From: "David G. Conroy" Newsgroups: alt.sys.pdp10 Subject: PDP-10 progress Organization: Used only for posts to newsgroups X-Newsreader: Microsoft Outlook Express Macintosh Edition - 4.5 (0410) Mime-version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-transfer-encoding: 7bit Lines: 18 Message-ID: NNTP-Posting-Host: 66.120.160.49 X-Complaints-To: abuse@prodigy.net X-Trace: newssvr14.news.prodigy.com 1025327279 ST000 66.120.160.49 (Sat, 29 Jun 2002 01:07:59 EDT) NNTP-Posting-Date: Sat, 29 Jun 2002 01:07:59 EDT X-UserInfo1: Q[R_@SVDLRUORWDYNCOF_W\@PJ_^PBQLGPQRJWQHBATBTSUBYFWEAE[YJLYPIWKHTFCMZKVMB^[Z^DOBRVVMOSPFHNSYXVDIE@X\BUC@GTSX@DL^GKFFHQCCE\G[JJBMYDYIJCZM@AY]GNGPJD]YNNW\GSX^GSCKHA[]@CCB\[@LATPD\L@J\\PF]VR[QPJN Date: Sat, 29 Jun 2002 05:07:59 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.imp.ch!news.imp.ch!fr.clara.net!heighliner.fr.clara.net!skynet.be!skynet.be!prodigy.com!newsmst01.news.prodigy.com!prodigy.com!postmaster.news.prodigy.com!newssvr14.news.prodigy.com.POSTED!3a2e9f9f!not-for-mail Xref: chonsp.franklin.ch alt.sys.pdp10:11563 Just in case anyone is interested, as of tonight I have all of the processor instruction diagnostics running in an emulation of my PDP-10 design's microprocessor running real microcode. Getting the LSB correct on the last FDVR case was really nasty! The instruction set implemented by the current version of the microcode is small; it's the intersection of the KA and the KL/KS, which is sufficient to run ITS. Everything else is complete and seems to work; interrupts, traps, paging, page fail on trap instructions, XCTR, the microcoded console, and so on. The hardware should be easy to design (the microprocessor is dead simple) but I don't plan to start designing hardware until ITS runs in the emulator. dgc ###### From: klh@panix.com (Ken Harrenstien) Newsgroups: alt.sys.pdp10 Subject: Re: PDP-10 progress Date: 30 Jun 2002 16:58:56 -0400 Organization: PANIX Public Access Internet and UNIX, NYC Lines: 9 Message-ID: References: NNTP-Posting-Host: panix2.panix.com X-Trace: reader2.panix.com 1025470736 13334 166.84.1.2 (30 Jun 2002 20:58:56 GMT) X-Complaints-To: abuse@panix.com NNTP-Posting-Date: Sun, 30 Jun 2002 20:58:56 +0000 (UTC) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.imp.ch!news.imp.ch!eusc.inter.net!news.stealth.net!news.stealth.net!jfk3-feed1.news.algx.net!allegiance!panix!panix2.panix.com!not-for-mail Xref: chonsp.franklin.ch alt.sys.pdp10:11559 In article , David G. Conroy wrote: >The hardware should be easy to design (the microprocessor >is dead simple) but I don't plan to start designing hardware until ITS >runs in the emulator. Just wondering, what's the plan for devices? (eg disk, tape, net) --Ken ###### Path: chonsp.franklin.ch!not-for-mail From: Neil Franklin Newsgroups: alt.sys.pdp10 Subject: Re: PDP-10 progress Date: 01 Jul 2002 01:59:01 +0200 Organization: My own Private Self Lines: 21 Message-ID: <6u7kkg2u9m.fsf@chonsp.franklin.ch> References: NNTP-Posting-Host: chonsp.franklin.ch X-Trace: chonsp.franklin.ch 1025481541 947 10.0.3.2 (30 Jun 2002 23:59:01 GMT) X-Complaints-To: news@chonsp.franklin.ch NNTP-Posting-Date: 30 Jun 2002 23:59:01 GMT X-Newsreader: Gnus v5.7/Emacs 20.4 Xref: chonsp.franklin.ch alt.sys.pdp10:11566 "David G. Conroy" writes: > my PDP-10 design's microprocessor running real microcode. > > The instruction set implemented by the current version of the microcode So yours is going to be microcoded. Where/how do you store microcode? In the FPGA (if so what FPGA for that amount of space?) or externally (if what is the impact on pins and speed?)