Re: tcp bw in 2.6
A few notes to the discussion. I've seen one e1000 bug that ended up being a crappy AMD pre-opteron SMP chipset with a totally useless PCI bus implementation, which limited performance quite a bit-totally depending on what you plugged in and in which slot. 10e milk-and-bread-store 32/33 gige nics actually were better than server-class e1000's in those, but weren't that great either. That could well be my problem, this is a dual processor (not core) athlon (not opteron) tyan motherboard if I recall correctly. Check your interrupt rates for the interface. You shouldn't be getting anywhere near 1 interrupt/packet. If you are, something is badly wrong :). The acks (because I'm sending) are about 1.5 packets/interrupt. When this box is receiving it's moving about 3x ass much data and has a _lower_ (absolute, not per packet) interrupt load. -- --- Larry McVoylm at bitmover.com http://www.bitkeeper.com - To unsubscribe from this list: send the line unsubscribe netdev in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: tcp bw in 2.6
On Tue, Oct 02, 2007 at 06:52:54PM +0800, Herbert Xu wrote: One of my clients also has gigabit so I played around with just that one and it (itanium running hpux w/ broadcom gigabit) can push the load as well. One weird thing is that it is dependent on the direction the data is flowing. If the hp is sending then I get 46MB/sec, if linux is sending then I get 18MB/sec. Weird. Linux is debian, running First of all check the CPU load on both sides to see if either of them is saturating. If the CPU's fine then look at the tcpdump output to see if both receivers are using the same window settings. tcpdump is a good idea, take a look at this. The window starts out at 46 and never opens up in my test case, but in the rsh case it starts out the same but does open up. Ideas? 08:08:06.033305 IP hp-ia64.bitmover.com.49614 work-cluster.bitmover.com.31235: S 2756874880:2756874880(0) win 32768 mss 1460,wscale 0,nop 08:08:06.05 IP work-cluster.bitmover.com.31235 hp-ia64.bitmover.com.49614: S 3360532803:3360532803(0) ack 2756874881 win 5840 mss 1460,nop,wscale 7 08:08:06.047924 IP hp-ia64.bitmover.com.49614 work-cluster.bitmover.com.31235: . ack 1 win 32768 08:08:06.048218 IP work-cluster.bitmover.com.31235 hp-ia64.bitmover.com.49614: . 1:2921(2920) ack 1 win 46 08:08:06.048426 IP hp-ia64.bitmover.com.49614 work-cluster.bitmover.com.31235: . ack 1461 win 32768 08:08:06.048446 IP work-cluster.bitmover.com.31235 hp-ia64.bitmover.com.49614: . 2921:5841(2920) ack 1 win 46 08:08:06.048673 IP hp-ia64.bitmover.com.49614 work-cluster.bitmover.com.31235: . ack 4381 win 32768 08:08:06.048684 IP work-cluster.bitmover.com.31235 hp-ia64.bitmover.com.49614: . 5841:10221(4380) ack 1 win 46 08:08:06.049047 IP hp-ia64.bitmover.com.49614 work-cluster.bitmover.com.31235: . ack 8761 win 32768 08:08:06.049057 IP work-cluster.bitmover.com.31235 hp-ia64.bitmover.com.49614: . 10221:16061(5840) ack 1 win 46 08:08:06.049422 IP hp-ia64.bitmover.com.49614 work-cluster.bitmover.com.31235: . ack 14601 win 32768 08:08:06.049429 IP work-cluster.bitmover.com.31235 hp-ia64.bitmover.com.49614: P 16061:18981(2920) ack 1 win 46 08:08:06.049462 IP work-cluster.bitmover.com.31235 hp-ia64.bitmover.com.49614: . 18981:20441(1460) ack 1 win 46 08:08:06.049484 IP work-cluster.bitmover.com.31235 hp-ia64.bitmover.com.49614: . 20441:23361(2920) ack 1 win 46 08:08:06.049924 IP hp-ia64.bitmover.com.49614 work-cluster.bitmover.com.31235: . ack 21901 win 32768 08:08:06.049943 IP work-cluster.bitmover.com.31235 hp-ia64.bitmover.com.49614: . 23361:32121(8760) ack 1 win 46 08:08:06.050549 IP hp-ia64.bitmover.com.49614 work-cluster.bitmover.com.31235: . ack 30661 win 32768 08:08:06.050559 IP work-cluster.bitmover.com.31235 hp-ia64.bitmover.com.49614: P 32121:39421(7300) ack 1 win 46 08:08:06.050592 IP work-cluster.bitmover.com.31235 hp-ia64.bitmover.com.49614: . 