Oh - to discover a real compiler problem

IWETHEY v. 0.3.0 | TODO

1,095 registered users | 0 active users | 0 LpH | Statistics
Login | Create New User

Welcome to IWETHEY!

Post #73,075

by deSitter

1/6/03 9:36:04 PM

Oh - to discover a real compiler problem

..is real cool! Once I had to write a big string processing program in FORTRAN (!!) so I stumbled across a problem with the IBM VS FORTRAN compiler regarding character data - man those guys hopped to it! A tape with the fixed compiler on it was in the lab the next day. It was a real "code red" for the compiler people - I was stoked about that :)

-drl

Post #73,093

by admin

1/6/03 10:18:20 PM

This is not considered a compiler problem by the GCC folks.

Basically, it is this:

IEEE designates that doubles should be held as 64-bits in memory. The Intel FPU, however, has 80 bits available for calculations. So it comes down to 1) being aware of the platform on which you are running, and 2) using a programming strategy that ensures the FPU is used as you need it to be used.

For many applications, the extra 16 bits of precision is a good thing. For some applications, such as mine, a much greater control over rounding and precision is needed, and means are provided for making use of what you need to use.

Regards,

-scott anderson

"Welcome to Rivendell, Mr. Anderson..."

C decimal question: - (admin) - (53) - Jan. 3, 2003, 05:53:14 PM EST

Equality on Real/Double is always a crapshoot - (tablizer) - Jan. 3, 2003, 06:04:04 PM EST

Don't know... - (Simon_Jester) - Jan. 3, 2003, 06:04:45 PM EST

Re: C decimal question: - (deSitter) - (1) - Jan. 3, 2003, 09:26:50 PM EST

Makes no difference. - (admin) - Jan. 6, 2003, 10:15:39 AM EST

The boolean compare operators - (ChrisR) - (36) - Jan. 4, 2003, 02:12:52 AM EST

I've done the subtract - (admin) - (35) - Jan. 4, 2003, 08:50:00 AM EST

Assembly dump? - (ChrisR) - (22) - Jan. 4, 2003, 12:57:27 PM EST

I did my dump with GDB - (Arkadiy) - Jan. 5, 2003, 02:11:21 PM EST

What I get: - (admin) - (20) - Jan. 6, 2003, 10:34:38 AM EST

Speculating... - (ChrisR) - (19) - Jan. 6, 2003, 10:49:00 AM EST

gcc 2.95.4 results - (admin) - (18) - Jan. 6, 2003, 11:06:04 AM EST

Re: gcc 2.95.4 results - (deSitter) - (1) - Jan. 6, 2003, 11:35:48 AM EST

EBP is base of stack frame - (Arkadiy) - Jan. 6, 2003, 01:18:01 PM EST

I86 Assembly is not my specialty... - (ChrisR) - Jan. 6, 2003, 12:30:53 PM EST

ASM Comments - (ChrisR) - (13) - Jan. 6, 2003, 02:12:59 PM EST

Errrr... - (admin) - (12) - Jan. 6, 2003, 03:08:52 PM EST

Getting ASM from VC6 - (deSitter) - (11) - Jan. 6, 2003, 04:06:35 PM EST

Results - (deSitter) - (10) - Jan. 6, 2003, 04:25:49 PM EST

Request... - (ChrisR) - (9) - Jan. 6, 2003, 04:34:54 PM EST

Re: Request... - (deSitter) - (8) - Jan. 6, 2003, 04:36:04 PM EST

Trimming it down... - (ChrisR) - (7) - Jan. 6, 2003, 04:57:34 PM EST

Re: Trimming it down... - (deSitter) - (6) - Jan. 6, 2003, 04:57:55 PM EST

Scratches head... - (ChrisR) - (5) - Jan. 6, 2003, 06:07:55 PM EST

Re: Scratches head... - (deSitter) - (4) - Jan. 6, 2003, 08:51:33 PM EST

Thanks... - (ChrisR) - (3) - Jan. 6, 2003, 09:27:16 PM EST

BTW - (deSitter) - (2) - Jan. 6, 2003, 09:29:58 PM EST

Hmmmm - (ChrisR) - (1) - Jan. 6, 2003, 10:10:48 PM EST

puts("42"); -NT - (deSitter) - Jan. 6, 2003, 10:21:20 PM EST

gcc 2.95.26 - (ChrisR) - Jan. 6, 2003, 06:13:49 PM EST

Did you try it with optimizations off? -NT - (deSitter) - (7) - Jan. 4, 2003, 01:12:28 PM EST

A really good optimizing compiler... - (ChrisR) - (6) - Jan. 4, 2003, 01:21:38 PM EST

Yes.. - (deSitter) - Jan. 4, 2003, 01:35:10 PM EST

Compiler has no way to predict how pow() is implemented - (Arkadiy) - (4) - Jan. 5, 2003, 02:13:15 PM EST

IEEE - (deSitter) - (2) - Jan. 5, 2003, 02:23:03 PM EST

pow() is library, not hardware - (Arkadiy) - (1) - Jan. 5, 2003, 07:14:41 PM EST

True - (deSitter) - Jan. 5, 2003, 07:38:57 PM EST

Re: Compiler has no way to predict how pow() is implemented - (deSitter) - Jan. 5, 2003, 08:57:17 PM EST

An option of interest? - (ChrisR) - (3) - Jan. 4, 2003, 01:13:29 PM EST

I'll look into that. Datum: gcc 3.2.1 on glibc 2.3.1 -NT - (admin) - Jan. 4, 2003, 05:56:33 PM EST

Doesn't seem to make any difference. - (admin) - (1) - Jan. 6, 2003, 10:12:59 AM EST

Re: Doesn't seem to make any difference. - (deSitter) - Jan. 6, 2003, 11:36:57 AM EST

Not a direct answer, but still on-topic. - (static) - Jan. 4, 2003, 04:52:08 AM EST

RESOLUTION: - (admin) - (10) - Jan. 6, 2003, 07:34:23 PM EST

A caveat from Borland: - (a6l6e6x) - (6) - Jan. 6, 2003, 08:14:49 PM EST

This was slightly different. - (admin) - (5) - Jan. 6, 2003, 08:17:18 PM EST

Re: This was slightly different. - (deSitter) - Jan. 6, 2003, 09:11:32 PM EST

Oh - to discover a real compiler problem - (deSitter) - (1) - Jan. 6, 2003, 09:36:04 PM EST

This is not considered a compiler problem by the GCC folks. - (admin) - Jan. 6, 2003, 10:18:20 PM EST

Scott, that's a bug worthy of catching - (Simon_Jester) - (1) - Jan. 6, 2003, 10:34:21 PM EST

Fun to track down, at least. - (admin) - Jan. 6, 2003, 10:37:27 PM EST

Re: RESOLUTION: - (deSitter) - (2) - Jan. 6, 2003, 09:03:59 PM EST

Historically speaking... - (ChrisR) - (1) - Jan. 6, 2003, 09:53:48 PM EST

Re: Historically speaking... - (deSitter) - Jan. 6, 2003, 10:04:39 PM EST

iwethey.org

Drop to the deck and flop like a fish!
334 ms