? > The hardware should be easy to design (the microprocessor > is dead simple) So more KS than KL? Or even simpler? -- Neil Franklin, neil@franklin.ch.remove http://neil.franklin.ch/ Hacker, Unix Guru, El Eng HTL/BSc, Programmer, Archer, Roleplayer - Make your code truely free: put it into the public domain ###### From: "David G. Conroy" Newsgroups: alt.sys.pdp10 Subject: Re: PDP-10 progress References: Organization: Used only for posts to newsgroups X-Newsreader: Microsoft Outlook Express Macintosh Edition - 4.5 (0410) Mime-version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-transfer-encoding: 7bit Lines: 48 Message-ID: NNTP-Posting-Host: 66.120.163.35 X-Complaints-To: abuse@prodigy.net X-Trace: newssvr14.news.prodigy.com 1025492779 ST000 66.120.163.35 (Sun, 30 Jun 2002 23:06:19 EDT) NNTP-Posting-Date: Sun, 30 Jun 2002 23:06:19 EDT X-UserInfo1: Q[R_@SVDLRUORWDYNCOF_W\@PJ_^PBQLGPQRJWQHBATBTSUBYFWEAE[YJLYPIWKHTFCMZKVMB^[Z^DOBRVVMOSPFHNSYXVDIE@X\BUC@GTSX@DL^GKFFHQCCE\G[JJBMYDYIJCZM@AY]GNGPJD]YNNW\GSX^GSCKHA[]@CCB\[@LATPD\L@J\\PF]VR[QPJN Date: Mon, 01 Jul 2002 03:06:19 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.imp.ch!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!cpk-news-hub1.bbnplanet.com!news.gtei.net!newscon02.news.prodigy.com!newsmst01.news.prodigy.com!prodigy.com!postmaster.news.prodigy.com!newssvr14.news.prodigy.com.POSTED!3a2e9f9f!not-for-mail Xref: chonsp.franklin.ch alt.sys.pdp10:11571 To Ken: Plan is to do the same thing I did for my pdp-8 and pdp-4 machines, which is to use modern devices and fix the software (this isn't quite what I did on the pdp-8; there I designed a pdp-8 compatible uart because apps knew how it worked, especially the non-destructive read). In some sense doing this on a -10 is easy because the software can actually be rebuilt from sources; I can create, compile, and execute fortran programs on my pdp-4, and all I had was binaries (and binaries for the pdp-9/15 at that), which meant hours of reading disassembler output. Currently the spec says ... The console is an industry-standard UART (the spec says 16450/16550 today, and that's what the emulator emulates. The disk is IDE. Commands are PIO, and data is either PIO or DMA. The disk DMA channel splits each 36-bit word into 4 9-bit frames and puts each in a 16-bit IDE data word. The DMA channel does scatter/gather with a channel chain (in ITS all I/O is to aligned 1K blocks). The NET interface is built with an NSC chip which stores its buffer rings in dedicated memory. A microcode assist copies data in and out of the packet buffers, with the 32-bits of useful data stored in the 36-bit words in an ITS-friendly way. To Neil: The microcode is off chip, stored in 3 32Kx8 3.3V SRAMS. The 12ns version of these chips cost about $1.00, and come from many vendors (cypress, alliance, winbond, paradigm, perhaps others). Much of the control store is empty; the basic decode is a 512 way dispatch into slots which are big enough to handle fairly complex instructions, and many instructions are 2 or 3 instructions in length. The external access sets the cycle time of the microprocessor to something in the 33MHz range, with the SDRAMS