Unless you have plenty of $$, there is no reason, none , at all, to get the Killer NIC over the Intel NIC. Even if the Killer was priced at 100$$.
Bigfoot is for Bigfools.
Bigfoot is for Bigfools.
- Status
Uhm. is it not still plugged into a pci bus, arguing its way to a function? The net is like diarhea, the lan card is toilet paper...there is no flush, there is a trash can. Back to the 15 dollar 100mbit lan card for me...
The changing latencies is pci arguing more data. My pci nic is also 35ms in intense gamng on 3mbit connect. It changes to the net, but sits there like the tank it is on its own little clock. The killer nic is a gimmick, altho, some big servers may get a giggle out of a rare useful function.
The changing latencies is pci arguing more data. My pci nic is also 35ms in intense gamng on 3mbit connect. It changes to the net, but sits there like the tank it is on its own little clock. The killer nic is a gimmick, altho, some big servers may get a giggle out of a rare useful function.
As everyone said. Why compare only to the Realtek??????????? ?x1,000,000
Throw a Intel and Marvell into the mix. Or not, and just compare with the standard weaker controller. Sad.
And then the price. Compare this card to any other good NIC and see just how the price/performance comes out. P.S. It was a poor price/performance to begin with. Better use for your money.
Throw a Intel and Marvell into the mix. Or not, and just compare with the standard weaker controller. Sad.
And then the price. Compare this card to any other good NIC and see just how the price/performance comes out. P.S. It was a poor price/performance to begin with. Better use for your money.
husky mctarflash
Distinguished
- Dec 3, 2007
- 1
- 0
- 18,510
Perp--be sure to let us know how that works out. I will likely do the same.
Only question is do I buy the $24 PCI one or the $44 PCIe one? (although it is $35 at Tiger)
Only question is do I buy the $24 PCI one or the $44 PCIe one? (although it is $35 at Tiger)
The article clearly points out that lower end systems with slower CPUs may benefit more from the off loading. This makes perfect sense.
I wish you had also covered different connectivity scenario. Will the performance benefits from this NIC increase has available bandwidth goes up. Following the same logic you used for CPUs, I would imagine that in bandwidth limited situation (let's say 1.5Mbps/384Kbps DSL) the card can't really shine. This might be different on Fios or ADSL2+ in Europe, where bandwidth can exceed 20Mbps.
Finally, would this card trully shine in LAN party scenario, where everyone is connected at LAN speed (100Mbps if not GigE.
It would have been useful as well to provide details of the WAN environment use for your tests.
I wish you had also covered different connectivity scenario. Will the performance benefits from this NIC increase has available bandwidth goes up. Following the same logic you used for CPUs, I would imagine that in bandwidth limited situation (let's say 1.5Mbps/384Kbps DSL) the card can't really shine. This might be different on Fios or ADSL2+ in Europe, where bandwidth can exceed 20Mbps.
Finally, would this card trully shine in LAN party scenario, where everyone is connected at LAN speed (100Mbps if not GigE.
It would have been useful as well to provide details of the WAN environment use for your tests.
How exactly that 'K' shape thing become the heat sink. I understand it looks cool but not efficient for a heat sink.
as old timer say.. your internet connection only as fast as your slowest one on the network..
which mean the slow down most due to the router. only thing this NIC may do is free up maybe 2~3% of CPU power?? if it work at all.. therefore 200$ is wayyyy to much.
tcp/ip is tcp/ip, if you jump 20 node before it reach to target is going to slooooow no matter what.
which mean the slow down most due to the router. only thing this NIC may do is free up maybe 2~3% of CPU power?? if it work at all.. therefore 200$ is wayyyy to much.
tcp/ip is tcp/ip, if you jump 20 node before it reach to target is going to slooooow no matter what.
