From owner-freebsd-smp Sun Oct 19 07:28:21 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id HAA08371 for smp-outgoing; Sun, 19 Oct 1997 07:28:21 -0700 (PDT) (envelope-from owner-freebsd-smp) Received: from math.berkeley.edu (math.Berkeley.EDU [128.32.183.94]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id HAA08355 for ; Sun, 19 Oct 1997 07:28:18 -0700 (PDT) (envelope-from dan@math.berkeley.edu) Received: (from dan@localhost) by math.berkeley.edu (8.8.7/8.8.7) id HAA13450; Sun, 19 Oct 1997 07:28:11 -0700 (PDT) Date: Sun, 19 Oct 1997 07:28:11 -0700 (PDT) From: dan@math.berkeley.edu (Dan Strick) Message-Id: <199710191428.HAA13450@math.berkeley.edu> To: freebsd-smp@freebsd.org Subject: multiple I/O APICs Cc: dan@math.berkeley.edu Sender: owner-freebsd-smp@freebsd.org X-Loop: FreeBSD.org Precedence: bulk The io_apic_setup() routine in mpapic.c doesn't like my motherboard. It says "panic io_apic_setup: apic #1". The problem seems to be that the motherboard has more than one I/O APIC and this is not supported. Please forgive this possibly stupid question, but I am not yet familiar with the Intel 1.4 MP spec. Can I configure the kernel to just ignore the second I/O APIC that it doesn't know how to use? If I set NAPIC to 1 in the config file, the kernel panics when it starts and tells me that I must reconfigure it with NAPIC set to 2. I am not certain about the motherboard's correct name or model number. I thought the machine was an Intel "AP450GX MP Server System", but none of the manuals or papers that came with it, including the rather bulky installation manual, carry any model name or number. The mptable program output follows. Dan Strick dan@math.berkeley.edu =============================================================================== MPTable, version 2.0.15 ------------------------------------------------------------------------------- MP Floating Pointer Structure: location: BIOS physical address: 0x000f75a0 signature: '_MP_' length: 16 bytes version: 1.4 checksum: 0x6c mode: Virtual Wire ------------------------------------------------------------------------------- MP Config Table Header: physical address: 0x000f75b0 signature: 'PCMP' base table length: 508 version: 1.4 checksum: 0xe1 OEM ID: 'INTEL ' Product ID: 'ALDER ' OEM table pointer: 0x00000000 OEM table size: 0 entry count: 52 local APIC address: 0xfec08000 extended table length: 240 extended table checksum: 201 ------------------------------------------------------------------------------- MP Config Base Table Entries: -- Processors: APIC ID Version State Family Model Step Flags 0 0x11 BSP, usable 6 1 9 0xfbff 4 0x11 AP, usable 6 1 9 0xfbff 1 0x11 AP, usable 6 1 9 0xfbff 2 0x11 AP, usable 6 1 9 0xfbff -- Bus: Bus ID Type 0 PCI 1 PCI 18 EISA -- I/O APICs: APIC ID Version State Address 14 0x11 usable 0xfec00000 13 0x11 usable 0xfec01000 -- I/O Ints: Type Polarity Trigger Bus ID IRQ APIC ID PIN# INT active-hi edge 18 1 14 1 INT active-hi edge 18 0 14 2 INT active-hi edge 18 3 14 3 INT active-hi edge 18 4 14 4 INT active-hi edge 18 5 14 5 INT active-hi edge 18 6 14 6 INT active-hi edge 18 7 14 7 INT active-hi edge 18 8 14 8 INT active-hi edge 18 9 14 9 INT active-hi level 18 10 14 10 INT active-hi level 18 11 14 11 INT active-hi edge 18 12 14 12 INT active-hi edge 18 13 14 13 INT active-hi edge 18 14 14 14 INT active-hi edge 18 15 14 15 INT active-hi level 0 11:A 13 0 INT active-hi level 0 12:A 13 1 INT active-hi level 0 13:A 13 2 INT active-hi level 1 10:A 13 3 INT active-hi level 1 13:A 13 4 INT active-hi level 0 11:B 13 5 INT active-hi level 0 12:C 13 5 INT active-hi level 0 13:D 13 5 INT active-hi level 1 10:B 13 5 INT active-hi level 1 13:C 13 5 INT active-hi level 1 14:D 13 5 INT active-hi level 0 11:C 13 6 INT active-hi