Jump to content

Talk:Nord-100

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Is the numbers for Whetstone MWIPS correct for the ND-100 / ND-110 comparision? The later improved processor have a lower rating and it seems odd. A citation needed is in place. ScandinavianRockguy (talk) 02:48, 7 February 2009 (UTC)[reply]

ND100/110 comparison

[edit]

Inserted fact tags in one ND100/110 column each. Citation needed for the figures, in particular MWIPS. Minimum number of ucycles per instruction and ucycle time point to a factor 4.5 improvement of performance (namely (150ns/100ns) * (3cycles/1cycle)) in case the ucycles/instruction ratio would be typical. So the 110 should be up to a factor 4.5 faster than the 100. Now MWIPS seems to be a float benchmark, so how did they compare those - FP emulation by ucode in the 110 case vs hard FP for the 100? And is the MWIPS figure really representative of performance for the typical application/system load? The calculation of these figures needs to be cited and explained, just accepting this w.o. the benchmark setup and conditions makes it pointless. If this is to be a performance comparison it needs more work, otherwise MWIPS figures to be deleted as meaningless. The whole article needs citations. 70.137.148.19 (talk) 03:50, 11 May 2009 (UTC)[reply]