe to send out the echo response as soon as you get the echo request. So in this situation, you could peg the main CPU and then have the NIC doing the mind numbingly boring task of sending out echo responses without going through the CPU, and in this case you might see a latency improvement of a few milliseconds. But in general the CPU is going to have to do some processing and formulate the correct response anyway, so having a "smart" network card doesn't help.

No different from any other decent server NIC (3, Insightful)

Colin Smith (2679) | more than 7 years ago | (#17178926)

Server systems have been using high quality NICs which offload network processing for years. Decades even... I think $250 is a bit steep. But then I'm not a l33t gamer. Kudos to them if they can get people to pay $250 for a $50 server NIC. I call that good marketing.

Of course they also need to be running 15,000rpm SCSI drives on a decent SCSI HBA as well as a top of the line CPU and loads of RAM and top of the range graphics card.
 

Re:No different from any other decent server NIC (1)

Arker (91948) | more than 7 years ago | (#17179012)

Oh, absolutely. I have a hard time believing this will perform better than currently existing *good* NICs, so the price may be excessive. But I have no problem believing it will outperform the POS NICs that are most common, even in fairly high-end gaming rigs.

Re:No different from any other decent server NIC (4, Informative)

Arker (91948) | more than 7 years ago | (#17179238)

Aha, found a better source with some real info on this thing. Here. [anandtech.com]

Re:How ... (1)

faragon (789704) | more than 7 years ago | (#17179102)

For these cases there is an alternate, cheap, and usually effective, solution: increase the thread priority of the thread doing TCP 'recv/read'. You're welcome.

P.S. please, be careful when playing with thread priority throttling, it is better keep priority untouched in case you have chances of achieving thread priority inversion (!).

Re:How ... (1)

scdeimos (632778) | more than 7 years ago | (#17179234)

So having a card that does all the network processing itself, without relying on the CPU, would avoid that slowdown.

Last time I checked, network packets get built in main memory by your Berkley/Winsock stack and the final packets with headers and checksums get transferred to your NIC by DMA for transmission down the wire. In a modern gaming machine we're talking a (typically) 800MHz/128-bit memory bus, versues a 33MHz/32-bit PCI bus. How does having a Killer NIC improve this situation?

All I can think of is that Killer NIC is tweaking the Priority and QoS attributes of packets to help your upstream packets reach the server more quickly, but it's not going to do anything for your downstream traffic (originating from a server that's not going to tweak Priority and QoS) which is going to be far larger in terms of bytecount. The only traffic being sent from you is your own position/action updates, but the server is sending you the state of everything in the zone you need to know about (all other players, active object states, etc).

Re:How ... (1)

maxwell demon (590494) | more than 7 years ago | (#17178452)

It has a built-in time machine which sends the packets a few milliseconds to the past.

Re:How ... (1)

CommunistHamster (949406) | more than 7 years ago | (#17178510)

The next version further refines the technology to give you a negative ping

Re:How ... (1)

SuperBanana (662181) | more than 7 years ago | (#17178506)

How can a NIC decrease the latency in any noticable way?

It can't. It is most likely doing some traffic shaping tricks to make it look like ping times are lower.

If you want lower latency and you're using a custom firewall- make sure polling is off on the cards so packet processing is interrupt driven, and check what timer value the kernel is using. Turn off polling on the gaming system as well, but increased interrupts may hurt performance in-game.

This will, however, result in so little improvement it will most likely be unnoticeable.

Re:How ... (5, Insightful)

TodMinuit (1026042) | more than 7 years ago | (#17178576)

How can a NIC decrease the latency in any noticable way?

You'd be surprised what marketers can do.

Re:How ... (0, Troll)

arivanov (12034) | more than 7 years ago | (#17178722)

It is the in-game "ping", not ICMP. In most games this includes the RTT for traversing the game own network API all the way to the event loop, through the event loop and back.

1. Anything that can interface at a higher API level compared to the normal Winhoze network stack it is likely to reduce application level round trip.