I have to wonder when the hype for this card is ever going to let up. All I have to say is this manufacturer must have deep pockets to not only continue to promote this... what I would call a marginal product... but to also continue to exist considering that people cannot be beating down their doors to buy it! There was a lot of hype about this ...Mmmmm....two years ago now? At that time the consensus from all the various online tech sites was negligible performance gains at best. Now we have Toms and a couple of other sites rehashing the exact same product again. Have there been any noticeable improvements in this card over the past couple of years? Is this the next generation of Killer? We should make a pact 
No more testing this thing until they actually send out a revision, or completely new card that actually has an impact worth 250.00 to 300.00
No more testing this thing until they actually send out a revision, or completely new card that actually has an impact worth 250.00 to 300.00
zenmaster
Distinguished
- Feb 21, 2006
- 41
- 0
- 18,590
At best, I would say their tests were significantly flawed.
How exactly they are flawed, I am not sure but they did something SERIOUSLY wrong.
I use a lowly onboard NIC on my system.
The Ping Response Time from my local router to my system is less than 1ms while doing a PING loop.
The only role the local NIC plays in the process is the transmittal and receival of the packet. What happens while the ping request is travelling on other systems plays ZERO role.
Hence, the potential improvement in the ping test is less than 1ms.
And what is up with some 'Human 1-2-3' response?
You setup a PING to run 10,000 times, record the computer calculated response. Run these tests at the same time on two different systems and you are done.
If I were to give them the benefeit of the doubt, is that the connection between then and where they were testing changed between the time the tests were being run.
It's a terrible article.
Once they realize that their PING results show a significant problem they need to address that issue and once it's resolved retest.
Perhaps they had duplex set wrong on the NIC.
But something was severley amiss.
How exactly they are flawed, I am not sure but they did something SERIOUSLY wrong.
I use a lowly onboard NIC on my system.
The Ping Response Time from my local router to my system is less than 1ms while doing a PING loop.
The only role the local NIC plays in the process is the transmittal and receival of the packet. What happens while the ping request is travelling on other systems plays ZERO role.
Hence, the potential improvement in the ping test is less than 1ms.
And what is up with some 'Human 1-2-3' response?
You setup a PING to run 10,000 times, record the computer calculated response. Run these tests at the same time on two different systems and you are done.
If I were to give them the benefeit of the doubt, is that the connection between then and where they were testing changed between the time the tests were being run.
It's a terrible article.
Once they realize that their PING results show a significant problem they need to address that issue and once it's resolved retest.
Perhaps they had duplex set wrong on the NIC.
But something was severley amiss.
zenmaster
Distinguished
- Feb 21, 2006
- 41
- 0
- 18,590
I would also like to note their discussion of CPU utilization....
I seen PI and PII servers w/o advanced NICs handle 1000s of user reading/writing files with the CPU sitting at 1-2%.
Now, Much of that 1-2% was actually the OS handling reading/writing files and processing requests. Not forming packets.
Considering the number of packets required to play the game would be far less, I can't imagine this NICs processor being of any significant use.
Common guys.
ANALYSE WHAT YOU WRITE!
I seen PI and PII servers w/o advanced NICs handle 1000s of user reading/writing files with the CPU sitting at 1-2%.
Now, Much of that 1-2% was actually the OS handling reading/writing files and processing requests. Not forming packets.
Considering the number of packets required to play the game would be far less, I can't imagine this NICs processor being of any significant use.
Common guys.
ANALYSE WHAT YOU WRITE!
If you want to see how well it helps offload from the CPU,
try rendering a 3d scene with 3d studio or Maya (CPU 100%)...etc while using another machine to send/receive huge amounts of data accrossed the network
If the claims are true you should see a much larger difference with the M1 NIC. Standard NICs should be at the mercy of the CPU and the M1 should hardly be phased.
Just a thought
try rendering a 3d scene with 3d studio or Maya (CPU 100%)...etc while using another machine to send/receive huge amounts of data accrossed the network
If the claims are true you should see a much larger difference with the M1 NIC. Standard NICs should be at the mercy of the CPU and the M1 should hardly be phased.