39421:40881(1460) ack 1 win 46 08:08:06.050614 IP work-cluster.bitmover.com.31235 hp-ia64.bitmover.com.49614: . 40881:42341(1460) ack 1 win 46 08:08:06.051170 IP hp-ia64.bitmover.com.49614 work-cluster.bitmover.com.31235: . ack 40881 win 32768 08:08:06.051188 IP work-cluster.bitmover.com.31235 hp-ia64.bitmover.com.49614: . 42341:54021(11680) ack 1 win 46 08:08:06.051923 IP hp-ia64.bitmover.com.49614 work-cluster.bitmover.com.31235: . ack 52561 win 32768 08:08:06.051932 IP work-cluster.bitmover.com.31235 hp-ia64.bitmover.com.49614: P 54021:58401(4380) ack 1 win 46 08:08:06.051942 IP work-cluster.bitmover.com.31235 hp-ia64.bitmover.com.49614: . 58401:67161(8760) ack 1 win 46 08:08:06.052671 IP hp-ia64.bitmover.com.49614 work-cluster.bitmover.com.31235: . ack 65701 win 32768 08:08:06.052680 IP work-cluster.bitmover.com.31235 hp-ia64.bitmover.com.49614: P 67161:74461(7300) ack 1 win 46 08:08:06.052719 IP work-cluster.bitmover.com.31235 hp-ia64.bitmover.com.49614: . 74461:77381(2920) ack 1 win 46 08:08:06.052752 IP work-cluster.bitmover.com.31235 hp-ia64.bitmover.com.49614: . 77381:81761(4380) ack 1 win 46 08:08:06.053549 IP hp-ia64.bitmover.com.49614 work-cluster.bitmover.com.31235: . ack 80301 win 32768 08:08:06.053566 IP work-cluster.bitmover.com.31235 hp-ia64.bitmover.com.49614: P 81761:97821(16060) ack 1 win 46 08:08:06.054423 IP hp-ia64.bitmover.com.49614 work-cluster.bitmover.com.31235: . ack 96361 win 32768 08:08:06.054433 IP work-cluster.bitmover.com.31235 hp-ia64.bitmover.com.49614: P 97821:113881(16060) ack 1 win 46 08:08:06.054476 IP work-cluster.bitmover.com.31235 hp-ia64.bitmover.com.49614: . 113881:115341(1460) ack 1 win 46 08:08:06.055422 IP hp-ia64.bitmover.com.49614 work-cluster.bitmover.com.31235: . ack 113881 win 32768 08:08:06.055438 IP work-cluster.bitmover.com.31235 hp-ia64.bitmover.com.49614: P 115341:131401(16060) ack 1 win 46 08:08:06.056421 IP hp-ia64.bitmover.com.49614 work-cluster.bitmover.com.31235: . ack 131401 win 32768 08:08:06.056432 IP work-cluster.bitmover.com.31235
Re: tcp bw in 2.6
Interesting data point. My test case is like this: server bind listen while (newsock = accept...) transfer() client connect transfer If the server side is the source of the data, i.e, it's transfer is a write loop, then I get the bad behaviour. If I switch them so the data flows in the other direction, then it works, I go from about 14K pkt/sec to 43K pkt/sec. Can anyone else reproduce this? I can extract the test case from lmbench so it is standalone but I suspect that any test case will do it. I'll try with the one that John sent. Yup, s/read/write/ and s/write/read/ in his two files at the appropriate places and I get exactly the same behaviour. So is this a bug or intentional? -- --- Larry McVoylm at bitmover.com http://www.bitkeeper.com - To unsubscribe from this list: send the line unsubscribe netdev in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: tcp bw in 2.6
If the server side is the source of the data, i.e, it's transfer is a write loop, then I get the bad behaviour. ... So is this a bug or intentional? For whatever it is worth, I believed that we used to get better performance from the same hardware. My guess is that it changed somewhere between 2.6.15-1-k7 and 2.6.18-5-k7. -- --- Larry McVoylm at bitmover.com http://www.bitkeeper.com - To unsubscribe from this list: send the line unsubscribe netdev in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: tcp bw in 2.6
Isn't this something so straightforward that you would have tests for it? This is the basic FTP server loop, doesn't someone have a big machine with 10gig cards and test that sending/recving data doesn't regress? Sounds like a bug to me, modulo the above caveat of making sure that it's not some hw/driver/switch kind of difference. Pretty unlikely given that we've changed the switch, the card works fine in the other direction, and I'm 95% sure that we used to get better perf before we switched to a more recent kernel. I'll try and find some other gig ether cards and try them. -- --- Larry McVoylm at bitmover.com http://www.bitkeeper.com - To unsubscribe from this list: send the line unsubscribe netdev in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: tcp bw in 2.6