running at 66MHz. The micropipeline is 3 stages long (fetch + 2 execute) and all microbranches have a 1-cycle delay. If I wanted to make the machine faster I would not work on the cycle time; I'd build more parallelism into the instruction fetch and decode. I did some work on this, but it got quite hairy quite quickly (there's executing from the ACs, there's writing into the next instruction) so I decided to punt. There really isn't much to the hardware. I can see places where the timing will be interesting (the address computation for the read port of the AC ram, which has to deal with the current/previous block stuff and do the "+1" needed for ASHC and DADD and things like that) but nothing seems unmanageable. ###### From: "Zane H. Healy" Newsgroups: alt.sys.pdp10 Subject: Re: PDP-10 progress Date: 1 Jul 2002 21:28:02 GMT Organization: Aracnet Lines: 10 Message-ID: References: NNTP-Posting-Host: p-135.newsdawg.com User-Agent: tin/1.4.4-20000803 ("Vet for the Insane") (UNIX) (Linux/2.2.19 (i686)) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!newscore.univie.ac.at!newsgate.cistron.nl!news.tele.dk!small.news.tele.dk!216.218.192.242!news.he.net!cyclone-sf.pbi.net!199.106.71.17!pln-w!spln!dex!extra.newsguy.com!newsp.newsguy.com!not-for-mail Xref: chonsp.franklin.ch alt.sys.pdp10:11568 David G. Conroy wrote: > If I wanted to make the machine faster I would not work What speed to you see it being at in comparasion to real PDP-10 models? Zane PS I hope once you get ITS up that you also add the ability to run TOPS-10/TOPS-20. ###### From: "David G. Conroy" Newsgroups: alt.sys.pdp10 Subject: Re: PDP-10 progress References: Organization: Used only for posts to newsgroups X-Newsreader: Microsoft Outlook Express Macintosh Edition - 4.5 (0410) Mime-version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-transfer-encoding: 7bit Lines: 33 Message-ID: <%I8U8.770$t42.67749176@newssvr14.news.prodigy.com> NNTP-Posting-Host: 66.120.163.35 X-Complaints-To: abuse@prodigy.net X-Trace: newssvr14.news.prodigy.com 1025578939 ST000 66.120.163.35 (Mon, 01 Jul 2002 23:02:19 EDT) NNTP-Posting-Date: Mon, 01 Jul 2002 23:02:19 EDT X-UserInfo1: OHZACY_DVZVS@^LYMRKNOPDA[X_LPO@FFA_J]Q]KEYUNDQUCCNSUAACY@L[ZX__HGFD]JBJNSFXTOOGA_VWY^_HG@FW_HUTHOH]TBPGCO\P^PLP^@[GLHUK@WLECKFVL^TYG[@RMWQXIWM[SDDYWNLG_G[_BWUCHFY_Y@AS@Q[B\APPF@DCZM_PG_VSCPQZM Date: Tue, 02 Jul 2002 03:02:19 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.imp.ch!news.imp.ch!nntp-out.monmouth.com!newspeer.monmouth.com!newscon02.news.prodigy.com!newsmst01.news.prodigy.com!prodigy.com!postmaster.news.prodigy.com!newssvr14.news.prodigy.com.POSTED!3a2e9f9f!not-for-mail Xref: chonsp.franklin.ch alt.sys.pdp10:11579 Simple ops between 2 and 3 MIPS, most of that waiting for read data from memory. The simple design aggravates the memory latency situation; reads are kicked off late in the pipe but read data is needed early in the pipe, so cycles get wasted. I actually have a section in the spec which talks about what needed to be done to make a TOPS-20 port possible; it's mainly writing the TOPS-20 paging microcode, and filling in the gaps in the instruction set. This is more work than anything else, and the micromachine is easy to program; the microcode is highly vertical, so it's like writing assembly language for a somewhat weird general purpose computer. There's lots of space in the control