Just a thought
zenmaster
Distinguished
- Feb 21, 2006
- 41
- 0
- 18,590
[citation][nom]ntrceptr[/nom]If you want to see how well it helps offload from the CPU,try rendering a 3d scene with 3d studio or Maya (CPU 100%)...etc while using another machine to send/receive huge amounts of data accrossed the networkIf the claims are true you should see a much larger difference with the M1 NIC. Standard NICs should be at the mercy of the CPU and the M1 should hardly be phased.Just a thought[/citation]
That would be a good point.
As a result, you would want to check to see if your game maxed all of your cores.
If you have more cores than it can Max, then it would not help much.
However, if you are short on cores it may help.
That would be a good point.
As a result, you would want to check to see if your game maxed all of your cores.
If you have more cores than it can Max, then it would not help much.
However, if you are short on cores it may help.
The results show this card would be more of a benefit to people with less then cutting edge computers then to people who already have the top of the line equipment. I would rather spend the 200-250$ on an upgrade to my CPU or GPU. On the other hand if this card drops to under 100$ and can show a marked improvement for gaming using more budget equipment possibly the use of this card and ones like it will grow.
I own one of these cards and never seen any decrease in ping when playing wow, also.. you see that massive heatsink on the chip.. well its there for a very good reason.. this is one of the hottest chips ive ever not had the pleasure of having in my computer.. you seat one of these below your every expensive v/card and temps on your gpu/ram will go up ... way up.. i do not recomment this to anyone... not even those with watercooled rigs as your case temps will increase with this garbage..
If you're going to use conviction and the voice of 'authority' to try to get your points across in a review, it would serve you well to know what you're talking about. For one thing, there's no "minimum legal" IP packet size, and, if there were, it would be 20 bytes (the size of an IP header), not 64. I can very well send a 28 byte ping, though it won't give me time statistics under UNIX/Linux, since they embed a 1-byte UNIX timestamp (though Windows uses an internal tick clock [which is wildly inaccurate] to measures its RTT).
Second, I don't know what "layers" of the "TCP/IP" stack you're talking about, but I don't have them on my computer. For a start, there's no "TCP/IP stack." I'm not sure where that gem got started, but it's a real peeve of mine. An IP stack, sure, but TCP has very little to do with IP itself. It's just another transport protocol (number SIX, in fact). ICMP, IP protocol number ONE, is much more integral to IP than any transport protocol. Second, the only "layers" to IP might be whatever you want to stack ontop of it, but there's nothing inherently layered about IP itself.
Usually, when someone uses that kind of language, they're referring to the OSI model, where Layer 1-3 would mean Physical, Data Link, and Network, and, not, as I guess you were attempting to go for (?), 'IP, TCP, and Session.'
As far as the predictability stupid comments go, particularly the one I've seen re-iterated time and time again about "MY LAN CARDZ SEEZ ONE MILLISECONDS SO HOW CAN IT IMPROEVS THE INTERNET???," uh. Yeah, sure. Your LAN is 0.2 ms. Now, think that through. A connection to the internet generally isn't, particularly under load.
Second, I don't know what "layers" of the "TCP/IP" stack you're talking about, but I don't have them on my computer. For a start, there's no "TCP/IP stack." I'm not sure where that gem got started, but it's a real peeve of mine. An IP stack, sure, but TCP has very little to do with IP itself. It's just another transport protocol (number SIX, in fact). ICMP, IP protocol number ONE, is much more integral to IP than any transport protocol. Second, the only "layers" to IP might be whatever you want to stack ontop of it, but there's nothing inherently layered about IP itself.
Usually, when someone uses that kind of language, they're referring to the OSI model, where Layer 1-3 would mean Physical, Data Link, and Network, and, not, as I guess you were attempting to go for (?), 'IP, TCP, and Session.'
As far as the predictability stupid comments go, particularly the one I've seen re-iterated time and time again about "MY LAN CARDZ SEEZ ONE MILLISECONDS SO HOW CAN IT IMPROEVS THE INTERNET???," uh. Yeah, sure. Your LAN is 0.2 ms. Now, think that through. A connection to the internet generally isn't, particularly under load.