2. The results will be more pronounced on a more loaded system.

3. If you read till the end of the article the NIC is actually an embedded Linux system so you can turn off your winhoze firewall completely and let Linux do the firewalling.

Overall: 3 alone on a heavily loaded system will give seriously improved application response rates and not just for gaming. 2 and 3 can boost that even further and allow applications that are not feasible under windows. F.E. winhoze sucks rotten eggz as far as Layer 3 based HA, anicast and routing and this will allow you to bolt on that on top of a Winhoze application by running a routing protocol stack on the embedded engine. In fact there is a lot of non-gaming potential in this thing.

Re:How ... (1)

lintux (125434) | more than 7 years ago | (#17179020)

3. If you read till the end of the article the NIC is actually an embedded Linux system so you can turn off your winhoze firewall completely and let Linux do the firewalling.

Wow... I'd think that such a thing would actually add extra overhead. It has to traverse an additional full TCP/IP stack. I suppose it only helps if the CPU is busy with other things then.

Re:How ... (1)

Haeleth (414428) | more than 7 years ago | (#17179308)

F.E. winhoze sucks rotten eggz as far as Layer 3 based HA, anicast and routing and this will allow you to bolt on that on top of a Winhoze application by running a routing protocol stack on the embedded engine.

That's fascinating, but I use Windows, not Winhoze. Will this affect performance for me?

Re:How ... (1)

thegrassyknowl (762218) | more than 7 years ago | (#17179352)

It can't reduce ping times in the real world - ping is almost entirely a function of the network bus, not the network card.

Sure, it could do neat things with the queueing so that ICMP echo to get it to the wire earlier, but you can do that in the OS anyway.

The one killer function you need in a NIC (gig and above) is the ability to calculate and verify checksums so the OS doesn't have to. A huge buffer on the card so it can do big bursts of DMA instead of lots of smaller ones would help too.

Those things would make all real world apps more snappy 'cos the OS wouldn't have to be continually wasting time doing them - it could be busy shuffling data to/from disk and doing things an OS is supposed to do.

All I can say is... (2, Funny)

MoralHazard (447833) | more than 7 years ago | (#17178326)

this card is for the gamer that needs two clocks: one set to tomorrow and one set to Tokyo time, so he knows when to drift race.

(sorry for mangling to PA quote)

Seriously, $200? WTF?

ONLY useful for world class gamers/richie rich (4, Insightful)

Bananatree3 (872975) | more than 7 years ago | (#17178772)

This card is most useful for worldclass gamers who could use a 10ms deduction in ping/lag times. For most gamers this is very neglagable, but some gamers can take good advantage of that 10ms. Also it looks like some games have more benefit from this than others. The article lists FEAR and WoW as the most positively effected while Counterstrike and Quake 4 as minimally effected. They list an example in TFA of an _expert_ FEAR gamer:

I [have] watched him repeatedly go into DM games and use nothing but dual pistols and own the map. Joshua saw distinct differences between his Killer and non-Killer sessions. He was easily able to identify his gaming session using the Killer NIC with confidence. One of the more notable things he conveyed to me was that his "machine was ahead of what was on the server." He explained this was allowing him to take shots and get cover before others had time to react to his presence. Joshua said the the onboard NIC felt the same as his machine at home, but the Killer NIC gave him a better experience in that it, "felt smoother with less lag kills." He went on to note that the overall reaction felt better and the Killer NIC supplied a smoothness of play he did not get with the onboard NIC.

These kinds of "professional" gamers could use a fancy NIC with lower times. Or if your Richie Rich and you need some extras for your already pimped out gaming rig.

Re:ONLY useful for world class gamers/richie rich (1)

DAldredge (2353) | more than 7 years ago | (#17179116)

I am sure the stack of money in front of the 'killer' NIC enable machine helps as well.

Check their home machines. (4, Insightful)

khasim (1285) | more than 7 years ago | (#17179250)

These kinds of "professional" gamers could use a fancy NIC with lower times.

In all the "reviews" of this that get posted here, I notice a few recurring items.

