richmon@astrovax.UUCP (Michael Richmond) (05/18/85)
I went to play with the 7300 at my local computer store yesterday and have
some (dis?)heartening news for people like me who want to do a lot of
program development on the machine. The compile times for a simple program
consisting of nothing but lines like
printf("hello there\n");
is (for 7300 in window environment, 7300 out of window environment and VAX
11/750 (out of window environment :-) ))
7300 w/ windows 7300 VAX
65 statements :28 :09
250 statements 1:58 :55 :21
The number that really gets me down is the figure for 7300 performance
with windows (the 'user agent') running. It seemed that there was a lot of
disk activity in the window environment even when I was doing nothing -
I'm not sure what the CPU was doing, but it ate up cycles like crazy.
Perhaps nice-ing up large compiles would help, but by how much? Has anyone
else done tests to compare with this rather simple-minded one?
I was expecting somewhat faster times from the 7300, especially in the
window environment where I'll be doing most of my work (oh, yes, I decided
to get one anyway), but perhaps I was being unrealistic. I'd be willing to
bet that a comparison of execution times for would put the 7300 in a better
light, but for development that isn't nearly as important. Ah well.
--
Michael Richmond Princeton University, Astrophysics
{allegra,akgua,burl,cbosgd,decvax,ihnp4,noao,princeton,vax135}!astrovax!richmonrichmon@astrovax.UUCP (Michael Richmond) (05/20/85)
In response to several questions I've received, I ought to have mentioned
that the times I posted were for a 7300 with (so I was told) 1 Meg and the
20 Meg hard disk. I shudder to think of the performance with 512 K.
--
Michael Richmond Princeton University, Astrophysics
{allegra,akgua,burl,cbosgd,decvax,ihnp4,noao,princeton,vax135}!astrovax!richmonmash@mips.UUCP (John Mashey) (05/20/85)
Michael Richmond writes: > I went to play with the 7300 at my local computer store yesterday and have > some (dis?)heartening news for people like me who want to do a lot of > program development on the machine. The compile times for a simple program .... > is (for 7300 in window environment, 7300 out of window environment and VAX > 11/750 (out of window environment :-) )) > 7300 w/ windows 7300 VAX > 65 statements :28 :09 > 250 statements 1:58 :55 :21 It's hard to tell what this means, because the 7300 times are significantly higher than for its ancestor-Miniframe, even at the same clock speed and using the same compiler. What I'd suspect is a high paging rate on a small system. I'd suggest redoing the experiment, doing the following: 1) Specify the size of the 7300. For whatever reasons, doing C compiles has a fierce knee inthe performance curve. 512K is tight, 1MB is OK. 2) What's OS on VAX? Be warned that the System V linker can often take 50-65% of the time on a minimal compile, whereas BSD linkers are not as powerful, but are faster. To measure this, try timing cc -c. 3) It's also worth reporting cpu as well as real. -- -john mashey UUCP: {decvax,ucbvax,ihnp4}!decwrl!mips!mash DDD: 415-960-1200 USPS: MIPS Computer Systems, 1330 Charleston Rd, Mtn View, CA 94043