Re: Pro*C compile hogs my system

From: Nick Price <nick_at_falstaff>
Date: 2 Nov 1994 10:19:22 GMT
Message-ID: <397p3b$gcs_at_news.INbe.net>


Steve Edwards (steve_at_tmw.com) wrote:
: I'm trying to develop Oracle Pro*C programs on UnixWare using the GNU
: compiler (version 2.3.3). This is on a 90Mhz Pentium with 16mb RAM.

So am I - same platform - same OS - same compiler.

: My problem is that during the compile (which takes just under 5
: minutes), gcc takes over the entire system for about 1.5 minutes. Even
: keystrokes are not echoed! Beyond what this is doing to my productivity,
: it's making me very unpopular with my co-workers. It's going to get even
: worse when we get another Pro*C programmer.
 

: Am I that short of memory or is this just a fact of life? If it is some
: resource shortage, how do I determine the resource and how can I get
: objective numbers to show my boss?

You said it - 16MB is badly underconfigured for the task. If you add things up you will find your UNIX kernel + system processes will take at least 8MB, and the ORACLE instance including its shared memeory area (the SGA) will take a minimum of 8MB too. So you have no memory to work in, the system spends all its time swapping, and your users start getting mad !

I use 32MB RAM on my system, and its still not enough. I believe 64MB is more reasonable. In these days of relatively cheap RAM, it is crazy to cripple your system for the sake of a thousand dollars.

Another point though - unless its absolutely necessary, its not a good idea to mix development and production environments on the same system.

If you want proof of the system swapping, you can use 'sar' or get a copy of 'svr4mon'.

Nick

--
+==============================================================================+
+    Nick Price                                                                +
+    Internet: nprice_at_innet.be                                                 +
+    Tel/Fax : +32 2 347 1101                                                  +
+==============================================================================+
Received on Wed Nov 02 1994 - 11:19:22 CET

Original text of this message