Article: Wolves in CISC Clothing
By: Paul DeMone (pdemone.delete@this.igs.net), January 7, 2004 10:34 am
Room: Moderated Discussions
José Javier Zarate (jzaratae@unav.es) on 1/7/04 wrote:
---------------------------
>Only two comments:
>
>On Transmeta
>I seem to recall that when Transmeta started announcing its processor they suggested
>the possibility of having an "universal processor" and that they would develop different
>CMSs as needed-demanded by the market. But Mr. DeMone seems to state the opposite
>in the last paragraph of the fourth page of his article.
I didn't say other ISAs couldn't be supported. I meant
that other ISAs couldn't be supported as efficiently
as x86. The Crusoe hardware isn't neutral, it is highly
organized around emulating x86. The Crusoe is far from
being a "universal processor".
Can a Crusoe support MC88000 comparison result generation
as efficiently as x86 status code generation? Can Crusoe
support VAX FP formats and semantics as efficiently as it
does x87?
>
>
>About IA-32EL:
>
>I read an article about HP's Dynamo, very similar internal working only for PA-RISC/HP-UX
>native programs. It appeared to me that Dynamo was much more aggressive when marking
>a block as "Hot". Also I ask myself if IA-64 could benefit from an EL devoted to its native programs.
>
>It would be very nice if any publically available documents on IA-32 EL were posted.
www.microarch.org/micro36/html/pdf/ goldenberg-IA32ExecutionLayer.pdf
---------------------------
>Only two comments:
>
>On Transmeta
>I seem to recall that when Transmeta started announcing its processor they suggested
>the possibility of having an "universal processor" and that they would develop different
>CMSs as needed-demanded by the market. But Mr. DeMone seems to state the opposite
>in the last paragraph of the fourth page of his article.
I didn't say other ISAs couldn't be supported. I meant
that other ISAs couldn't be supported as efficiently
as x86. The Crusoe hardware isn't neutral, it is highly
organized around emulating x86. The Crusoe is far from
being a "universal processor".
Can a Crusoe support MC88000 comparison result generation
as efficiently as x86 status code generation? Can Crusoe
support VAX FP formats and semantics as efficiently as it
does x87?
>
>
>About IA-32EL:
>
>I read an article about HP's Dynamo, very similar internal working only for PA-RISC/HP-UX
>native programs. It appeared to me that Dynamo was much more aggressive when marking
>a block as "Hot". Also I ask myself if IA-64 could benefit from an EL devoted to its native programs.
>
>It would be very nice if any publically available documents on IA-32 EL were posted.
www.microarch.org/micro36/html/pdf/ goldenberg-IA32ExecutionLayer.pdf
| Topic | Posted By | Date |
|---|---|---|
| New Silicon Insider article | Dean Kent | 2004/01/05 11:39 PM |
| New Silicon Insider article | Michael Menietti | 2004/01/06 01:18 AM |
| New Silicon Insider article | Paul DeMone | 2004/01/07 10:28 AM |
| New Silicon Insider article | Singh, S.R. | 2004/01/06 01:41 AM |
| New Silicon Insider article | José Javier Zarate | 2004/01/06 11:03 AM |
| New Silicon Insider article | Tim Hughes | 2004/01/06 06:41 AM |
| Thanks, glad you liked it (NT) | Paul DeMone | 2004/01/07 10:35 AM |
| New Silicon Insider article | José Javier Zarate | 2004/01/06 10:59 AM |
| CMS/IA-32EL/Dynamo | José Javier Zarate | 2004/01/07 05:44 AM |
| CMS/IA-32EL/Dynamo | Paul DeMone | 2004/01/07 10:34 AM |
| CMS/IA-32EL/Dynamo | José Javier Zarate | 2004/01/07 12:14 PM |
| CMS/IA-32EL/Dynamo | Paul DeMone | 2004/01/07 01:21 PM |
| CMS/IA-32EL/Dynamo | PiedPiper | 2004/01/12 01:57 PM |
| CMS/IA-32EL/Dynamo | anon y pus | 2004/01/07 05:48 PM |
| New Silicon Insider article | José Javier Zarate | 2004/01/10 12:34 AM |
| New Silicon Insider article | PiedPiper | 2004/01/12 01:52 PM |
| New Silicon Insider article | Anonymous | 2004/01/17 04:52 PM |
| New Silicon Insider article | FR | 2004/01/13 01:48 AM |