1/1
instruction set problem on openrisc
by Unknown on Jan 29, 2004 |
Not available! | ||
Hi all,
There are two classes of instruction in the openrisc, class I and class II. If a modified openrisc just implments class I instruction (for saving resource), how can I avoid to generated the instruction of class II in the compiler toolchain. I means if there are any options for avoiding to generate the class II instructions. One more question, If I want to add custom instruction set e.g. multiply - accumulate in to the openrisc, is there any easy way to make the complier to generate the instruction set except modifying the compiler? Thank you very much. Best Regards, Stephen |
instruction set problem on openrisc
by Unknown on Jan 29, 2004 |
Not available! | ||
----- Original Message -----
From: whli_interqos@yahoo.com.hk>
To: openrisc@opencores.org>
Sent: Thursday, January 29, 2004 5:11 AM
Subject: [openrisc] instruction set problem on openrisc
Hi all,
There are two classes of instruction in the openrisc, class I and class II. If a modified openrisc just implments class I instruction (for saving resource), how can I avoid to generated the instruction of class II in the compiler toolchain. I means if there are any options for avoiding
to generate the class II instructions.
Class II instructions by default are not generated by compiler. One more question, If I want to add custom instruction set e.g. multiply - accumulate in to the openrisc, is there any easy way to make the complier to generate the instruction set except modifying the compiler? That;s the only way. You have to modify compiler to generate your custom insns. regards, Damjan
Thank you very much.
Best Regards,
Stephen
_______________________________________________
http://www.opencores.org/mailman/listinfo/openrisc
|
instruction set problem on openrisc
by Unknown on Jan 29, 2004 |
Not available! | ||
> One more question, If I want to add custom instruction set e.g.
> multiply - accumulate in to the openrisc, is there any easy way to make > the complier to generate the instruction set except modifying the > compiler? That;s the only way. You have to modify compiler to generate your custom insns. You can also make #define MAC(x,y), which will generate via asm statement "word 0xblabla" instruction. Marko |
instruction set problem on openrisc
by Unknown on Jan 30, 2004 |
Not available! | ||
Hi Marko,
Thank you for the advice.
I put the the statement,#define MAC(x,y), into my source code. But it seems that there is not multiply-accumulate instruction in the asm code. Did I do anything wrong? Can you give more details on it? Thank you very much.
Regards,
Stephen
Marko Mlinar markom@opencores.org> wrote:
You can also make #define MAC(x,y), which will generate via asm statement
"word 0xblabla" instruction.
Marko
_______________________________________________
http://www.opencores.org/mailman/listinfo/openrisc
Shining Friends¡B¦n¤ß¦n³ø¡B·³¤ë¦pºq...
®öº©¹aÃn ±¡¤ß³sô
http://ringtone.yahoo.com.hk/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.opencores.org/forums/openrisc/attachments/20040130/efedb44e/attachment.htm
> One more question, If I want to add custom instruction set e.g.
> multiply - accumulate in to the openrisc, is there any easy way to make > the complier to generate the instruction set except modifying the > compiler? That;s the only way. You have to modify compiler to generate your custom insns. |
instruction set problem on openrisc
by Unknown on Jan 30, 2004 |
Not available! | ||
I think Marko was suggesting that when you have implemented your own custom
multiply-accumulate in your own special version of the CPU, you can access
it from C code by using a #define. In this way, you don't have to modify the
compiler. As you rightly say, there is no MAC instruction in the current
instruction set.
Robert Cragie, Design Engineer
_______________________________________________________________
Jennic Ltd, Furnival Street, Sheffield, S1 4QT, UK
http://www.jennic.com Tel: +44 (0) 114 281 2655
_______________________________________________________________
-----Original Message-----
From: openrisc-bounces@opencores.org
[mailto:openrisc-bounces@opencores.org]On Behalf Of li stephen
Sent: 30 January 2004 03:28
To: List about OpenRISC project
Subject: Re: [openrisc] instruction set problem on openrisc
Hi Marko,
Thank you for the advice.
I put the the statement,#define MAC(x,y), into my source code.
But it seems that there is not multiply-accumulate instruction in the asm
code. Did I do anything wrong? Can you give more details on it? Thank you
very much.
Regards,
Stephen
Marko Mlinar markom@opencores.org> wrote:
> > One more question, If I want to add custom instruction set e.g.
> > multiply - accumulate in to the openrisc, is there any easy way to
make
> > the complier to generate the instruction set except modifying the
> > compiler?
>
> That;s the only way. You have to modify compiler to generate your
custom
> insns.
You can also make #define MAC(x,y), which will generate via asm
statement
"word 0xblabla" instruction.
Marko
_______________________________________________
http://www.opencores.org/mailman/listinfo/openrisc
Shining Friends¡B¦n¤ß¦n³ø¡B·³¤ë¦pºq...
®öº©¹aÃn ±¡¤ß³sô
http://ringtone.yahoo.com.hk/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.opencores.org/forums/openrisc/attachments/20040130/b813ad2b/attachment.htm
|
instruction set problem on openrisc
by Unknown on Jan 30, 2004 |
Not available! | ||
There is l.mac in the current insn set, but it doesn't get generated by the
compiler. I tried twice to have it generated by compiler and compiler just
didn't want to emit the l.mac insn. If somebody else can figure out how to
generated it directly with the compiler (in the or32.md machine description
file) it would be great.
regards,
Damjan
----- Original Message -----
From: "Robert Cragie" rcc@jennic.com>
To: "List about OpenRISC project" openrisc@opencores.org>
Sent: Friday, January 30, 2004 9:27 AM
Subject: RE: [openrisc] instruction set problem on openrisc
I think Marko was suggesting that when you have implemented your own
custom
multiply-accumulate in your own special version of the CPU, you can access
it from C code by using a #define. In this way, you don't have to modify the
compiler. As you rightly say, there is no MAC instruction in the current
instruction set.
Robert Cragie, Design Engineer
_______________________________________________________________
Jennic Ltd, Furnival Street, Sheffield, S1 4QT, UK
http://www.jennic.com Tel: +44 (0) 114 281 2655
_______________________________________________________________
-----Original Message-----
From: openrisc-bounces@opencores.org
[mailto:openrisc-bounces@opencores.org]On Behalf Of li stephen
Sent: 30 January 2004 03:28
To: List about OpenRISC project
Subject: Re: [openrisc] instruction set problem on openrisc
Hi Marko,
Thank you for the advice.
I put the the statement,#define MAC(x,y), into my source code.
But it seems that there is not multiply-accumulate instruction in the asm
code. Did I do anything wrong? Can you give more details on it? Thank you
very much.
Regards,
Stephen
Marko Mlinar markom@opencores.org> wrote:
> > One more question, If I want to add custom instruction set e.g.
> > multiply - accumulate in to the openrisc, is there any easy way to make
> > the complier to generate the instruction set except modifying the
> > compiler? > > That;s the only way. You have to modify compiler to generate your custom
> insns.
You can also make #define MAC(x,y), which will generate via asm
statement
"word 0xblabla" instruction.
Marko
_______________________________________________
http://www.opencores.org/mailman/listinfo/openrisc
Shining Friends¡B¦n¤ß¦n³ø¡B·³¤ë¦pºq...
®öº©¹aÃn ±¡¤ß³sô
http://ringtone.yahoo.com.hk/
---------------------------------------------------------------------------- ----
_______________________________________________
http://www.opencores.org/mailman/listinfo/openrisc
|
1/1