One of the most interesting to me is that they want the "gamers" to test the NIC as part of their entire box. But the real gamers would already have a box built to their specs that they were familiar with ... their home gaming machine.

Yet the "gamers" never seem to insist that they be allowed to compare the KillerNIC in their own box, against their existing NIC. And if they're serious gamers, they've already spent money replacing the on-board NIC if their motherboard came with it.

Kind of like if a tire company wants you to like new tires, but they won't let you drive them on your own car. You have to use their car. And you have to compare it to a different car that they have without the tires. And people accept that.

Under those conditions, I can show you improved ping times using nothing more than cool stickers for your case.

Bashing (1)

Broken scope (973885) | more than 7 years ago | (#17178330)

We were ripping at it a few months back... so we are afraid to talk about it.

However, was this on a net connection, across a cable line? Or was it on a high end switch setup with good cables running at gigabit speeds? No I did not read TFA, I'm just raising a question so I can be yelled at and answered.

Re:Bashing (5, Funny)

Reverend528 (585549) | more than 7 years ago | (#17178472)

It requires a fiber-optic cable because it needs to be able to send the photons in a superposition of states. By the time your program gets around to sending a packet, the photons are most of the way there and merely need to collapse into the same state as the packet. The naysayers who claim that the card can't actually improve latency are only thinking in terms of classical physics.

Re:Bashing (1)

maxwell demon (590494) | more than 7 years ago | (#17178670)

Of course also quantum mechanics doesn't allow you to do FTL communication. See http://en.wikipedia.org/wiki/No-communication_theo rem [wikipedia.org]

Re:Bashing (4, Funny)

Reverend528 (585549) | more than 7 years ago | (#17178706)

Clearly you didn't read the review. The card works.

Maybe you shouldn't accept everything you read on wikipedia as scientific fact.

Re:Bashing (4, Funny)

grommit (97148) | more than 7 years ago | (#17178820)

Yeah, you definitely need to have a Cisco router with Monster Cable CAT 6 running between it and your computer to make full use of Gigabit. Anything less and you're dropping packet fidelity through crosstalk line expansion. Not to mention the undervoltage latencies.

Been done before (5, Funny)

Anonymous Coward | more than 7 years ago | (#17178338)

Killer network cards have been around for so long, there's actually a Localtalk version here:
http://www.fiftythree.org/etherkiller/ [fiftythree.org]

Game studio did their own test... (5, Informative)

bl4nk (607569) | more than 7 years ago | (#17178346)

I have a friend who works at a game studio who had their IT guy perform several tests to see if it did anything... chat log:

> killer NIC. bad.
> file transfers = 1/4 of the speed of a normal NIC
> the drivers are fucking TERRIBLE to install/uninstall/update, you have to reboot. then it'll let you do whate you need to do. then reboot AGAIN...
> and when it does start working, there is literally no difference in either framerate or ping, even on the games they say it specifically improves

This wasn't entirely scientific... (4, Insightful)

nxtw (866177) | more than 7 years ago | (#17178360)

Where's the comparison between different onboard gigabit chipsets? (eg Broadcom, nForce, etc.) Where's the comparison between different PCI, PCI-X, and PCI Expressgigabit NICs?

If applicable, what are the settings for the onboard NICs being tested? Many have options for various CPU offload settings and optimizations for throughput or CPU usage.

Until we see these, how can we be sure if a high-end regular PCI-e NIC won't work just as well?

Mod parent up! (4, Insightful)

khasim (1285) | more than 7 years ago | (#17178968)

And what is this about PLAYING the games? You don't fucking PLAY them! You SCRIPT them so you have as close to the exact same environment as you can get.

From TFA:
We watched the engineers as they played through a Counter Strike deathmatch round.

That is just idiotic.

While this is not the most scientific way to gather data, it certainly did a good job to reflect what would happen in real world gaming situations.

If you aren't going to do it right, then you are doing it WRONG. So it did NOT "reflect what would happen in real world gaming situations".

Bigfoot allowed us to barge into their offices on a beautiful Saturday morning with real gamers in tow.

