Wes Peters writes: | On Friday 07 March 2003 09:16, Doug Ambrisko wrote: | You did something truly bizarre. I've tested similar cards on many | machines ranging from K6-2 400MHz to P4 2.4GHz and the RealTek | performance has always been at or near the bottom of the heap. On the | slower processors, the overhead of aligning packets shows heavily, but it | can be noticed on any system.
Depends on what your systems is doing. We are PCI bus limited. | A number of the chips folded into the dc(4) driver are horrible and | perform right down there with the RealTek, but a few are fairly good. Agreed. We've tested the common 21143 and some clones. We also ran the tests a few times to the the dc(4) chip to get the TX delays right adjusted so they don't have FIFO under-runs since that adjustment kill performance. | The 3com 3c905s are generally quite good using the xl(4) driver, as are | the Intel EEPro's using fxp. I've read of others struggling with both | but never encountered this myself. I tend to be quite conservative about | throwing random versions of FreeBSD at systems, though, and many of those | complaints come from people at various points on -stable, rather than a | known release point. We've had good success with the fxp(4) chips except for a strange bug on an onboard motherboard version. There seems to be a bug in the eeprom setting for it that I patch in the fxp(4) driver. Unfortunately I'm guess at the correction since we haven't been able to get the definition of that register. Since Intel sets to that value and makes our bug go away we just do it. Makes me nervous though. | > So I'd say given a sufficiently fast CPU and memory the Realteks work | > pretty darn good. | | For a sufficient engine RPM, that escort will do 85 MPH in first gear, | too. ;^) Yep, and if you never have to turn a corner and the engine can handle it then it is okay. Our '87 Porsche 911 can't turn in a "normal" sense very fast due to cronic understeer. However, with a rear-weight bias it spins very fast. So to turn fast you just spin the car into the direction you need, gas it to stop the spin and off you go. Side benefit it that you don't need to brake as much going into a corner since when you are going side ways you are braking so you just factor that in. Is a Porsche 911 a performance car? In the right hands it is otherwise it's going backwards out of a corner which can be an interesting feeling! Sounds like a Corvair. | > To date we haven't had any trouble with them and we've shipped a bunch. | | Give me 1 second and I can easily bring any of your systems to their | knees, regardless of which cards you have installed. Everything is | relative. Were you watching the system load while performing your | testing? Was the cpu doing anything but routing? Is it required to for | your application? These and many others are all important questions, and | tend to have different answers for every application. For a desktop | workstation with undemanding network application requirements (email, web | browsing, occasional software updates) RealTek or any other card that | successfully attach to the network and correctly autonegotiate with your | hub (shudder) or switch is fine. Even a RealTek. ;^) Hmm, I thought I had said "benchmark in your environment". We have a closed box that is sort-of a router and a bridge. So your only inputs is really network traffic. That is what we tune the box for. So it would be interesting to see you kill it in 1s. Again our issue is PCI bus. Now that the P4 Serverworks chipset is out we finally have a machine that holds the current gig with crypto records by a lot (faster then a couple of P4 Xeon machines we have). With a 32bit/33Mhz we are pegged at the PCI chipset limits. One of the challenges of testing crypto (IPsec) stuff is having clients that can keep up. I'be been told there is a paper in the works for HW crypto performance based on this and other HW. So results of this should be published. Doug A. To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