On Tue, Oct 02, 2007 at 09:47:26AM -0700, Stephen Hemminger wrote: On Tue, 2 Oct 2007 09:25:34 -0700 [EMAIL PROTECTED] (Larry McVoy) wrote: If the server side is the source of the data, i.e, it's transfer is a write loop, then I get the bad behaviour. ... So is this a bug or intentional? For whatever it is worth, I believed that we used to get better performance from the same hardware. My guess is that it changed somewhere between 2.6.15-1-k7 and 2.6.18-5-k7. For the period from 2.6.15 to 2.6.18, the kernel by default enabled TCP Appropriate Byte Counting. This caused bad performance on applications that did small writes. It's doing 1MB writes. Is there a sockopt to turn that off? Or /proc or something? -- --- Larry McVoylm at bitmover.com http://www.bitkeeper.com - To unsubscribe from this list: send the line unsubscribe netdev in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: tcp bw in 2.6
I have a more complex configuration application, but I don't see this problem in my testing. Using e1000 nics and modern hardware I'm using a similar setup, what kernel are you using? I am purposefully setting the socket send/rx buffers, as well has twiddling with the tcp and netdev related tunables. Ben sent those to me, see below, they didn't make any difference. I tried diddling the socket send/recv buffers to 10MB, that didn't help. The defaults didn't help. 1MB didn't help and 64K didn't help. -- --- Larry McVoylm at bitmover.com http://www.bitkeeper.com - To unsubscribe from this list: send the line unsubscribe netdev in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: tcp bw in 2.6
On Tue, Oct 02, 2007 at 10:14:11AM -0700, Rick Jones wrote: Larry McVoy wrote: A short summary is can someone please post a test program that sources and sinks data at the wire speed? because apparently I'm too old and clueless to write such a thing. WRT the different speeds in each direction talking with HP-UX, perhaps there is an interaction between the Linux TCP stack (TSO perhaps) and HP-UX's ACK avoidance heuristics. If that is the case, tweaking tcp_deferred_ack_max with ndd on the HP-UX system might yield different results. I doubt it because I see the same sort of behaviour when I have a group of Linux clients talking to the server. The HP box is in the mix simply because it has a gigabit card and that makes driving the load simpler. But if I do several loads from 100Mbit clients I get the same packet throughput. WRT the small program making a setsockopt(SO_*BUF) call going slower than the rsh, does rsh make the setsockopt() call, or does it bend itself to the will of the linux stack's autotuning? What happens if your small program does not make setsockopt(SO_*BUF) calls? I haven't tracked down if rsh does that but I've tried doing it with values of default, 64K, 1MB, and 10MB with no difference. *) depending on the quantity of CPU around, and the type of test one is These are fast CPUs and they are running at 93% idle while running the test. -- --- Larry McVoylm at bitmover.com http://www.bitkeeper.com - To unsubscribe from this list: send the line unsubscribe netdev in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: tcp bw in 2.6
I'm currently on 2.6.20, and have also tried 10gbe nics on 2.6.23 with My guess is that it is a bug in the debian 2.6.18 kernel. Have you tried something like ttcp, iperf, or even regular ftp? Yeah, I've factored out the code since BitKeeper, my test program, and John's test program all exhibit the same behaviour. Also switched switches. Checked your nics to make sure they have no errors and are negotiated to full duplex? Yup and yup. -- --- Larry McVoylm at bitmover.com http://www.bitkeeper.com - To unsubscribe from this list: send the line unsubscribe netdev in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: tcp bw in 2.6
