By: Singh, S.R. (swaranNOrajSPAMsinghPLEASE.delete@this.hotmail.com), February 10, 2003 9:34 pm
Room: Moderated Discussions
William L. (wle4c@yahoo.com) on 2/10/03 wrote:
---------------------------
>Could you give a little more explanation as to how IBM's design is more automated
>than other companies? You may have to dumb it down a little for me as I am not
>quite as well versed as most here.
In short, IBM uses cell libraries as opposed to full-custom circuit design to design their MPUs, including a heavier reliance on automated placement & routing tools in order to complete the chip layout.
What this means is that IBM develops standard cell libraries for commonly-used circuits which the HDL compiler will use when synthesizing a netlist. This approach isn't used in all parts of the chip (ie: IBM has claimed this approach is good for adders and comparators, but poor for register files), though, giving IBM the ability to optimize some of the more important paths in the chip.
---------------------------
>Could you give a little more explanation as to how IBM's design is more automated
>than other companies? You may have to dumb it down a little for me as I am not
>quite as well versed as most here.
In short, IBM uses cell libraries as opposed to full-custom circuit design to design their MPUs, including a heavier reliance on automated placement & routing tools in order to complete the chip layout.
What this means is that IBM develops standard cell libraries for commonly-used circuits which the HDL compiler will use when synthesizing a netlist. This approach isn't used in all parts of the chip (ie: IBM has claimed this approach is good for adders and comparators, but poor for register files), though, giving IBM the ability to optimize some of the more important paths in the chip.
| Topic | Posted By | Date |
|---|---|---|
| New Article Available | David Kanter | 2003/02/02 01:44 AM |
| Excellent Article, Paul (NT) | Arcadian | 2003/02/02 03:57 PM |
| Yikes! Slashdotted! | Dean Kent | 2003/02/02 06:53 PM |
| Yikes! Slashdotted! | Singh, S.R. | 2003/02/02 11:38 PM |
| Yikes! Slashdotted! | Dean Kent | 2003/02/02 11:53 PM |
| Yikes! Slashdotted! | Anonymous | 2003/02/03 06:03 AM |
| Yikes! Slashdotted! | Dean Kent | 2003/02/03 08:13 AM |
| FYI, Paul: | Peter Gerassimoff | 2003/02/03 12:38 AM |
| FYI, Paul: | Bill Todd | 2003/02/03 02:18 AM |
| FYI, Paul: | Marc M. | 2003/02/03 09:08 AM |
| Opteron SPEC Performance | Arcadian | 2003/02/03 11:15 AM |
| Opteron SPEC Performance | Marc M. | 2003/02/03 11:18 AM |
| i want to see those links pls!! (NT) | waitressInGaza | 2003/02/03 12:15 PM |
| i want to see those links pls!! (NT) | José Javier Zarate | 2003/02/05 08:23 AM |
| i want to see those links pls!! (NT) | waitressInGaza | 2003/02/05 01:50 PM |
| i want to see those links pls!! (NT) | tecate | 2003/02/05 04:09 PM |
| gender should be irrelevant | waitressInGaza | 2003/02/05 05:36 PM |
| gender should be irrelevant | Dean Kent | 2003/02/05 06:03 PM |
| gender should be irrelevant | tecate | 2003/02/05 08:27 PM |
| well it is all about your viewpoint | waitressInGaza | 2003/02/05 11:44 PM |
| well it is all about your viewpoint | doriangrey | 2003/02/06 11:39 PM |
| Please be a girl. :p | NIKOLAS | 2003/02/06 05:28 AM |
| FYI, Paul: | Paul DeMone | 2003/02/03 07:43 AM |
| larger cache for POWER4+? | Anil Maliyekkel | 2003/02/03 04:08 AM |
| larger cache for POWER4+? | Thu Nguyen | 2003/02/03 05:35 AM |
| well written indeed | hobold | 2003/02/03 06:41 AM |
| well written indeed | tecate | 2003/02/03 08:05 AM |
| well written indeed | hobold | 2003/02/03 08:12 AM |