level 0 12:D 13 6 INT active-hi level 0 13:B 13 6 INT active-hi level 1 14:B 13 6 INT active-hi level 1 10:C 13 6 INT active-hi level 1 13:D 13 6 INT active-hi level 0 12:B 13 7 INT active-hi level 0 13:C 13 7 INT active-hi level 0 11:D 13 7 INT active-hi level 1 13:B 13 7 INT active-hi level 1 14:C 13 7 INT active-hi level 1 10:D 13 7 INT active-hi level 1 11:A 13 12 INT active-hi level 1 12:A 13 13 INT active-hi level 1 14:A 13 15 -- Local Ints: Type Polarity Trigger Bus ID IRQ APIC ID PIN# ExtINT active-hi edge 18 0 255 0 NMI active-hi edge 0 0:A 255 1 ------------------------------------------------------------------------------- MP Config Extended Table Entries: -- bus ID: 0 address type: memory address address base: 0xa0000 address range: 0x20000 -- bus ID: 0 address type: prefetch address address base: 0xfc500000 address range: 0x100000 -- bus ID: 0 address type: memory address address base: 0xfc600000 address range: 0x2500000 -- bus ID: 0 address type: I/O address address base: 0xf000 address range: 0x1000 -- bus ID: 1 address type: prefetch address address base: 0xfc300000 address range: 0x100000 -- bus ID: 1 address type: memory address address base: 0xfc400000 address range: 0x100000 -- bus ID: 1 address type: I/O address address base: 0xe000 address range: 0x1000 -- bus ID: 0 address type: memory address address base: 0x10000000 address range: 0xec300000 -- bus ID: 0 address type: memory address address base: 0xfeb00000 address range: 0x1500000 -- bus ID: 0 address type: I/O address address base: 0x0 address range: 0xe000 -- bus ID: 18 bus info: 0x01 parent bus ID: 0-- bus ID: 0 address modifier: add predefined range: 0x00000000-- bus ID: 0 address modifier: add predefined range: 0x00000001-- bus ID: 1 address modifier: subtract predefined range: 0x00000000-- bus ID: 1 address modifier: subtract predefined range: 0x00000001 ------------------------------------------------------------------------------- # SMP kernel config file options: # Required: options SMP # Symmetric MultiProcessor Kernel options APIC_IO # Symmetric (APIC) I/O # Optional (built-in defaults will work in most cases): #options NCPU=4 # number of CPUs #options NBUS=3 # number of busses #options NAPIC=2 # number of IO APICs #options NINTR=41 # number of INTs =============================================================================== From owner-freebsd-smp Sun Oct 19 08:42:05 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id IAA14536 for smp-outgoing; Sun, 19 Oct 1997 08:42:05 -0700 (PDT) (envelope-from owner-freebsd-smp) Received: from Ilsa.StevesCafe.com (Ilsa.StevesCafe.com [205.168.119.129]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id IAA14528 for ; Sun, 19 Oct 1997 08:41:59 -0700 (PDT) (envelope-from fbsd@Ilsa.StevesCafe.com) Received: from Ilsa.StevesCafe.com (localhost [127.0.0.1]) by Ilsa.StevesCafe.com (8.8.7/8.8.5) with ESMTP id JAA12019; Sun, 19 Oct 1997 09:41:54 -0600 (MDT) Message-Id: <199710191541.JAA12019@Ilsa.StevesCafe.com> X-Mailer: exmh version 2.0gamma 1/27/96 From: Steve Passe To: dan@math.berkeley.edu (Dan Strick) cc: freebsd-smp@freebsd.org Subject: Re: multiple I/O APICs In-reply-to: Your message of "Sun, 19 Oct 1997 07:28:11 PDT." <199710191428.HAA13450@math.berkeley.