Date: Tue, 12 Dec 2006 11:00:23 +0900 From: Pyun YongHyeon <pyunyh@gmail.com> To: "Bruce M. Simpson" <bms@FreeBSD.org> Cc: freebsd-current@FreeBSD.org Subject: Re: Call for Marvell/SysKonnect Yukon II Gigabit Ethernet testers. Message-ID: <20061212020023.GA9698@cdnetworks.co.kr> In-Reply-To: <457DF011.9010701@FreeBSD.org> References: <20060926002916.GA5975@cdnetworks.co.kr> <ekikdg$aqu$1@sea.gmane.org> <20061129013052.GC71523@cdnetworks.co.kr> <457DF011.9010701@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, Dec 11, 2006 at 11:56:01PM +0000, Bruce M. Simpson wrote: > Hi, > > I successfully tested this driver under 7-CURRENT as of today on an ASUS > Vintage AH-1 based system. > > lspci has the following to say about it: > > 02:00.0 Ethernet controller: Marvell Technology Group Ltd. 88E8053 PCI-E > Gigabit Ethernet Controller (rev 19) > Subsystem: ASUSTeK Computer Inc. Marvell 88E8053 Gigabit > Ethernet controller PCIe (Asus) > Flags: bus master, fast devsel, latency 0, IRQ 18 > Memory at fe4fc000 (64-bit, non-prefetchable) > I/O ports at c800 > Expansion ROM at fe4c0000 [disabled] > Capabilities: [48] Power Management version 2 > Capabilities: [50] Vital Product Data > Capabilities: [5c] Message Signalled Interrupts: 64bit+ > Queue=0/1 Enable- > Capabilities: [e0] Express Legacy Endpoint IRQ 0 > > A few sample netperf runs between this system (AMD Athlon64 3000+) and > an Intel Dothan 1.8Ghz based Lenovo T43 with bge(4) interconnected via a > 3Com 5-port gigabit workgroup switch reveal the following. > > anglepoise# netperf -t UDP_STREAM -H 192.168.123.18 > UDP UNIDIRECTIONAL SEND TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to > 192.168.123.18 (192.168.123.18) port 0 AF_INET > Socket Message Elapsed Messages > Size Size Time Okay Errors Throughput > bytes bytes secs # # 10^6bits/sec > > 9216 9216 10.00 129443 1135712 954.30 > 42080 10.00 128739 949.11 > > > anglepoise# netperf -t UDP_RR -H 192.168.123.18 > UDP REQUEST/RESPONSE TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to > 192.168.123.18 (192.168.123.18) port 0 AF_INET > Local /Remote > Socket Size Request Resp. Elapsed Trans. > Send Recv Size Size Time Rate > bytes Bytes bytes bytes secs. per sec > > 9216 42080 1 1 10.00 2804.30 > 9216 42080 > > anglepoise# netperf -t TCP_STREAM -H 192.168.123.18 > TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 192.168.123.18 > (192.168.123.18) port 0 AF_INET > Recv Send Send > Socket Socket Message Elapsed > Size Size Size Time Throughput > bytes bytes bytes secs. 10^6bits/sec > > 65536 32768 32768 10.00 642.80 > > anglepoise# netperf -t TCP_RR -H 192.168.123.18 > TCP REQUEST/RESPONSE TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to > 192.168.123.18 (192.168.123.18) port 0 AF_INET > Local /Remote > Socket Size Request Resp. Elapsed Trans. > Send Recv Size Size Time Rate > bytes Bytes bytes bytes secs. per sec > > 32768 65536 1 1 10.00 2790.69 > 32768 65536 > > > anglepoise# netperf -t TCP_CRR -H 192.168.123.18 > TCP Connect/Request/Response TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET > to 192.168.123.18 (192.168.123.18) port 0 AF_INET > Local /Remote > Socket Size Request Resp. Elapsed Trans. > Send Recv Size Size Time Rate > bytes Bytes bytes bytes secs. per sec > > 32768 65536 1 1 10.00 1350.91 > 32768 65536 > > anglepoise# netperf -t TCP_MAERTS -H 192.168.123.18 > TCP MAERTS TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 192.168.123.18 > (192.168.123.18) port 0 AF_INET > Recv Send Send > Socket Socket Message Elapsed > Size Size Size Time Throughput > bytes bytes bytes secs. 10^6bits/sec > > 65536 32768 32768 10.00 407.51 > > And a statistically tighter test: > > anglepoise# netperf -t TCP_STREAM -I 99 -i 30,10 -H 192.168.123.18 > TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 192.168.123.18 > (192.168.123.18) port 0 AF_INET : +/-49.5% @ 99% conf. > Recv Send Send > Socket Socket Message Elapsed > Size Size Size Time Throughput > bytes bytes bytes secs. 10^6bits/sec > > 65536 32768 32768 10.00 644.98 > > Another test, this time between the same box and a dual PIII 933Mhz > running 6.1-RELEASE with an em(4) card and 66Mhz 64-bit PCI data path on > the same workgroup switch: > > anglepoise# netperf -t TCP_STREAM -I 99 -i 30,10 -H 192.168.123.6 > TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 192.168.123.6 > (192.168.123.6) port 0 AF_INET : +/-49.5% @ 99% conf. > Recv Send Send > Socket Socket Message Elapsed > Size Size Size Time Throughput > bytes bytes bytes secs. 10^6bits/sec > > 262144 32768 32768 10.00 161.17 > > ...probably harder on the old machine than anything, it is very unlikely > it could saturate at line rate. > > Thank you for all the excellent work on this driver, I hope this data is > useful. > Thanks for testing. The main focus for msk(4) was for getting working native driver. Performance was not heavily tested and highly likely to be lower than that of optimal performance. It seems that myk(4) has several workarounds for better performance but that magic code is hard to verify wihtout errata information from vendor. :-( Btw, I'll commit msk(4) in two days if there is no breakage report for e1000phy(4). > regards, > BMS -- Regards, Pyun YongHyeon
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20061212020023.GA9698>