Again, you script it. You do not play it.

Taking all of our testers one at a time, we allowed them about 45 minutes of gameplay in their chosen title. Our testers were given time to setup their own keyboards, mice, and needed add-on software so that they had a system close to what they were used to using at home.

I'll give the KillerNIC people this, they certainly know how to pick their suckers.

Seriously. They didn't even bring their own PC's? They used the "testing machines" provided for them. And they think this has anything to do with "real world" performance?

From there, the gameplay was divide in half. One half of the online play used the Killer, the other half used the motherboard's onboard NIC.

A far, far better test, even under these biased conditions, would have been for them to use their own PC's. It cannot be that difficult to swap a NIC, can it?

In a blind taste test, more people preferred Coke over the Pepsi that I had previously pissed in.

For some strange reason, all I ever see in these "reviews" are the KillerNIC people insisting that the games be run on THEIR machines. And people who are "reviewing" it accepting this strange requirement. And not even scripting it so that they can compare it with their home machines.

Re:This wasn't entirely scientific... (0)

Anonymous Coward | more than 7 years ago | (#17179388)

Of course it wasn't scientific. It was an advertisement on HardOCP, who then advertised it on Slashdot.

of course (1)

tyroney (645227) | more than 7 years ago | (#17178368)

If you're comparing it to an onboard that shares main cpu it can't help but be better.

Re:of course (2, Insightful)

ZachPruckowski (918562) | more than 7 years ago | (#17179076)

Most games are GPU-bound anyways. And most games are single threaded. Add it up, and you find that most people who can afford a $250 NIC have CPU power to spare (most likely dual core, too). Therefore, it'd be useful if you were on a budget system, but at $250, you would almost always spend that on the GPU or RAM.

killer nic vs onboard nics (1)

El_Muerte_TDS (592157) | more than 7 years ago | (#17178372)

but no word on what the onboard nics were. And there didn't compare it with other nics either.
So for what it's worth, the 'killer' nic is probably better than the most cheap ass realtek nic. A 3com 3c905 nic costs hardly costs $30 and performs way better than the average realtek stuff.

Re:killer nic vs onboard nics (1)

dave562 (969951) | more than 7 years ago | (#17178598)

A 3com 3c905 nic costs hardly costs $30 and performs way better than the average realtek stuff.

The last time I did an sort of comparison with NICs I used that exact card. I had an Asus motherboard with an onboard NIC. The NIC fried so I scavanged a 3c905b out of my old Pentium III. My ping times in America's Army were SIGNIFICANTLY better. Those 3c905s are GOOD NICs. There may be other 100BT NICs that are better, but I doubt that those that are better are better by much.

Re:killer nic vs onboard nics (1)

Zonekeeper (458060) | more than 7 years ago | (#17178936)

Wow, someone else who loves 3Com. Personally, after working with and installing 3Com hardware from nics to managed switches for years, I now won't use ANYTHING that 3Com makes, if I can at all avoid it. I never dealt with such dodgy hardware in all my life.

Re:killer nic vs onboard nics (1)

profplump (309017) | more than 7 years ago | (#17179118)

I don't love 3com in general, just the 3c905. It's got a well-maintained driver in OS X, Windows, Linux, FreeBSD and OpenBSD, going back to versions of those systems from 1998 or earlier (where applicable). Recent versions (i.e. rev B and later) include the acceleration features you'd expect in a modern NIC (checksuming, segmenting, etc.). And you can buy them used for $7 a pop. It's a hard deal to beat when you just need to add a network card to a box.

I prefer the Intel EE Pro cards for new machines, for the same reasons. Good driver in almost every OS. Available in PCI/PCI-E/PCI-X versions. Available in 10/100 and 10/100/1000 versions. Not terribly expensive.

For a better experience (1)

Bullfish (858648) | more than 7 years ago | (#17178392)

You are better off buying a good set of speakers and a good sound card. Especially in gaming, the aural experience will make the game more entertaining and the card will take a load off your cpu which will buy a few FPS's. This smacks of the physics cards and we know where that market went. So the top gamers will see a benefit. So what? The fact that they are in that category means they wax the rest of us regularly. How will this make that more enjoyable.

