| Title: | Alpha Support Conference |
| Notice: | This is a new Alphanotes, please read note 2.2 |
| Moderator: | VAXAXP::BERNARDO |
| Created: | Thu Jan 02 1997 |
| Last Modified: | Fri Jun 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 128 |
| Total number of notes: | 617 |
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 26.1 | Look att www.tpc.org | HERON::BLOMBERG | Trapped inside the universe | Tue Jan 21 1997 07:29 | 15 |
| 26.2 | COL01::VSEMUSCHIN | Duck and Recover ! | Tue Jan 21 1997 07:50 | 14 | |
| 26.3 | bandwidth intro, spec scaling... | PERFOM::HENNING | Tue Jan 21 1997 07:52 | 20 | |
| 26.4 | Apples & Oranges?? | TROOA::GILLAM | Mon Apr 28 1997 14:47 | 25 | |
I have a customer with a floating point intense, (chemical) process
optimization application which currently takes about an hour to run on
a DEC 3000 Model 800. The goal is to find a new platform in the same
class that can run the application in under 15 minutes.
Now setting the application aside and just looking at what's available
in the mid-range over 400MHz systems there are some interesting
comparisons when looking at SPECfp95 numbers.
System Cache Memory SPECfp95
AlphaServer 1000 5/500 8 512 17.8
AlphaServer 4100 5/466 4 128 19.2
AlphaStation 500 5/500 8 128 20.4
I am admittedly an novice, but I would of thought the 1000 with 4 x the
memory (and twice the cache of the 4100) would've blown the doors off
of either of the other systems, but didn't. Is the 1000A's internal
I/O the bottleneck (this question is for my own personal interest)?
Based on these numbers I would guess if the 4100 had a 500MHz CPU it
would be the SPECfp95 performance leader.
Bill
| |||||
| 26.5 | DECCXL::OUELLETTE | mudseason into blackfly season | Mon Apr 28 1997 15:24 | 4 | |
The bottleneck on the 1000A is interface to main memory. SPEC does no appreciable I/O. In fact SPEC95's benchmarks are chosen to fit into 64MB of memory, so the additional memory on the 1000A is of no help whatsoever. | |||||