David Wang (dwang@realworldtech.com) on 6/20/03 wrote:
---------------------------
>Bill Todd (billtodd@metrocast.net) on 6/20/03 wrote:
>---------------------------
>>David Wang (dwang@realworldtech.com) on 6/19/03 wrote:
>>---------------------------
>>>Bill Todd (billtodd@metrocast.net) on 6/19/03 wrote:
>>>---------------------------
>
>>>>So if I managed to get past all that gushing appreciation to the heart of the matter,
>>>>it sounds as if your analysis predicts a SPECint_base score for the 1.5 GHz Madison
>>>>of something like 1175 (assuming unchanged compiler technology). I just keep thinking
>>>>back to Arcadian's glowing assurances that Madison would beat the bejazus out of
>>>>Opteron and wondering if it can really be that low.
>>>
>>>Did Arcadian predict that Madison would beat the "bejezus" out of Opteron in SPECxxx?
>>
>>While it was difficult to pin him down to specific claims of SPECint superiority,
>>he said "It's not exactly part of my job, but it's closely enough related so that
>>I can get the skinny. I can't get into any details, but I feel pretty safe about
>>Madison outperforming any competitive CPU in 2003, with the possible exception of
>>Power5." Why he thought that POWER5 would be available this year I don't know,
>>but the statement would certainly seem to include not only the Opterons available
>>now but those available through year's end - and it was made in the specific *context*
>>of a SPECint discussion (though honesty compels me to admit that my own projections
>>for EV7's likely SPECint performance in another part of the same discussion turned
>>out to have been very overoptimistic as well).
>
>I think I recall the thread now.
>
>If he just said "outperforming competitive CPU in 2003" (in SPECint), that appears
>to be a relatively mild statement rather than the more boastful "beat the bejazus out of" (in SPECint).

Indeed - my memory embellished the degree of emphasis (as I just noted in another post).

>
>AMD's best SPECint numbers as reported on spec.org seems to be the 1095 base and
>1170 peak numbers. Beating those numbers would also seem to be well within reach
>of the 1.5 GHz Madison, and that would support Arcadian's "modest claim" of "beating".

No, since they were made in the immediate context of Fred Weber's report of a SPECint score of 1202 for the 2 GHz part: that's the score Madison needs to beat for Arcadian's statement to stand (and with a bit of compiler advance it well may) - and even then it needs to remain standing for the next six months.

>At least until AMD could ship 2+ GHz Opterons and/or perhaps get better compiler support to pull those numbers up.

Exactly. And in this case advances in compilation would seem to be at least as applicable to Hammer 64-bit performance as they are so often claimed to be to getting EPIC up to speed.

- bill