If nothing else, when it comes to enjoying the game, it sounds like the money is better spent on a good single malt to sip on during.

Re:For a better experience (1)

Broken scope (973885) | more than 7 years ago | (#17178484)

Physics cards have a chance. Though with the huge boost were going to see in CPUs in the next few years, I think your right.

Re:For a better experience (1)

Proud like a god (656928) | more than 7 years ago | (#17179014)

Go google GPGPU, Nvidia Quantum, Havok FX.

The more interesting thing (2, Interesting)

the_humeister (922869) | more than 7 years ago | (#17178394)

What's more interesting is that the card is actually a single-board computer with PowerPC processor and 64 MB of RAM!

Re:The more interesting thing (1)

Square Snow Man (985909) | more than 7 years ago | (#17178476)

i lol'd

Re:The more interesting thing (2, Funny)

Anonymous Coward | more than 7 years ago | (#17178540)

What's more interesting is that the card is actually a single-board computer with PowerPC processor and 64 MB of RAM!

So can you run Linux on it then? You could ssh to it from Windows and get some real work done.

Re:The more interesting thing (1)

tomhudson (43916) | more than 7 years ago | (#17178738)

Actualy, it DOES run linux - and they'll pay you to write applications that run on it.

There's one out already - an iptables-based firewall.

Re:The more interesting thing (1)

maxwell demon (590494) | more than 7 years ago | (#17178774)

So can you run Linux on it then? You could ssh to it from Windows and get some real work done.
According to TFA it does run Linux:
To put it simply, the Killer NIC is actually a Linux system-on-chip. That basically means that it is its own little Linux computer inside what is likely your Windows box. The engineers at Bigfoot have given you access to this machine.

Obligitory (0, Redundant)

Watson Ladd (955755) | more than 7 years ago | (#17178578)

Can it run Linux?

Re:Obligitory (1)

Schraegstrichpunkt (931443) | more than 7 years ago | (#17178782)

Apparently so.

Why? (5, Funny)

Anonymous Coward | more than 7 years ago | (#17178420)

Business Development Guy: We need a new product. Something... niche. Something overpriced. I know! Don't we still have a bunch of boxes of old network cards?
Hardware Engineer: Uh, yeah. We were about to offload them on eBay to make room in our closet...
Business Development Guy: No! Lets tack on some parts from China and sell them! We'll call it, hmm.. the Killer NIC! Since no one wants to buy NIC cards, we'll overprice them for no apparent reason! $250 a pop!
* Hardware Engineer bashes his forehead on the desk.
Hardware Engineer: You've got to be kidding me. Isn't that, like, fraud?
Business Development Guy: Not at all. We'll just never say how it works, only that it works. The processor will be for like, decorative purposes. Consumers love that kind of stuff!

Snake oil (3, Insightful)

pilkul (667659) | more than 7 years ago | (#17178438)

Latency is 99% percent due to delays over the Internet, not anything that happens on your local machine. What does this card do, sprinkle magic fairy dust over packets so they go faster through the wire?

This reminds me of gold-plated power cords for sound systems. Guaranteed to create richer, deeper sound!

It's not magic fairy dust... (2)

countvlad (666933) | more than 7 years ago | (#17178500)

it's lubricant, so your rockets travel through the pipes faster than the other guys.

Its not lubricant or pipes ignore Sen Ted Stephens (1)

bug1 (96678) | more than 7 years ago | (#17178994)

What happens is that your rockets get loaded a truck and an internet delivers those trucks to the server, its faster because your rockets are traveling en-mass.

