From owner-freebsd-hackers Fri Mar 24 09:43:29 1995 Return-Path: hackers-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id JAA13375 for hackers-outgoing; Fri, 24 Mar 1995 09:43:29 -0800 Received: from cs.weber.edu (cs.weber.edu [137.190.16.16]) by freefall.cdrom.com (8.6.10/8.6.6) with SMTP id JAA13369 for ; Fri, 24 Mar 1995 09:43:28 -0800 Received: by cs.weber.edu (4.1/SMI-4.1.1) id AA09749; Fri, 24 Mar 95 10:35:52 MST From: terry@cs.weber.edu (Terry Lambert) Message-Id: <9503241735.AA09749@cs.weber.edu> Subject: Re: DEC Alpha Multia To: kuku@gilberto.physik.rwth-aachen.de Date: Fri, 24 Mar 95 10:35:51 MST Cc: aledm@relay-europe.ps.net, freebsd-hackers@freefall.cdrom.com In-Reply-To: <199503241558.QAA18500@gilberto.physik.rwth-aachen.de> from "Christoph Kukulies" at Mar 24, 95 04:58:45 pm X-Mailer: ELM [version 2.4dev PL52] Sender: hackers-owner@FreeBSD.org Precedence: bulk > > Anyone thought about a port of FreeBSD to the DEC Alpha processor? The > > new Multia machine seems to offer good value for money hardware (166MHz > > 64bit Alpha processor, 24Mb RAM, 340Mb SCSI-2 HDD, PCI bus, and Ethernet > > for $3000) > > > > Perhaps the BSD consortium could fund a development machine if we had a > > volunteer to do the port? > > There is a NetBSD port to the APX architecture. This is for the DEC AXP150. An EISA machine with an Adaptec 1742 controller; there is supposedly no X support, and a number of other problems. There is, however, the basic code for the Alpha processor with the OSF Alpha microcode, and *significant* work on many parts of the 4.4 system to make it 64 bit clean. On the other hand, EISA is being replaced by PCI and the AHA1742 is a dead piece of hardware (so is the AXP150, for that matter). The new DEC Alpha PCI motherboard ($1170 from DEC Direct), although it wants PS/2 style SIMMs (bletch) and the default microcode requires 16M of memory, seems a much better deal. There is a DEC-available-to-askers-only microcode update disk that allows it to run with only 8M (still the OSF microcode otherwise) and in fact, Linux is already booting from floppy on this hardware configuration (the guy doing the work lives about 17 minutes away from me). Personally, I would prefer that at *least* kernel multithreading and *preferrably* SMP support were in prior to ossifying the kernel and VM structures into multiple architectures which would then have to be individually retrofitted as interface changes were made to put them in. In the limit, I think it possible that the job might become to large to *ever* tackle otherwise. Terry Lambert terry@cs.weber.edu --- Any opinions in this posting are my own and not those of my present or previous employers.