Make sure you don't have slab debugging turned on. It kills performance. It's a stock debian kernel, so unless they turn it on it's off. -- --- Larry McVoylm at bitmover.com http://www.bitkeeper.com - To unsubscribe from this list: send the line unsubscribe netdev in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: tcp bw in 2.6
On Tue, Oct 02, 2007 at 11:01:47AM -0700, Rick Jones wrote: has anyone already asked whether link-layer flow-control is enabled? I doubt it, the same test works fine in one direction and poorly in the other. Wouldn't the flow control squelch either way? -- --- Larry McVoylm at bitmover.com http://www.bitkeeper.com - To unsubscribe from this list: send the line unsubscribe netdev in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: tcp bw in 2.6
Looks like you have TSO enabled. Does it behave differently if it's disabled? It cranks the interrupts/sec up to 8K instead of 5K. No difference in performance other than that. I think Rick Jones is on to something with the HP ack avoidance. I sincerely doubt it. I'm only using the HP box because it has gigabit so it's a single connection. I can produce almost identical results by doing the same sorts of tests with several linux clients. One direction goes fast and the other goes slow. 3x performance difference depending on the direction of data flow: # Server is receiving, goes fast $ for i in 22 24 25 26; do rsh -n glibc$i dd if=/dev/zero|dd of=/dev/null done load free cach swap pgin pgou dk0 dk1 dk2 dk3 ipkt opkt int ctx usr sys idl 0.98 0000 00 0 0 0 30K 15K 8.1K 68K 12 66 22 0.98 0000 00 0 0 0 29K 15K 8.2K 67K 11 64 25 0.98 0000 00 0 0 0 29K 15K 8.2K 67K 12 66 22 # Server is sending, goes slow $ for i in 22 24 25 26; do dd if=/dev/zero|rsh glibc$i dd of=/dev/null done load free cach swap pgin pgou dk0 dk1 dk2 dk3 ipkt opkt int ctx usr sys idl 1.06 0000 00 0 0 0 5.0K 10K 4.4K 8.4K 21 17 62 0.97 0000 00 0 0 0 5.1K 10K 4.4K 8.9K 2 15 83 0.97 0000 00 0 0 0 5.0K 10K 4.4K 8.6K 21 26 53 $ for i in 22 24 25 26; do rsh glibc$i cat /etc/motd; done | grep Welcome Welcome to redhat71.bitmover.com, a 2Ghz Athlon running Red Hat 7.1. Welcome to glibc24.bitmover.com, a 1.2Ghz Athlon running SUSE 10.1. Welcome to glibc25.bitmover.com, a 2Ghz Athlon running Fedora Core 6 Welcome to glibc26.bitmover.com, a 2Ghz Athlon running Fedora Core 7 $ for i in 22 24 25 26; do rsh glibc$i uname -r; done 2.4.2-2 2.6.16.13-4-default 2.6.18-1.2798.fc6 2.6.22.4-65.fc7 No HP in the mix. It's got nothing to do with hp, nor to do with rsh, it has everything to do with the direction the data is flowing. -- --- Larry McVoylm at bitmover.com http://www.bitkeeper.com - To unsubscribe from this list: send the line unsubscribe netdev in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: tcp bw in 2.6
More data, we've conclusively eliminated the card / cpu from the mix. We've got 2 ia64 boxes with e1000 interfaces. One box is running linux 2.6.12 and the other is running hpux 11. I made sure the linux one was running at gigabit and reran the tests from the linux/ia64 = hp/ia64. Same results, when linux sends it is slow, when it receives it is fast. And note carefully: we've removed hpux from the equation, we can do the same tests from linux to multiple linux clients and see the same thing, sending from the server is slow, receiving on the server is fast. -- --- Larry McVoylm at bitmover.com http://www.bitkeeper.com - To unsubscribe from this list: send the line unsubscribe netdev in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: tcp bw in 2.6