Killer the only good thing it does is KILL YOUR WALLET. Purchased one could see no differance when playing BF2 went back to my onboard network connection, Motherboard Asus M2N-SLI Delux AMD 6400 Black Box , 4 Gig Mem 8800GTX and tried 8800GT.
Worst investment i have made.
Worst investment i have made.
It's interesting that the unsatisfied customers have been using Athlon 64s and Tom's Hardware tested on an Intel CPU. AMD CPUs have faster memory access due to their integrated memory controllers (and Intel CPUs will soon), which might negate one of the obvious benefits of this card: its dedicated memory. The other obvious benefit of this card, the added compute power, raises the question of whether it would be more cost effective to add one of these, or to add an additional CPU core to handle network-related computation. To resolve this question, I would like to see a review comparing a Phenom X3 + Bigfoot NIC with a Phenom X4 + regular NIC. A quick look at newegg shows that the cost of the additional core at 2.4 GHz is $35 -- a far cry from the price of the Bigfoot NIC. And we all know that CPUs of the future will incorporate more and more cores, more than the games of today can utilize.
Strange that Tom's chose a mid range machine to run these tests on. One would think if you are appealing to gamers who run high end machines that your would also be high end as well. Showing that latency was reduced on dual core machines with a single 8600GT under the hood hardly constitutes a high end gaming rig at this moment.
At the very least I would have expected to see a quad core processor and a geforce 8800GT. I agree comparing to onboard NICs is also unfair, but not even using quad cores which are quickly becoming standard, then stating it will reduce latency by unloading from the CPU... hardly fair.
Also a very crafty move: they change the subject referring to the processor instead of the NIC in the last paragraph so anyone skimming the article would see the final statement: "We?re impressed!"
Certainly not condoning those who skim through articles and make their decisions based on the final paragraph, or summary, but c'mon guys! I have always trusted Tom's for an unbiased and fair review of new technology... Until this blatant attempt to push a $200+ NIC on people without even providing any evidence it will actually do any real good. If you want to offload CPU calls from the networking API, I agree a simple PCI or PCIe NIC would probably give something near par performance gains. But if you are running a high end set up I really doubt you would see any added gains, and certainly non that would warrent the price tag.
This NIC has what look to be a lot of cool features, that the price is way to steep. It will appeal to those who only skim articles, those who like cool heatsinks (assuming their Sli / Triple Sli system can even fit a PCI NIC... quickly becoming impossible), and people who are nieve enough to think that this slanted farce of a review is in any way objective.
Cheers!
At the very least I would have expected to see a quad core processor and a geforce 8800GT. I agree comparing to onboard NICs is also unfair, but not even using quad cores which are quickly becoming standard, then stating it will reduce latency by unloading from the CPU... hardly fair.
Also a very crafty move: they change the subject referring to the processor instead of the NIC in the last paragraph so anyone skimming the article would see the final statement: "We?re impressed!"
Certainly not condoning those who skim through articles and make their decisions based on the final paragraph, or summary, but c'mon guys! I have always trusted Tom's for an unbiased and fair review of new technology... Until this blatant attempt to push a $200+ NIC on people without even providing any evidence it will actually do any real good. If you want to offload CPU calls from the networking API, I agree a simple PCI or PCIe NIC would probably give something near par performance gains. But if you are running a high end set up I really doubt you would see any added gains, and certainly non that would warrent the price tag.
This NIC has what look to be a lot of cool features, that the price is way to steep. It will appeal to those who only skim articles, those who like cool heatsinks (assuming their Sli / Triple Sli system can even fit a PCI NIC... quickly becoming impossible), and people who are nieve enough to think that this slanted farce of a review is in any way objective.
Cheers!
- Status
TRENDING THREADS
-
Question Which games are you playing right now please suggest me some- Started by happyclub
- Replies: 0
-
-
Question I need a best partition manager software for PC Windows 11
- Started by synthia
- Replies: 3
-
-
-
Question local broadcast TV constant shifting - Why??
- Started by oldtimer114
- Replies: 1
-
Facebook
Twitter