| Title: | Hudson VLSI |
| Notice: | For Digital Chip Data - CHIPBZ::PRODUCTION$:[DS_INFO...] |
| Moderator: | RICKS::PHIPPS |
| Created: | Wed Feb 12 1986 |
| Last Modified: | Fri Jun 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 701 |
| Total number of notes: | 4658 |
Information regarding the planned performance level for Merced is starting to leak through speaches and customer/partners being briefed by HP and Intel. One technical journalist (Jan Sandred at Datateknik) who was at Compaq Innovation 97 heard Andy Grove giving the following data points: Launch 1999 40 milj transistors 900 mhz Voltage >1V 2500 MIPS 60 SPECint I also have 2500 MFLOPS from an HP partner working in High Performance Computing. This is for the high performance variantion aimed at servers, they will make some variants for the desktop with less complexity and performance. Sound very much like an EV6 type of chip made in 0.18 micron with the added transistor budget for the x86 translator and maybe som PA-RISC translation as well. Given that we should be at 60 SPECint/90 SPECfp with EV67 by sometime 1998 and 2/3 towards EV7 by 1999 Alpha could well remain the hing end alternative to Intel even after Merced hits the streets. It's time to start challening the Merced myth that Alpha will end it's story when the big unknown arrives. Let's not make the mistake and let them get away with it. What clould we lose by challenging this alrady now? Nothing. Bj�rn Bj�rn [Posted by WWW Notes gateway]
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 699.1 | BGSDEV::POEGEL | Thu May 29 1997 13:36 | 7 | ||
>> <<< Note 699.0 by NNTPD::"[email protected]" "Bjorn Fehrm" >>> >> -< Merced not that impressive >- Anyone see a floating point estimate? Garry | |||||
| 699.2 | YIELD::HARRIS | Thu May 29 1997 15:53 | 4 | ||
> Anyone see a floating point estimate?
They keep getting floating point errors when the try an calculate it.
| |||||
| 699.3 | GLDX02::ALLBERY | Jim | Thu May 29 1997 16:15 | 3 | |
re: Anyone see a floating point estimate?
.0 had an indirect reference to 2500 MFLOPs.
| |||||
| 699.4 | MFLOPS probably theoretical... | CERN::HOBBS | Congrats to the Ignoble Peace Prize winner! (http://www.eecs.harvard.edu/ig_nobel) | Fri May 30 1997 04:04 | 7 |
> 900 mhz > 2500 MIPS > 2500 MFLOPS Gee, 833 MHz with 3 integer units and 3 FP units would give 2500 MIPS/MFLOPS SPECfp estimates would be more useful | |||||