store and enough space in the scratchpad for the various bases and things. dgc ---------- In article , "Zane H. Healy" wrote: > David G. Conroy wrote: >> If I wanted to make the machine faster I would not work > > What speed to you see it being at in comparasion to real PDP-10 models? > > Zane > > PS I hope once you get ITS up that you also add the ability to run > TOPS-10/TOPS-20. > ###### From: "Zane H. Healy" Newsgroups: alt.sys.pdp10 Subject: Re: PDP-10 progress Date: 2 Jul 2002 03:40:32 GMT Organization: Aracnet Lines: 11 Message-ID: References: <%I8U8.770$t42.67749176@newssvr14.news.prodigy.com> NNTP-Posting-Host: p-907.newsdawg.com User-Agent: tin/1.4.4-20000803 ("Vet for the Insane") (UNIX) (Linux/2.2.19 (i686)) Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.imp.ch!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!pln-e!spln!dex!extra.newsguy.com!newsp.newsguy.com!not-for-mail Xref: chonsp.franklin.ch alt.sys.pdp10:11576 David G. Conroy wrote: > I actually have a section in the spec which talks about > what needed to be done to make a TOPS-20 port possible; it's mainly > writing the TOPS-20 paging microcode, and filling in the > gaps in the instruction set. This is more work than anything else, Out of curiosity, have you put the spec online anywhere? I checked your website and only see the the PDP-4/X and PDP-8/X pages. The spec sounds like it might be an interesting read. Zane ###### From: "David G. Conroy" Newsgroups: alt.sys.pdp10 Subject: Re: PDP-10 progress References: <%I8U8.770$t42.67749176@newssvr14.news.prodigy.com> Organization: Used only for posts to newsgroups X-Newsreader: Microsoft Outlook Express Macintosh Edition - 4.5 (0410) Mime-version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-transfer-encoding: 7bit Lines: 20 Message-ID: NNTP-Posting-Host: 66.120.163.35 X-Complaints-To: abuse@prodigy.net X-Trace: newssvr13.news.prodigy.com 1025583881 ST000 66.120.163.35 (Tue, 02 Jul 2002 00:24:41 EDT) NNTP-Posting-Date: Tue, 02 Jul 2002 00:24:41 EDT X-UserInfo1: FKPO@MC@@S@KRZPX@ZKBOFTBTR\B@GXLN@GZ_GYO^ZUDUWYAKVUOPCW[ML\JXUCKVFDYZKBMSFX^OMSAFNTINTDDMVW[X\THOPXZRVOCJTUTPC\_JSBVX\KAOTBAJBVMZTYAKMNLDI_MFDSSOLXINH__FS^\WQGHGI^C@E[A_CF\AQLDQ\BTMPLDFNVUQ_VM Date: Tue, 02 Jul 2002 04:24:41 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!feedme.news.mediaways.net!andromeda.datanet.hu!newsfeed.gamma.ru!Gamma.RU!news.maxwell.syr.edu!nntp.flash.net!newscon02.news.prodigy.com!prodigy.com!newsmst01.news.prodigy.com!prodigy.com!postmaster.news.prodigy.com!newssvr13.news.prodigy.com.POSTED!3a2e9f9f!not-for-mail Xref: chonsp.franklin.ch alt.sys.pdp10:11580 Not yet. Maybe soon, work permitting. It's in a "me only" state right now, and it doesn't describe the micromachine at all (that is only in my design notes and a very cryptic include file shared by the simulator and the assembler). ---------- In article , "Zane H. Healy" wrote: > David G. Conroy wrote: >> I actually have a section in the spec which talks about >> what needed to be done to make a TOPS-20 port possible; it's mainly >> writing the TOPS-20 paging microcode, and filling in the >> gaps in the instruction set. This is more work than anything else, > > Out of curiosity, have you put the spec online anywhere? I checked your > website and only see the the PDP-4/X and PDP-8/X pages. The spec sounds > like it might be an interesting read. > > Zane