home *** CD-ROM | disk | FTP | other *** search
/ NetNews Usenet Archive 1992 #26 / NN_1992_26.iso / spool / comp / arch / 10692 < prev    next >
Encoding:
Internet Message Format  |  1992-11-13  |  2.9 KB
  1. Path: sparky!uunet!noc.near.net!hri.com!snorkelwacker.mit.edu!bloom-picayune.mit.edu!athena.mit.edu!solman
  2. From: solman@athena.mit.edu (Jason W Solinsky)
  3. Newsgroups: comp.arch
  4. Subject: Re: More Alpha AXP Performance Numbers
  5. Message-ID: <1992Nov13.140907.2304@athena.mit.edu>
  6. Date: 13 Nov 92 14:09:07 GMT
  7. References: <1992Nov10.204139.5148@ryn.mro4.dec.com>
  8. Sender: news@athena.mit.edu (News system)
  9. Organization: Massachusetts Institute of Technology
  10. Lines: 35
  11. Nntp-Posting-Host: m4-035-18.mit.edu
  12.  
  13. In article <1992Nov10.204139.5148@ryn.mro4.dec.com>, bhandarkar@wrksys.enet.dec.com (Dileep Bhandarkar) writes:
  14. |> This is in response to e-mail requests for more benchmark data.
  15. |> 
  16. |>     Alpha AXP OpenVMS (version 1.0) SYSTEMS PERFORMANCE DATA 
  17. |> 
  18. |>  |---------------------------------------------------------------------------|
  19. |>  |METRIC              | 3000/400  | 3000/500 |   4000   | 7000      | 10000  |
  20. |>  |====================|===========|==========|==========|===========|========|
  21. |>  |SPECint89           |   65.8    |   73.5   |   83.7   |   95.1    | 104.5  |
  22. |>  |SPECfp89            |  150.6    |  169.9   |  188.4   |  244.2    | 268.6  |
  23. |>  |SPECmark89          |  108.1    |  121.5   |  136.2   |  167.4    | 184.1  |
  24. |>  |--------------------|------------------------------------------------------|
  25. |>  |SPECint92           |   65.3    |   74.3   |   83.5   |   96.6    | 106.5  |
  26. |>  |SPECfp92            |  112.2    |  126.0   |  143.1   |  182.1    | 200.4  |
  27. |>  |--------------------|------------------------------------------------------|
  28. |>  |Dhrystone(DMIPS)V1.1|  129.9    |  146.7   |  158.8   |  177.3    | 194.5  |
  29. |>  |--------------------|------------------------------------------------------|
  30. |>  |Linpack 100x100     |   26.4    |   30.2   |   36.3   |   38.6    |  42.5  |
  31. |>  |       1000x1000    |   70.8    |   79.9   |   86.4   |  102.1    | 111.6  |
  32. |>  |--------------------|------------------------------------------------------|
  33. |>  |Perfect BM Suite    |   18.1    |   20.4   |   22.9   |   26.0    |  28.6  |
  34. |>  | Mflops(Geom. Mean) |           |          |          |           |        |
  35. |>  |--------------------|------------------------------------------------------|
  36. |>  |DN&R Labs           |           |          |          |           |        |
  37. |>  |CPU2    (MVUPS)     |  159.1    |  179.7   |  193.2   |  217.9    | 239.3  |
  38. |>  |--------------------|------------------------------------------------------|
  39. |>  |Livermore Loop Mflop|   18.7    |   21.3   |   22.9   |   25.6    |  28.1  |
  40. |>  |--------------------|------------------------------------------------------|
  41.  
  42. Based on this data, it is extraordinarily unclear why somebody would be willing
  43. to go for the 10000 box when the 3000 delivers better than half the performance
  44. at a TINY fraction of the price. Also, why aren't there any multi-processing
  45. configurations available?
  46.  
  47. Jason W. Solinsky
  48.