When is the Bigfoot Networks Killer NIC review going to be completed?

[nweaver]

Originally posted by: Goosemaster

Originally posted by: nweaver
btw, send me the card, I know how to test network based products...I get paid to do this every single day....

yeah..and send me an x1900xt..I am a great reviewer of expensive and hard-to-find-times

yeah, I wasn't looking for an expensive piece of crap, I hav eno such use for that card. I am, however, a professional QA engineer, and we do tons of network testing for a few compines you may have heard of...like Cisco, Novell, Citrix, Intel, Broadcom, Atheros, HP, etc. Writing a test plan to test that wouldn't have been too tough, but then, I have several years of experience, nad the backing of severeal very good network engineers and software engineers.

What I would hav edone, in a nutshell....

Simulate the WAN, tweak/change it with very specific settings, i.e. X latency, Y packetloss across Z hops.
Script mouse movements to simulate a full server, so you run the game serve,r have 15 join with scripted setting on where to run, when to shoot, etc. Add the 16th (assuming max of 16, but can go up to 200ish pretty easily) from the WAN side of things, and do some precise measurements using multiple tools. Course, that gets a bit spendy, assuming all the scripting and setup, you are looking at 12-60K depending on exact requirements and number of test runs.

 

[spidey07]

Originally posted by: nweaver
What I would hav edone, in a nutshell....

Simulate the WAN, tweak/change it with very specific settings, i.e. X latency, Y packetloss across Z hops.
Script mouse movements to simulate a full server, so you run the game serve,r have 15 join with scripted setting on where to run, when to shoot, etc. Add the 16th (assuming max of 16, but can go up to 200ish pretty easily) from the WAN side of things, and do some precise measurements using multiple tools. Course, that gets a bit spendy, assuming all the scripting and setup, you are looking at 12-60K depending on exact requirements and number of test runs.

You.....you stop with your proper network testing methodolgy (Hats off to AT for understanding the things that they cannot control). It was a great review, but it lacked true analysis and focused on the client's perspective. eitherway, the suckernic is a bogus product targeted to gamers.

 

[spidey07]

Originally posted by: Giffen
An Anandtech person mentioned they were working on a review of this beast....when can we expect the review to be published?

oh, by the way.....

look at join date. Look at the post. Draw your own conclusions. Gotta give props to the shamsters. They understand marketing.

 

[n0cmonkey]

Originally posted by: nweaver
eh....not sure about that Pro 1000 PT, but all my Intel server adapters rock...

and clearly you don't understand that quality means more then speed MachFive...I need uptime, reliablity, and support (are there *nix drivers for the nvidia's yet?). And that nvidia firewall was quality??

forcedeth on Linux
nfe(4) on OpenBSD
nve(4) on FreeBSD
nfe(4) on NetBSD

The nfe is written by Jonathan Gray and Damien Bergamini without documentation. Ain't reverse engineering grand?

The FreeBSD driver, I believe, uses a blob.

I heard the forcedeth driver started out by reverse engineering the chips, but later nVidia employees chipped in with a bit of work.

 

[nweaver]

Sorry to Gary, I didn't mean to sound too harsh...I'm here because I DO like AT and they have pretty decent methedologies versus some other sites. I also wouldn't expect you to do what I would do, but then, someone is paying me lots of money to perform that testing, and they arn't for you guys. Given what you had, you did a fantastic job, and admited that it wasn't the easiest of reviews. I didn't mean to sound like I thought you guys don't know what you are doing, as you guys are a step up from any other site imho.

 

[Gary Key]

Originally posted by: spidey07


You.....you stop with your proper network testing methodolgy (Hats off to AT for understanding the things that they cannot control). It was a great review, but it lacked true analysis and focused on the client's perspective. eitherway, the suckernic is a bogus product targeted to gamers.

We could have tested the card eight different ways and probably tried four or five. A couple of pages of technical information were removed with some analysis as we thought it was too much already. We did complete server side and LAN testing. There just was not anything different in the results to post about but I should have talked about it more than I did.

What it came down to in the end was reviewing the product for the market it was designed for and its purpose in life. That is a client based card designed strictly for Online gaming (Linux apps on the side) and holding it to standards other than that would not have been fair. In regards to other analysis we did our best to crack to the drivers to figure out exactly how the process of bypassing the Windows Network Stack and inserting two layers into the OS worked from a Killer NIC viewpoint. We have our ideas about what they did and a few interesting results but we did not publish them due to "some" speculation on our part. I cleaned up the front page and few other areas so it is not so "wordy" now. We will revisit the card once the FNapps are released next month.

 

[Gary Key]

Originally posted by: nweaver
Sorry to Gary, I didn't mean to sound too harsh...I'm here because I DO like AT and they have pretty decent methedologies versus some other sites. I also wouldn't expect you to do what I would do, but then, someone is paying me lots of money to perform that testing, and they arn't for you guys. Given what you had, you did a fantastic job, and admited that it wasn't the easiest of reviews. I didn't mean to sound like I thought you guys don't know what you are doing, as you guys are a step up from any other site imho.

I put your check in the mail but reduced it for not mentioning my first place finish in the 100 meter race in fifth grade. Thanks for the comments and I did not take any offense to the statements. I am always open to new ideas or trying something different around here.

 

[n0cmonkey]

I'm upgrading all of our servers to Killer NICs.

 

[nweaver]

Originally posted by: n0cmonkey
I'm upgrading all of our servers to Killer NICs.

I didn't think it would help, so I only upgraded all our client PC's, so we can frag that samba server faster