| well written indeed | Paul DeMone | 2003/02/03 08:15 AM |
| well written indeed | hobold | 2003/02/03 11:04 AM |
| well written indeed | Paul DeMone | 2003/02/03 11:58 AM |
| about Jim Keller | Marc M. | 2003/02/03 12:50 PM |
| about Jim Keller | Paul DeMone | 2003/02/03 01:22 PM |
| Nope but will now (NT) | Marc M. | 2003/02/03 01:50 PM |
| patent info... | Marc M. | 2003/02/03 01:52 PM |
| about the article... | Dean Kent | 2003/02/03 03:49 PM |
| Er, I just have to point something out. | Anonymous | 2003/02/04 09:19 PM |
| Next time check the date. | Paul DeMone | 2003/02/05 12:10 AM |
| Yes, I know it was written in 2000. | Anonymous | 2003/02/05 11:42 AM |
| Yes, I know it was written in 2000. | Paul DeMone | 2003/02/05 12:20 PM |
| about Jim Keller | Alejandro G. Belluscio | 2003/02/09 12:09 PM |
| about Jim Keller | Dean Kent | 2003/02/09 12:40 PM |
| about Jim Keller | Paul DeMone | 2003/02/09 01:00 PM |
| about Jim Keller | David Wang | 2003/02/09 01:18 PM |
| about Jim Keller | Paul DeMone | 2003/02/09 02:47 PM |
| about Jim Keller | Interested in Illinois | 2003/02/10 10:11 AM |
| about Jim Keller | Dean Kent | 2003/02/10 11:06 AM |
| about Jim Keller | Alejandro Belluscio | 2003/02/11 05:53 PM |
| about Jim Keller | Dean Kent | 2003/02/11 08:52 PM |
| about Jim Keller | Alejandro Belluscio | 2003/02/12 10:06 AM |
| about Jim Keller | Interested in Illinois | 2003/02/13 03:19 PM |
| about Jim Keller | Interested in Illinois | 2003/02/11 10:35 PM |
| about Jim Keller | David Kanter | 2003/02/11 11:24 PM |
| Taxation systems | Alejandro Belluscio | 2003/02/12 09:54 AM |
| Taxation systems | Interested in Illinois | 2003/02/13 06:23 AM |
| Taxation systems | Alejandro Bellusco | 2003/02/13 09:53 AM |
| Taxation systems | Interested in Illinois | 2003/02/13 02:47 PM |
| Taxation systems | Aleajdnro G. Belluscio | 2003/02/13 04:56 PM |
| Taxation systems | Jouni Osmala | 2003/02/14 07:03 AM |
| Taxation systems | Alejandro G. Belluscio | 2003/02/14 05:13 PM |
| Taxation systems | David Kanter | 2003/02/14 11:18 PM |
| Taxation systems | Alejandro Belluscio | 2003/02/15 07:36 AM |
| well written indeed | hobold | 2003/02/03 01:40 PM |
| E8870 Chipset | José Javier Zarate | 2003/02/04 02:19 PM |
| E8870 Chipset | Arcadian | 2003/02/05 12:54 AM |
| E8870 Chipset | José Javier Zarate | 2003/02/05 08:13 AM |
| E8870 Chipset | Arcadian | 2003/02/05 11:49 AM |
| Minor question | José Javier Zarate | 2003/02/04 02:24 PM |
| well written indeed | Dean Kent | 2003/02/03 08:18 AM |
| well written indeed | hobold | 2003/02/03 08:36 AM |
| well written indeed | Dean Kent | 2003/02/03 09:33 AM |
| hear hear... | Marc M. | 2003/02/03 11:16 AM |
| well written indeed | hobold | 2003/02/03 02:02 PM |
| well written indeed | Dean Kent | 2003/02/03 03:48 PM |
| well written indeed | hobold | 2003/02/04 12:56 PM |
| well written indeed | Marc M. | 2003/02/03 09:19 AM |
| BTW Paul, it was VERY well written... BUT... =) | Marc M. | 2003/02/03 09:25 AM |
| Ditto (NT) | William L. | 2003/02/10 07:28 PM |
| New Article Available | JS | 2003/02/05 01:09 AM |
| Changes in MPU design methodologies? | Richard Stacpoole | 2003/02/06 04:46 AM |
| Changes in MPU design methodologies? | Paul DeMone | 2003/02/06 09:47 AM |
| Changes in MPU design methodologies? | doriangrey | 2003/02/06 11:41 PM |
| Changes in MPU design methodologies? | William L. | 2003/02/10 07:34 PM |
| Changes in MPU design methodologies? | Singh, S.R. | 2003/02/10 09:34 PM |
| Watch this | Singh, S.R. | 2003/02/10 09:42 PM |
| excellent article, just a nit or two | mulp | 2003/02/15 11:11 PM |