#RapidCAD
7 messages in this thread
I noticed the article in PC Week on page 30 about RapidCAD Engineering
Coprocesser Chip from Intel. Anyone out there seen this puppy run with
AutoCAD?
Hi Tommy, Intel has tested RapidCad on Autocad Release 11 using the BYTE
Benchmarks and have found the following results in comparison to the
i387DX-33.
On redraw Rapidcad is 8% faster
On PAN Rapidcad is 28% faster
On ZOOM RapidCad is 34% faster
On HIDE Rapidcad is 46% faster
If you want more information there is a news release in the Intel forum
which can be downloaded.
->Ricki (Intel Technical Support)
Thanks for the info. Is Intel going to offer a trade in policy for the CPU
and the Math Co-Pro that are removed? They hinted at this in the article.
Ricki,
Get one of the SysGods to stick it the DLs for us, that's some very
valuable information that should reside locally.
Dennis
Ricki,
My understanding is that RapidCAD is basically a 486DX chip that fits in a
386 socket. If that's so, why is it being marketed solely as a CAD
enhancement solution? Seems to me that any PC application could benefit
from a performance boost, even those without FP requirements.
Jake
Jake,
It fits in a 386 socket? What'd they do, rip some pins out, or totally
repackage the silicon? This sounds like a real interesting development. I
have an upgrade option on my AMI mo-bo to upgrade it to a 486, this could
save me some money. Tell me more, anybody…
Dennis
Jake:
I got to play with a RapidCAD just before it was shipped. It's a two chip
set. One chip goes into the 386 socket; the other replaces the math-copro
and handles exceptions that must go out across the bus. The major speed
improvement is due to the integration of the main CPU and math functions
into the same die. Intel notes that the processor/math co-processor
combination spends a full 30% of its time just handling I/O between the two
chips.
I ran the CADalyst benchmarks on the RapidCAD and on the same 386 with a
standard 386/387 chip set. The results, along with a "First Look" review,
appear on page 24 of the March issue of CADalyst.
David