Senator Ted Stephens started a rumor claiming that "the internet is not a big truck, its a series of tubes" dont listen to him, we all know rockets cant fly down tubes, they cant even steer.

http://en.wikipedia.org/wiki/Series_of_tubes [wikipedia.org]

Holy shit (5, Insightful)

Anonymous Coward | more than 7 years ago | (#17178442)

Wow, this 'article' is a real gem.
This more recent fast moving game was used by Keith, our software tester from Dell. He reported ping times during his Killer NIC session to be about 10ms faster on the server that was returning 50 to 70ms pings with the onboard NIC. When quizzed about a discernible difference in Q4 session quality, Keith did not express anything to us that really set one session off against the other. The Killer NIC and Quake 4 seemed to be a wash at best.
I think that's the only time they actually note anything about "latency reduction" in this advert. I leave it as an exercise to the reader to find all the flaws in this testing methodology. What is even more amazing, given the other 3 accounts do not give any confirmation whatsoever is that they come to the most ridiculous conclusion.
The Killer NIC does exactly what it is advertised to do. It will lower your pings and very likely give you marginally better framerates in real world gaming scenarios....
Paid product placement at its finest I suppose. But even the vetted boys at HardOCP realized they would have to blow some reality up everyone's ass after all that smoke, and added the following to be "fair and balanced."
...The Killer NIC is not for everyone as it is extremely expensive in this day and age of free onboard NICs. There are very likely other upgrades you can make to your computer for the same investment that will give you more in return.

Designed to seperate fools from their money (3, Insightful)

Pizaz (594643) | more than 7 years ago | (#17178494)

This NIC card has no clothes. But hey, phishing schemes and Nigerian con artists can be successful so why shouldn't this?

Re:Designed to seperate fools from their money (2, Funny)

maxwell demon (590494) | more than 7 years ago | (#17178880)

You probably just forgot to buy the gold-plated ethernet cables as well. BTW, that combination also improves the sound quality of audio files transmitted over the network.

Malfunctional card (1)

Anonymous Coward | more than 7 years ago | (#17178522)

What I understand about previous readings about this product, the card tampers with the ethernet protocol to send more than it's supposed to. Here's another great idea; make a card that detects when another player on the LAN is broadcasting and make a deliberate "collision" so that the data packet needs to be resent thus creating more lag. Great for LAN parties, coming soon from the same vendor, I presume.

So they only compared to onboard? (1)

Jezebeau (917315) | more than 7 years ago | (#17178560)

That seems a little... ridiculous. What are we going to see next, ATI insisting their new card is 10x more capable than the average onboard graphic processor?

This review was a sham. (1)

rei_slashdot (558039) | more than 7 years ago | (#17178570)

They performed the "tests" on the company's machines. What a joke.

Re:This review was a sham. (1)

supremebob (574732) | more than 7 years ago | (#17179248)

Considering that HardOCP's web site has been covered with those obnoxiously animated KillerNIC ads for the past three months, I wouldn't be at all surprised if they "tweaked" their testing methods in order to come up with a favorable opinion of the card.

People have already been giving Kyle Bennett a hard time about those ads already, so I'm surprised that he escalated the issue even further by being one of the few hardware review sites to give it a positive review. I can only imagine how many "HardOCP sells out!" style editorials are going to come out of this decision.

Anandtech == Better Review (5, Informative)

fineghal (989689) | more than 7 years ago | (#17178586)

Anandtech has a much better review here: Linky: http://www.anandtech.com/cpuchipsets/showdoc.aspx? i=2865 [anandtech.com]

**MOD PARENT UP** (1)

Bananatree3 (872975) | more than 7 years ago | (#17178802)

Anandtech article is FAR more extensive, listing actual benchmarks, test setup, etc.

Where are the benchmarks? (5, Insightful)

Animats (122034) | more than 7 years ago | (#17178588)

I don't see any benchmarks in that article. Here are some, [extremetech.com] , and they don't make the thing look all that impressive.

The only benefit in this thing, apparently, is that, for games which make too many "select()" polls, there's a faster no-data return. This is really a bug in the game, which ought to be multi-threaded by now. As games are revised for multi-core systems, this problem had better go away. In fact, it probably will go away in Vista, which has a multithreaded network stack.

