HiPE and SIGFPE

Joel Reymont joelr1@REDACTED
Tue Aug 15 12:15:05 CEST 2006


I saw this happening when the SSE2 instruction wasn't being skipped.  
The exception flags are cleared and the FPU evaluates the same  
instruction again. I don't have Linux/AMD64 to fix this.

On Aug 15, 2006, at 8:43 AM, Anton Berezin wrote:

> On a related note, there seems to be something wrong with FP exception
> handling code on amd64 (I am using FreeBSD port of Erlang, r11b0).
> Basically,
>
> 1/0.
>
> just hangs eating 100% CPU.  It produces an exception, as expected,  
> on i386.
>
> Is there more information you might want to debug the problem?
>
> \Anton.

--
http://wagerlabs.com/








More information about the erlang-questions mailing list