Article View: comp.sys.transputer
Article #679Mixed C40/transputer networks
From: Stephen J. Bradshaw
Date: Sat, 22 Aug 1992 15:56
Date: Sat, 22 Aug 1992 15:56
63 lines
3002 bytes
3002 bytes
> Never having used a C40 (this being comp.sys.transputer) the > following questions come to mind. I would have thought that the C40 is sufficiently similar to the transputer, at least in concept, that comp.sys.transputer is a suitable place for such discussions. After all, it gives people such as Mr. Doyle, of the Software Marketing Group at Inmos, the opportunity to tell us how much better the real thing will be :^) > 1. Its a newer chip then a T8 and clearly faster, but it an actual > multiprocessing system do you get more bang per buck? > (1) A 50 MFLOP CPU. (2) Six bi-directional 20 MBytes/s Communication Ports (Links) (3) Six channel DMA co-processor enabling concurrent operation of Communication Ports and CPU (4) Two 100 MBytes/s Memory Ports (5) On-chip analysis module with JTAG (IEEE1149.1) interface. Monitors activities of processor, enables control of breakpoints, and privides read/write access to full memory map. (6) TIM-40 standard, the TRAM standard of the C40 market. (7) Source compatibility offered by Helios and 3L > 2. Is the architecture up to the job? Context switch time is just as > important as comunications speed. To get the much sort after linear > increase in performace > > Performance | / > | / > | / > |/ > |-------- > No of processors > each processor must be able to fill up all its available time. To do > this it must be able to quickly do something else while it waits for > other processors i.e. Context switch The is obviously dependent on the hardware but ultimately a function of the software system being used, as the C40 does not support multi-tasking in hardware as does the tranpsuter. Spectron Microsystems, for example, quote a "task" switching time for SPOX of 18 processor cycles. This is the time to complete the last instruction of one task to being ready to execute the first instruction of the next task. Your comments about keeping the CPU busy should of course be inscribed on every transputer users brain (mine included). Keeping the CPU busy is also possible on the C40 since the Comm Ports can run autonomously with the CPU. The same rules apply but the numbers are different. Steve Bradshaw Radioactive cats have 18 half-lives ! Traquair Data Systems Judd Falls Plaza, Suite 80 Phone: (607) 272 4417 Ithaca, NY 14850 Fax: (607) 272 6211
Technical Details
Message-ID:
<9208221556.AA05214@theory.TC.CORNELL.EDU>Path:
arc1-ca.pugleaf.net!sparky!uunet!usc!zaphod.mps.ohio-state.edu!cis.ohio-state.edu!ucbvax!THEORY.TC.CORNELL.EDU!sjbShow Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!sparky!uunet!usc!zaphod.mps.ohio-state.edu!cis.ohio-state.edu!ucbvax!THEORY.TC.CORNELL.EDU!sjb From: sjb@THEORY.TC.CORNELL.EDU (Stephen J. Bradshaw) Newsgroups: comp.sys.transputer Subject: Mixed C40/transputer networks Message-ID: <9208221556.AA05214@theory.TC.CORNELL.EDU> Date: 22 Aug 92 15:56:44 GMT Sender: daemon@ucbvax.BERKELEY.EDU Distribution: world Organization: The Internet Lines: 64