I think I'm still missing some basic data here (probably because this thread did not originate on netdev). Let me try to nail down some of the basics. You have a linux ia64 box (running 2.6.12 or 2.6.18?) that sends slowly, and receives faster, but not quite a 1 Gbps? And this is true regardless of which peer it sends or receives from? And the behavior is different depending on which kernel? How, and which kernel versions? Do you have other hardware running the same kernel that behaves the same or differently? just got off the phone with Linus and he thinks it is the side that does the accept is the problem side, i.e., if you are the server, you do the accept, and you send the data, you'll go slow. But as I'm writing this I realize he's wrong, because it is the combination of accept send. accept recv goes fast. A trivial way to see the problem is to take two linux boxes, on each apt-get install rsh-client rsh-server set up your .rhosts, and then do dd if=/dev/zero count=10 | rsh OTHER_BOX dd of=/dev/null rsh OTHER_BOX dd if=/dev/zero count=10 | dd of=/dev/null See if you get balanced results. For me, I get 45MB/sec one way, and 15-19MB/sec the other way. I've tried the same test linux - linux and linux - hpux. Same results. The test setup I have is work: 2ghz x 2 Athlons, e1000, 2.6.18 ia64: 900mhz Itanium, e1000, 2.6.12 hp-ia64:900mhz Itanium, e1000, hpux 11 glibc*: 1-2ghz athlons running various linux releases all connected through a netgear 724T 10/100/1000 switch (a linksys showed identical results). I tested work - hp-ia64 work - ia64 ia64 - hp-ia64 and in all cases, one direction worked fast and the other didn't. It would be good if people tried the same simple test. You have to use rsh, ssh will slow things down way too much. Alternatively, take your favorite test programs, such as John's, and make a second pair that reverses the direction the data is sent. So one pair is server sends, the other is server receives, try both. That's where we started, BitKeeper, my stripped down test, and John's test all exhibit the same behavior. And the rsh test is just a really simple way to demonstrate it. Wayne, Linus asked for tcp dumps from just one side, with the first 100 packets and then wait 10 seconds or so for the window to open up, and then a snap shot of the another 100 packets. Do that for both directions and send them to the list. Can you do that? I want to get lunch, I'm starving. -- --- Larry McVoylm at bitmover.com http://www.bitkeeper.com - To unsubscribe from this list: send the line unsubscribe netdev in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: tcp bw in 2.6
We fixed a lot of bugs in TSO last year. It would be really great to see numbers with a more recent kernel than 2.6.18 More data, sky2 works fine (really really fine, like 79MB/sec) between Linux dylan.bitmover.com 2.6.18.1 #5 SMP Mon Oct 23 17:36:00 PDT 2006 i686 Linux steele 2.6.20-16-generic #2 SMP Sun Sep 23 18:31:23 UTC 2007 x86_64 So this is looking like a e1000 bug. I'll try to upgrade the kernel on the ia64 box and see what happens. -- --- Larry McVoylm at bitmover.com http://www.bitkeeper.com - To unsubscribe from this list: send the line unsubscribe netdev in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: tcp bw in 2.6
On Tue, Oct 02, 2007 at 02:16:56PM -0700, David Miller wrote: We absolutely depend upon people like you to report when there are anomalies like this. It's the only thing that scales. Well cool, finally doing something useful :) Is this issue no test setup? Because this does seem like something we'd want to have work well. FWIW I have a t1000 Niagara box and an Ultra45 going through a netgear gigabit switch. I'm getting 85MB/sec in one direction and 10MB/sec in the other (using bw_tcp from lmbench3). Note that bw_tcp mucks with SND/RCVBUF. It probably shouldn't, it's been 12 years since that code went in there and I dunno if it is still needed. Both are using identical broadcom tigon3 gigabit chips and identical current kernels so that is a truly strange result. I'll investigate, it may be the same thing you're seeing. Wow, sounds very similar. In my case I was seeing pretty close to 3x consistently. You're more like 8x, but I was all e1000 not broadcom. And note that sky2 doesn't have this problem. Does the broadcom do TSO? And sky2 not? I noticed a much higher CPU load for sky2. -- --- Larry McVoylm at bitmover.com http://www.bitkeeper.com - To unsubscribe from this list: send the line unsubscribe netdev in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: tcp bw in 2.6