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Sun, 19 Oct 1997 09:41:54 -0600 Sender: owner-freebsd-smp@freebsd.org X-Loop: FreeBSD.org Precedence: bulk Hi, > The io_apic_setup() routine in mpapic.c doesn't like my motherboard. > It says "panic io_apic_setup: apic #1". The problem seems to be that > the motherboard has more than one I/O APIC and this is not supported. > > Please forgive this possibly stupid question, but I am not yet familiar > with the Intel 1.4 MP spec. Can I configure the kernel to just ignore > the second I/O APIC that it doesn't know how to use? If I set NAPIC > to 1 in the config file, the kernel panics when it starts and tells > me that I must reconfigure it with NAPIC set to 2. > > I am not certain about the motherboard's correct name or model number. > I thought the machine was an Intel "AP450GX MP Server System", but none > of the manuals or papers that came with it, including the rather bulky > installation manual, carry any model name or number. there should be an option in the BIOS to disable the second APIC. -- Steve Passe | powered by smp@csn.net | Symmetric MultiProcessor FreeBSD From owner-freebsd-smp Wed Oct 22 02:28:57 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id CAA19209 for smp-outgoing; Wed, 22 Oct 1997 02:28:57 -0700 (PDT) (envelope-from owner-freebsd-smp) Received: from btp1da.phy.uni-bayreuth.de (btp1da.phy.uni-bayreuth.de [132.180.20.32]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id CAA19198 for ; Wed, 22 Oct 1997 02:28:44 -0700 (PDT) (envelope-from werner@btp1da.phy.uni-bayreuth.de) Received: (from werner@localhost) by btp1da.phy.uni-bayreuth.de (8.8.7/8.7.3) id LAA02074; Wed, 22 Oct 1997 11:28:29 +0200 (MEST) Message-ID: <19971022112828.38489@btp1da.phy.uni-bayreuth.de> Date: Wed, 22 Oct 1997 11:28:28 +0200 From: Werner Griessl To: smp@freebsd.org Subject: smp-hardware ? Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.84e Sender: owner-freebsd-smp@freebsd.org X-Loop: FreeBSD.org Precedence: bulk Because the PPRO-180 is in the moment relatively cheap (~400DM) i think about a Dual-FBSD-System. Is anything wrong with this hardware ? Tyan S1662D Mainboard 2xPPRO 180 NCR-SCSI-Controller 4Gb-IBM-DCAS Intel EtherExpress 10/100B Matrox Millenium Soundblaster Vibra 16 Werner Please answer directly to wernerqbtp1da.phy.uni-bayreuth.de because i'm not ( yet ?) member of this list. From owner-freebsd-smp Wed Oct 22 09:44:34 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id JAA10384 for smp-outgoing; Wed, 22 Oct 1997 09:44:34 -0700 (PDT) (envelope-from owner-freebsd-smp) Received: from Ilsa.StevesCafe.com (Ilsa.StevesCafe.com [205.168.119.129]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id JAA10379 for ; Wed, 22 Oct 1997 09:44:31 -0700 (PDT) (envelope-from fbsd@Ilsa.StevesCafe.com) Received: from Ilsa.StevesCafe.com (localhost [127.0.0.1]) by Ilsa.StevesCafe.com (8.8.7/8.8.5) with ESMTP id KAA26461 for ; Wed, 22 Oct 1997 10:44:27 -0600 (MDT) Message-Id: <199710221644.KAA26461@Ilsa.StevesCafe.com> X-Mailer: exmh version 2.0gamma 1/27/96 From: Steve Passe To: smp@freebsd.org Subject: testing maillists Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Wed, 22 Oct 1997 10:44:27 -0600 Sender: owner-freebsd-smp@freebsd.org X-Loop: FreeBSD.org Precedence: bulk Hi, forgive this, I've received NO mail from any list in past 12 hours... -- Steve Passe | powered by smp@csn.net | Symmetric MultiProcessor FreeBSD From owner-freebsd-smp Thu Oct 23 16:54:53 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id QAA06892 for smp-outgoing; Thu, 23 Oct 1997 16:54:53 -0700 (PDT) (envelope-from owner-freebsd-smp) Received: from wally.eecs.harvard.edu (wally.eecs.harvard.edu [140.247.60.30]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id QAA06887 for ; Thu, 23 Oct 1997 16:54:51 -0700 (PDT) (envelope-from shieyuan@eecs.harvard.edu) Received: from steward.eecs.harvard.edu (shieyuan@steward.eecs.harvard.edu [140.247.60.20]) by wally.eecs.harvard.edu (8.8.6/8.8.6) with ESMTP id TAA18305 for ; Thu, 23 Oct 1997 19:54:45 -0400 (EDT) From: ShieYuan Wang Received: (from shieyuan@localhost) by steward.eecs.harvard.edu (8.8.5/8.8.5) id TAA27809 for freebsd-smp@freebsd.org; Thu, 23 Oct 1997 19:54:21 -0400 (EDT) Message-Id: <199710232354.TAA27809@steward.eecs.harvard.edu> Subject: SMP is slower when two processors are enabled To: freebsd-smp@freebsd.org Date: Thu, 23 Oct 97 19:54:21 EDT Mailer: Elm [revision: 70.85] Sender: owner-freebsd-smp@freebsd.org X-Loop: FreeBSD.org Precedence: bulk Hi, I am using SMP 3.0 with a two Pentium-Pro motherboard. I found that SMP does not give me speedup at all when I enable two processors at the same time. My system even becomes only 1/3 of its throughput when there is only one processor enabled. I am uisng "stcp" and "rtcp" to test how fast my system can use TCP/IP protocol stack to transfer unlimited data on the same local host. If I only enable one processor, the throughput can be 35 MB/sec. However if I enable the other processor hoping to get better throughput, I only get 12 MB/sec. I know that locking overhead between multiple processors may be a lot. But I never thought that it could be so high. Could anyone give me some clues about this bad performance? My SMP 3.0 was installed about 6 months ago. Does the more recent SMP have better performance and have solved the performance problem? Your suggestion will be highly appriciated. Shie-Yuan Wang CS Ph.D. student. Harvard University From owner-freebsd-smp Thu Oct 23 17:15:01 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id RAA08117 for smp-outgoing; Thu, 23 Oct 1997 17:15:01 -0700 (PDT) (envelope-from owner-freebsd-smp) Received: from dyson.iquest.net (dyson.iquest.net [198.70.144.127]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id RAA08096 for ; Thu, 23 Oct 1997 17:14:56 -0700 (PDT) (envelope-from toor@dyson.iquest.net) Received: (from root@localhost) by dyson.iquest.net (8.8.7/8.8.5) id TAA01451; Thu, 23 Oct 1997 19:14:48 -0500 (EST) From: "John S. Dyson" Message-Id: <199710240014.TAA01451@dyson.iquest.net> Subject: Re: SMP is slower when two processors are enabled In-Reply-To: <199710232354.TAA27809@steward.eecs.harvard.edu> from ShieYuan Wang at "Oct 23, 97 07:54:21 pm" To: shieyuan@eecs.harvard.edu (ShieYuan Wang) Date: Thu, 23 Oct 1997 19:14:48 -0500 (EST) Cc: freebsd-smp@freebsd.org X-Mailer: ELM [version 2.4ME+ PL31 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@freebsd.org X-Loop: FreeBSD.org Precedence: bulk ShieYuan Wang said: > protocol stack to transfer unlimited data on the same local host. If I > only enable one processor, the throughput can be 35 MB/sec. However if > I enable the other processor hoping to get better throughput, I only get > 12 MB/sec. > > I know that locking overhead between multiple processors may be a lot. But > I never thought that it could be so high. Could anyone give me some clues > about this bad performance? My SMP 3.0 was installed about 6 months ago. > Does the more recent SMP have better performance and have solved the > performance problem? Your suggestion will be highly appriciated. > Most likely it is the overhead of moving data between CPUs. Could be cache effects also. -- John dyson@freebsd.org jdyson@nc.com From owner-freebsd-smp Fri Oct 24 10:16:39 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id KAA04871 for smp-outgoing; Fri, 24 Oct 1997 10:16:39 -0700 (PDT) (envelope-from owner-freebsd-smp) Received: from deimos.nc.com (deimos.nc.com [207.88.167.190]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id KAA04861; Fri, 24 Oct 1997 10:16:36 -0700 (PDT) (envelope-from TYANG@nc.com) Received: (from oracle@localhost) by deimos.nc.com (8.8.4/8.8.4) id KAA12913; Fri, 24 Oct 1997 10:14:08 -0700 (PDT) Message-Id: <199710241714.KAA12913@deimos.nc.com> Date: 24 Oct 97 10:13:27 -0700 From: "Tina Yang" To: shieyuan@eecs.harvard.edu, owner-freebsd-smp@freebsd.org Subject: Re: SMP is slower when two processors are enabled Cc: freebsd-smp@freebsd.org MIME-Version: 1.0 X-Mailer: Oracle InterOffice (version 4.1.1.1.10) Content-Type: multipart/mixed; boundary="=_ORCL_2577328_0_1191971024111509_0" Sender: owner-freebsd-smp@freebsd.org X-Loop: FreeBSD.org Precedence: bulk --=_ORCL_2577328_0_1191971024111509_0 Content-Transfer-Encoding:quoted-printable Content-Type:text/plain; charset="iso-8859-1" >ShieYuan Wang said: >> protocol stack to transfer unlimited data on the same local host. If I >> only enable one processor, the throughput can be 35 MB/sec. However if >> I enable the other processor hoping to get better throughput, I only get >> 12 MB/sec. >> >> I know that locking overhead between multiple processors may be a lot. But >> I never thought that it could be so high. Could anyone give me some clues >> ABOUT THIS BAD Performance? My SMP 3.0 was installed about 6 months ago. >> Does the more recent SMP have better performance and have solved the >> performance problem? Your suggestion will be highly appriciated. >> >Most likely it is the overhead of moving data between CPUs. Could be >cache effects also. > >-- >John Interesting. Is it true that SMP 3.0 kernel is still single-threaded ? Since stcp/rtcp is io intensive, it's like you still only have one thread of execution most of the time (like UP), and in the mean time, you lost the UP cache efficiency as well when two processors alternately execute the kernel ?? - Tina --=_ORCL_2577328_0_1191971024111509_0 Content-Type:message/rfc822 Date: 23 Oct 97 17:14:48 From:"John S. Dyson " To:shieyuan@eecs.harvard.edu,(ShieYuan,Wang) Subject:Re: SMP is slower when two processors are enabled Cc:freebsd-smp@FreeBSD.ORG Return-Path: Received:from maildrop.nc.com (proxy@nc.com [207.88.167.98]) by deimos.nc.com (8.8.4/8.8.4) with SMTP id RAA23966 for ; Thu, 23 Oct 1997 17:50:15 -0700 (PDT) Received:(from proxy@localhost) by maildrop.nc.com (8.6.12/8.6.9) id RAA17164 for ; Thu, 23 Oct 1997 17:48:43 -0700 Received:from sarip.sol.net(169.207.30.120) by sebastian.nc.com via smap (V2.0) id xma017160; Thu, 23 Oct 97 17:48:17 -0700 Received:from hub.freebsd.org (hub.FreeBSD.ORG [204.216.27.18]) by sarip.sol.net (8.8.7/8.8.4) with ESMTP id TAA06466; Thu, 23 Oct 1997 19:36:29 -0500 (CDT) Received:(from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id RAA08117 for smp-outgoing; Thu, 23 Oct 1997 17:15:01 -0700 (PDT) (envelope-from owner-freebsd-smp) Received:from dyson.iquest.net (dyson.iquest.net [198.70.144.127]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id RAA08096 for ; Thu, 23 Oct 1997 17:14:56 -0700 (PDT) (envelope-from toor@dyson.ique Received:(from root@localhost) by dyson.iquest.net (8.8.7/8.8.5) id TAA01451; Thu, 23 Oct 1997 19:14:48 -0500 (EST) Message-Id:<199710240014.TAA01451@dyson.iquest.net> In-Reply-To:<199710232354.TAA27809@steward.eecs.harvard.edu> from ShieYuan Wang at "Oct 23, 97 07:54:21 pm" Sender:owner-freebsd-smp@FreeBSD.ORG X-Loop:FreeBSD.org Precedence: bulk MIME-Version: 1.0 Content-Type:text/plain; charset=US-ASCII Content-Transfer-Encoding:7bit ShieYuan Wang said: > protocol stack to transfer unlimited data on the same local host. If I > only enable one processor, the throughput can be 35 MB/sec. However if > I enable the other processor hoping to get better throughput, I only get > 12 MB/sec. > > I know that locking overhead between multiple processors may be a lot. But > I never thought that it could be so high. Could anyone give me some clues > about this bad performance? My SMP 3.0 was installed about 6 months ago. > Does the more recent SMP have better performance and have solved the > performance problem? Your suggestion will be highly appriciated. > Most likely it is the overhead of moving data between CPUs. Could be cache effects also. -- John dyson@freebsd.org jdyson@nc.com --=_ORCL_2577328_0_1191971024111509_0--