And the question I'd have (2, Insightful)

Sycraft-fu (314770) | more than 7 years ago | (#17179036)

Is not how does it compare against an onboard NIC, but how does it compare against a good NIC that does offloading, like an Intel server NIC? I mean $100 will get you Intel's copper 1000mbps server NIC which does support offloading of various functions and, unlike the killer NIC, has rock solid drivers (not just for Windows either). My bet? The Intel NIC probably does near as good a job, and doesn't have any problems, as well as saving you $150.

Comparing it only to cheap onboard NICs really isn't useful. I mean yes I'd be interested to know if it's better but the real question is if it's better than a high quality addon NIC that's already available.

stupid suggestion... (2, Insightful)

TheSHAD0W (258774) | more than 7 years ago | (#17178632)

Maybe where this NIC really belongs is on the MMORPG *servers*?

Re:stupid suggestion... (1)

JRW129 (823295) | more than 7 years ago | (#17178914)

hahaha, yes, that would make their fiber networks run faster

Numbers, please? (1)

Schraegstrichpunkt (931443) | more than 7 years ago | (#17178662)

Exactly how much is latency reduced by this card?

Re:Numbers, please? (1)

Bluesman (104513) | more than 7 years ago | (#17178788)

Well, normal latency reduction is around ten, and these cards, THESE ones right here, they go to eleven.

Just get an Intel NIC (2, Informative)

York the Mysterious (556824) | more than 7 years ago | (#17178696)

It seems a like a lot of what this NIC claims to do (offloading TCP work such as checksumming and segmentation) can be done by an Intel NIC on Windows and Linux (2.6.19 and up at least). So go get a gig Intel NIC for the $35 and invest the rest in a new mouse pad or something that might actually impact your gaming a bit more than that silly nic. http://shopping.yahoo.com/p:Intel%20PRO%2F1000%20P T%20Desktop%20Adapter:1993037566;_ylt=Apu.Yd0SJ9f3 4z.56feYYLIbFt0A;_ylu=X3oDMTBic2hxMGNhBGx0AzQEc2Vj A3Ny?clink=dmps/intel_pro.2f.1000_pt/ctx=mid:5,pid :1993037566,pdid:5,pos:2,spc:14489115,date:2006120 9,srch:kw,x [yahoo.com] :

Re:Just get an Intel NIC (1)

jours (663228) | more than 7 years ago | (#17179390)

Supposedly the card is optimized for UDP traffic, which is what most games use, rather than TCP. Not saying it works...and certainly not worth $279 to me...

This card has its own processor, right? (1)

Schraegstrichpunkt (931443) | more than 7 years ago | (#17178810)

How long until we have botnets made up of these things?

Killer NIC $250? (2)

JRW129 (823295) | more than 7 years ago | (#17178888)

Hey, Here's an idea: If you don't think it's worth $250 DON'T BUY IT! It's not a hard concept to grasp, I know I myself had a problem with that at one point when I was younger. Stop your bitching and STFU!

Re:Killer NIC $250? (1)

BillyBlaze (746775) | more than 7 years ago | (#17179154)

Here's a clue: if you don't want to read about whether a product is worth it, don't click on a Slashdot article about it. Slashdot would suck if nobody shared their opinions. It would be improved, however, if you reserved your complaining about complaints for articles about how often one should complain.

Re:Killer NIC $250? (1)

schon (31600) | more than 7 years ago | (#17179384)

Hey, Here's an idea: If you don't think it's worth $250 DON'T BUY IT!
Yeah, because god knows that when a company starts selling snake-oil, nobody should ever call them on it, right? I mean, when you see someone telling lies soley so they can make money from the ignorance of others, it's unethical to point that out to people who might otherwise be taken in.

Stop your bitching and STFU!
You first.

You guys completely missed the point (0)

Anonymous Coward | more than 7 years ago | (#17178918)

While overpriced for a NIC, it also has the side effect of having its own onboard RAM and reprogrammable (400mhz) processor. Thus making it possible to plug a USB hard drive into the back of the card, run a bittorrent client with its onboard proceesor, and never use a single cpu cycle.

That alone for some people would be worth it.