On Tue, Oct 02, 2007 at 03:32:16PM -0700, David Miller wrote: I'm starting to have a theory about what the bad case might be. A strong sender going to an even stronger receiver which can pull out packets into the process as fast as they arrive. This might be part of what keeps the receive window from growing. I can back you up on that. When I straced the receiving side that goes slowly, all the reads were short, like 1-2K. The way that works the reads were a lot larger as I recall. -- --- Larry McVoylm at bitmover.com http://www.bitkeeper.com - To unsubscribe from this list: send the line unsubscribe netdev in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: tcp bw in 2.6
On Sat, Sep 29, 2007 at 11:02:32AM -0700, Linus Torvalds wrote: On Sat, 29 Sep 2007, Larry McVoy wrote: I haven't kept up on switch technology but in the past they were much better than you are thinking. The Kalpana switch that I had modified to support vlans (invented by yours truly), did not store and forward, it was cut through and could handle any load that was theoretically possible within about 1%. Hey, you may well be right. Maybe my assumptions about cutting corners are just cynical and pessimistic. So I got a netgear switch and it works fine. But my tests are busted. Catching netdev up, I'm trying to optimize traffic to a server that has a gbit interface; I moved to a 24 port netgear that is all 10/100/1000 and I have a pile of clients to act as load generators. I can do this on each of the clients dd if=/dev/zero bs=1024000 | rsh work dd of=/dev/null and that cranks up to about 47K packets/second which is about 70MB/sec. One of my clients also has gigabit so I played around with just that one and it (itanium running hpux w/ broadcom gigabit) can push the load as well. One weird thing is that it is dependent on the direction the data is flowing. If the hp is sending then I get 46MB/sec, if linux is sending then I get 18MB/sec. Weird. Linux is debian, running Linux work 2.6.18-5-k7 #1 SMP Thu Aug 30 02:52:31 UTC 2007 i686 and dual e1000 cards: e1000: eth0: e1000_probe: Intel(R) PRO/1000 Network Connection e1000: eth1: e1000_probe: Intel(R) PRO/1000 Network Connection I wrote a tiny little program to try and emulate this and I can't get it to do as well. I've tracked it down, I think, to the read side. The server sources, the client sinks, the server looks like: 11689 accept(3, {sa_family=AF_INET, sin_port=htons(49376), sin_addr=inet_addr(10.3.1.38)}, [16]) = 4 11689 setsockopt(4, SOL_SOCKET, SO_RCVBUF, [1048576], 4) = 0 11689 setsockopt(4, SOL_SOCKET, SO_SNDBUF, [1048576], 4) = 0 11689 clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0xb7ddf708) = 11694 11689 close(4) = 0 11689 accept(3, unfinished ... 11694 write(4, \0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0..., 1048576) = 1048576 11694 write(4, \0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0..., 1048576) = 1048576 11694 write(4, \0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0..., 1048576) = 1048576 11694 write(4, \0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0..., 1048576) = 1048576 ... but the client looks like connect(3, {sa_family=AF_INET, sin_port=htons(31235), sin_addr=inet_addr(10.3.9.1)}, 16) = 0 read(3, \0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0..., 1048576) = 2896 read(3, \0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0..., 1048576) = 1448 read(3, \0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0..., 1048576) = 2896 