Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
| From | Michael Black <et472@ncf.ca> |
|---|---|
| Newsgroups | comp.misc |
| Subject | Re: Does a compiler use all x86 instructions? (2010) |
| Date | 2016-08-24 23:03 -0400 |
| Organization | National Capital Freenet, Ottawa, Ontario, Canada |
| Message-ID | <alpine.LNX.2.02.1608242258560.18471@darkstar.example.org> (permalink) |
| References | <nplgj9$36g$3@solani.org> |
On Thu, 25 Aug 2016, RS Wood wrote: > Old news to you maybe, but I learned something from it. > > http://pepijndevos.nl/2016/08/24/x86-instruction-distribution.html > > //--clip > On the Z80 CPU there are so few registers and instructions that you can > easily know and use them all and wish there were more of them. However, > half of the time it feels like the only on you really use is ld. I imagine > that if ld took half the number of clock cycles, average code would run > twice as fast. > Is that really true? I remembered opcodes for the 6502 precisely because I was using a small subset. I remember someone looked into this, and discovered that over a lot of software, a small handful of instructions were used, the rest seeing limited if any use. I think this wsa some of the optimizing done by Motorola for the 6809. They ran programs on a lot of available 6800 software, to find out what was used, and maybe find the bottlenecks, trying to improve the key points. There was a three part article by Motorola employees about 1978 in Byte that showed the process they used to create the 6809 architecture and instruction set. Don't forget that while some of the Z80 instructions did complicated things, they also used up a lot of clock cycles. I seem to recall looking into that, and some decided the 8080 instructions were faster, even if you had to use more of them to do the equivalent of the fancy Z80 instructions. Michael
Back to comp.misc | Previous | Next — Previous in thread | Next in thread | Find similar
Does a compiler use all x86 instructions? (2010) RS Wood <rsw@therandymon.com> - 2016-08-25 01:11 +0000
Re: Does a compiler use all x86 instructions? (2010) Rich <rich@example.invalid> - 2016-08-25 01:24 +0000
Re: Does a compiler use all x86 instructions? (2010) Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-08-25 13:29 +0100
Re: Does a compiler use all x86 instructions? (2010) Mark Storkamp <mstorkamp@yahoo.com> - 2016-08-25 08:51 -0500
Re: Does a compiler use all x86 instructions? (2010) Rich <rich@example.invalid> - 2016-08-25 16:20 +0000
Re: Does a compiler use all x86 instructions? (2010) Walter Banks <walter@bytecraft.com> - 2016-08-25 10:34 -0400
Re: Does a compiler use all x86 instructions? (2010) Michael Black <et472@ncf.ca> - 2016-08-24 23:03 -0400
Re: Does a compiler use all x86 instructions? (2010) Theo Markettos <theom+news@chiark.greenend.org.uk> - 2016-08-25 14:23 +0100
Re: Does a compiler use all x86 instructions? (2010) Mark Storkamp <mstorkamp@yahoo.com> - 2016-08-25 08:39 -0500
Re: Does a compiler use all x86 instructions? (2010) Richard Kettlewell <invalid@invalid.invalid> - 2016-08-25 16:40 +0100
Re: Does a compiler use all x86 instructions? (2010) BartC <bc@freeuk.com> - 2016-08-25 19:34 +0100
Re: Does a compiler use all x86 instructions? (2010) Marko Rauhamaa <marko@pacujo.net> - 2016-08-25 22:41 +0300
Re: Does a compiler use all x86 instructions? (2010) Rich <rich@example.invalid> - 2016-08-25 20:49 +0000
Re: Does a compiler use all x86 instructions? (2010) Richard Kettlewell <invalid@invalid.invalid> - 2016-08-25 22:26 +0100
csiph-web