Re:You guys completely missed the point (1, Insightful)

Anonymous Coward | more than 7 years ago | (#17179018)

What people? People who can't strain their 386's with the added weight of a torrent client?

For the price of this NIC, you can invest that money into a good upgrade.

Stupid.

NPU (1)

Bromskloss (750445) | more than 7 years ago | (#17179004)

Ooo, another processing unit to help out the "central" one with heavy mathematical computations! Let's see, we have the CPU, of course, the GPU, the DSP on the expensive sound card, myself, tagging along with my calculator and now the NPU! We're getting fast here.

Wait, so... (1)

rob1980 (941751) | more than 7 years ago | (#17179038)

Bigfoot allowed us to barge into their offices on a beautiful Saturday morning with real gamers in tow.

I'm calling bullshit on this article until they perform their tests someplace other than on the card manufacturer's network.

Pfft... (2, Funny)

flatt (513465) | more than 7 years ago | (#17179124)

While Weird Al is waxing his modem to make it go faster, I've found that cheetah blood rubbed on a network card will significantly reduce latency.

*.PU (1)

n1hilist (997601) | more than 7 years ago | (#17179130)

CPU, FPU, GPU, PhysixPU, NPU...

What's next, OPU? One Processor to Rule Them All (TM)?

Wow it hacks into my router and makes it better? (1)

Xest (935314) | more than 7 years ago | (#17179138)

I ping from my router to my ISP and get and average of 30ms ping, I do the same from my computer and get 31ms ping. I'm intrigued to know how it can possibly increase my latency without making some modification to my router to make it communicate faster with the exchange which let's face it, is practically impossible.

Of course they could claim my router is the bottleneck but then, seeing as it's impossible for me to get an ethernet port connected directly to some internet backbone then I have no use for this card anyway.

More to the point however, this is around the 4th time this has been posted on Slashdot, not another "Slashdot repeated news postings whine" but more intrigued to know who this KillerNIC company is paying to keep getting their lies spread all over the internet?

Game Servers (5, Informative)

dysfunct (940221) | more than 7 years ago | (#17179182)

If you see this and ponder buying this card for your game servers first try optimizing the Linux kernel. 1000 Hz ticks, big kernel lock preemption and other latency patches actually does wonders to ping times (and I do not mean ICMP echo or similar but ping packets answered by the game) and latency.

Many games have their own interesting capabilities for performance tuning. For instance Counter Strike 1.6 has the -pingboost setting which will switch between select() and alert() syscalls (10 ms reduction) or processing a frame for every packet. Other games have similar tuning options that will enhance performance. Then there's also tuning your network settings.

By the way, as far as I remember this Killer NIC is just some kind of offload engine. How *exactly* does this increase performance when most game specific packets are simple UDP packets that performance-wise are not as demanding as TCP packets (less checksums, no window scaling and other options easily tunable etc.)?

Imagine... (1)

DrJokepu (918326) | more than 7 years ago | (#17179220)

Imagine a Beowulf cluster of these!

Yes, but... (1)

The Real Nem (793299) | more than 7 years ago | (#17179224)

...can it lower /. refresh times?

I feel some first post karma coming my way...

If I got $250 for a NIC... (1)

Hymer (856453) | more than 7 years ago | (#17179310)

...I would buy a multiport server NIC.

Not sure (1)

warrior_s (881715) | more than 7 years ago | (#17179432)

I am not sure if someone already said this in this discussion. Lower ping times can be achieved if you abuse the implementation of TCP in your network stack. If your TCP implementation do not backoff upon congestion detection, you will starve other TCP flows but your latency will reduce.
Load More Comments
Slashdot Login

Need an Account?

Forgot your password?

Submission Text Formatting Tips

We support a small subset of HTML, namely these tags:

  • b
  • i
  • p
  • br
  • a
  • ol
  • ul
  • li
  • dl
  • dt
  • dd
  • em
  • strong
  • tt
  • blockquote
  • div
  • quote
  • ecode

"ecode" can be used for code snippets, for example:

<ecode>    while(1) { do_something(); } </ecode>