read(3, \0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0..., 1048576) = 2896 read(3, \0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0..., 1048576) = 2896 read(3, \0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0..., 1048576) = 2896 read(3, \0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0..., 1048576) = 2896 read(3, \0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0..., 1048576) = 2896 read(3, \0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0..., 1048576) = 2896 read(3, \0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0..., 1048576) = 1448 read(3, \0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0..., 1048576) = 2896 which I suspect may be the problem. I played around with SO_RCVBUF/SO_SNDBUF and that didn't help. So any ideas why a simple dd piped through rsh is kicking my ass? It must be something simple but my test program is tiny and does nothing weird that I can see. -- --- Larry McVoylm at bitmover.com http://www.bitkeeper.com - To unsubscribe from this list: send the line unsubscribe netdev in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: tcp bw in 2.6
On Mon, Oct 01, 2007 at 07:14:37PM -0700, Linus Torvalds wrote: On Mon, 1 Oct 2007, Larry McVoy wrote: but the client looks like connect(3, {sa_family=AF_INET, sin_port=htons(31235), sin_addr=inet_addr(10.3.9.1)}, 16) = 0 read(3, \0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0..., 1048576) = 2896 read(3, \0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0..., 1048576) = 1448 read(3, \0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0..., 1048576) = 2896 .. This is exactly what I'd expect if the machine is *not* under excessive load. That's fine, but why is it that my trivial program can't do as well as dd | rsh dd? A short summary is can someone please post a test program that sources and sinks data at the wire speed? because apparently I'm too old and clueless to write such a thing. -- --- Larry McVoylm at bitmover.com http://www.bitkeeper.com - To unsubscribe from this list: send the line unsubscribe netdev in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: tcp bw in 2.6
On Mon, Oct 01, 2007 at 08:50:50PM -0700, David Miller wrote: From: [EMAIL PROTECTED] (Larry McVoy) Date: Mon, 1 Oct 2007 19:20:59 -0700 A short summary is can someone please post a test program that sources and sinks data at the wire speed? because apparently I'm too old and clueless to write such a thing. You're not showing us your test program so there is no way we can help you out. Attached. Drop it into an lmbench tree and build it. My initial inclination, even without that critical information, is to ask whether you are setting any socket options in way? The only one I was playing with was SO_RCVBUF/SO_SNDBUF and I tried disabling that and I tried playing with the read/write size. Didn't help. In particular, SO_RCVLOWAT can have a large effect here, if you're setting it to something, that would explain why dd is doing better. A lot of people link to helper libraries with interfaces to setup sockets with all sorts of socket option settings by default, try not using such things if possible. Agreed. That was my first thought as well, I must have been doing something that messed up the defaults. But you did get the strace output, there wasn't anything weird there. You also shouldn't dork at all with the receive and send buffer sizes. They are adjusted dynamically by the kernel as the window grows. But if you set them to specific values, this dynamic logic is turned off. Yeah, dorking with those is left over from the bad old days of '95 when lmbench was first shipped. But I turned that all off and no difference. So feel free to show me where I'm an idiot in the code, but if you can't, then what would rock would be a little send.c / recv.c that demonstrated filling the pipe. -- --- Larry McVoylm at bitmover.com http://www.bitkeeper.com /* * bytes_tcp.c - simple TCP