Path: csiph.com!weretis.net!feeder6.news.weretis.net!panix!.POSTED.panix5.panix.com!not-for-mail From: Rich Alderson Newsgroups: alt.sys.pdp10,alt.folklore.computers Subject: Re: XKL (PDP10) Site Announcement Date: 25 Sep 2021 20:54:51 -0400 Organization: PANIX Public Access Internet and UNIX, NYC Lines: 70 Sender: alderson+news@panix5.panix.com Message-ID: References: <87lf3lcbn9.fsf@nightsong.com> Injection-Info: reader1.panix.com; posting-host="panix5.panix.com:166.84.1.5"; logging-data="13496"; mail-complaints-to="abuse@panix.com" X-Newsreader: Gnus v5.7/Emacs 22.3 Xref: csiph.com alt.sys.pdp10:8337 alt.folklore.computers:219041 Paul Rubin writes: > "Stephen M. Jones" writes: >> I have invested in a TOAD-2 machine running the PDP-10 architecture, and >> would like to make its capabilities available. The TOAD is manufactured by >> XKL LLC, and is a completely new hardware implementation of Digital's 36-bit >> PDP-10 architecture, not an emulator running on another platform. This >> machine I've purchased will be a repository for PDP10 public domain >> software... > This is cool, thanks for doing it, and I didn't realize that the Toad-2 > existed. Finding much info about it is non-trivial. Is it really built > around an ASIC? Or maybe an FPGA, that would be easier? When was it built? > Was there really much benefit by then of building a hardware implementation > instead of simply running an emulator? I've been slightly interested in > fooling with emulated PDP-10's for various purposes, but have never > understood the attraction of hardware reimplementations once emulators became > a lot faster than historical 10's. Hmm. Start with searches in these newsgroups from the 1990s, when I was (among other things) running sales support and customer support for XKL. Look for posts from "alderson@xkl.com". Note, please, that I have not been an employee of XKL since 2003, and do not now speak for them in any way, official or un-. XKL was started by Len Bosack, co-founder of cisco Systems (original spelling), to build what was originally intended to be cisco's premier product, a desktop PDP-10 clone which implemented things like the full 30 bit address space in hardware. The result was the introduction, in 1995, of the XKL Toad-1 System, an extended clone the size of a two drawer file cabinet. The Toad-1 featured a processor roughly 2x faster than the KL-10, a 4-port 10baseT Ethernet interface, a 4-port FASTWIDE SCSI-2 interface, and a 32MW memory card, in a 7 slot backplane. One of the tests we performed was to put 4 memory cards in, to see if our mods to TOPS-20 were correct; the system sat there happily using 128MW and never even hiccoughed. We sold a few systems, to MCI, Digital, and Paul Allen (as an individual, not to one of his companies); others were gifts to friends of the company. The XKL-1 processor board was built around an FPGA from Xilinx. The TOAD-2 processor is a reimplementation of the Toad-1 in a single FPGA. The primary file systems are on 2 microSD cards; the Ethernet is built in, and they provide a NAS application which runs on Linux to allow for much more than 2GB of user storage. The early version, which we had at Living Computers: Museum + Labs, was labeled "TOAD-2"; the current version is embedded in their DarkStar optical router boxes. The TOAD-2 is used as the configuration driver in their 40G/100G and faster optical networking products (where routing is done in the fiber header, not a general purpose processor); if no networking interfaces are detected, the system will assume that it should boot TOPS-20 instead. They are still using (bigger, faster) FPGAs in their designs. Len has always held that smart engineers can get the speeds they want from better designs, instead of wasting time on ASICs which will have to be competely replaced instead of being modified if a bug, or an enhancement, needs to be updated. At the time we build the Toad-1 (on which I did some of the operating system work as well as my other duties), PC clocks had not reached the 1GHz range yet, and it was a perfectly reasonable product line, given that PDP-10 users had at least 20 years invested in local software (like most mainframe shops, primary software loads were local). We even ported Tops-10 to the Toad-1 for a potential customer; no one apparently has asked for Tops-10 for the TOAD-2. -- Rich Alderson news@alderson.users.panix.com Audendum est, et veritas investiganda; quam etiamsi non assequamur, omnino tamen proprius, quam nunc sumus, ad eam perveniemus. --Galen