bandwidth source/sink * * server usage: bytes_tcp -s * client usage: bytes_tcp hostname [msgsize] * * Copyright (c) 1994 Larry McVoy. * Copyright (c) 2002 Carl Staelin. Distributed under the FSF GPL with * additional restriction that results may published only if * (1) the benchmark is unmodified, and * (2) the version in the sccsid below is included in the report. * Support for this development by Sun Microsystems is gratefully acknowledged. */ char *id = $Id$\n; #include bench.h #define XFER (1024*1024) int server_main(int ac, char **av); int client_main(int ac, char **av); void source(int data); void transfer(int get, int server, char *buf) { int c; while ((get 0) (c = read(server, buf, XFER)) 0) { get -= c; } if (c 0) { perror(bytes_tcp: transfer: read failed); exit(4); } } /* ARGSUSED */ int client_main(int ac, char **av) { int server; int get = 256 20; char buf[XFER]; char* usage = usage: %s -remotehost OR %s remotehost [msgsize]\n; if (ac != 2 ac != 3) { (void)fprintf(stderr, usage, av[0], av[0]); exit(0); } if (ac == 3) get = bytes(av[2]); server = tcp_connect(av[1], TCP_DATA+1, SOCKOPT_READ|SOCKOPT_REUSE); if (server 0) { perror(bytes_tcp: could not open socket to server); exit(2); } transfer(get, server, buf); close(server); exit(0); /*NOTREACHED*/ } void child() { wait(0); signal(SIGCHLD, child); } /* ARGSUSED */ int server_main(int ac, char **av) { int data, newdata; signal(SIGCHLD, child); data = tcp_server(TCP_DATA+1, SOCKOPT_READ|SOCKOPT_WRITE|SOCKOPT_REUSE); for ( ;; ) { newdata = tcp_accept(data, SOCKOPT_WRITE|SOCKOPT_READ); switch (fork()) { case -1: perror(fork); break; case 0: source(newdata); exit(0); default: close(newdata); break; } } } void source(int data) { char buf[XFER]; while (write(data, buf, sizeof(buf)) 0); } int main(int ac, char **av) { char* usage = Usage: %s -s OR %s -serverhost OR %s serverhost [msgsize]\n; if (ac 2 || 3 ac) { fprintf(stderr, usage, av[0], av[0], av[0]); exit(1); } if (ac == 2 !strcmp(av[1], -s)) { if (fork() == 0) server_main(ac, av); exit(0); } else { client_main(ac, av); } return(0); } /* * tcp_lib.c - routines for managing TCP connections. * * Positive port/program numbers are RPC ports, negative ones are TCP ports. * * Copyright (c) 1994-1996 Larry McVoy. */ #define _LIB /* bench.h needs this */ #include bench.h /* * Get a TCP socket, bind it, figure out the port, * and advertise the port as program prog. * * XXX - it would be nice if you could advertise ascii strings. */ int tcp_server(int prog, int rdwr) { int sock; struct sockaddr_in s; #ifdef LIBTCP_VERBOSE fprintf(stderr, tcp_server(%u, %u)\n, prog, rdwr); #endif if ((sock = socket(AF_INET, SOCK_STREAM, IPPROTO_TCP)) 0) { perror(socket); exit(1); } sock_optimize(sock, rdwr); bzero((void*)s, sizeof(s)); s.sin_family = AF_INET; if (prog 0) { s.sin_port = htons(-prog); } if (bind(sock, (struct sockaddr*)s, sizeof(s)) 0) { perror(bind); exit(2); } if (listen(sock, 100) 0) { perror(listen
on a different note
I do have a pretty nice cluster of linux boxes if you need lmbench results or something like that. Linux and the rest of the unix stuff for whatever that is worth... work ~/LMbench2/bin ls -1 alpha-glibc22-linux # need to upgrade to debian 4 hppa-glibc23-linux hppa-hpux11 ia64-glibc23-linux ia64-hpux11 mips-glibc23-linux mips-irix powerpc-aix powerpc-glibc23-linux powerpc-macosx sparc-glibc23-linux sparc-solaris x86-darwin8.10.1 x86-freebsd2 x86-freebsd3 x86-freebsd4 x86-freebsd5 x86-freebsd6 x86-glibc20-linux # redhat 5.2 x86-glibc21-linux # redhat 6.1 x86-glibc22-linux # redhat 7.2 x86-glibc23-linux # redhat 9 x86-glibc24-linux # debian from here on down x86-glibc25-linux x86-glibc26-linux x86-netbsd x86-openbsd x86-sco3.2v5.0.7# ha! x86-solaris x86_64-glibc23-linux -- --- Larry McVoylm at bitmover.com http://www.bitkeeper.com - To unsubscribe from this list: send the line unsubscribe netdev in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html