( ESNUG 445 Item 5 ) -------------------------------------------- [05/24/05]
Subject: ( ESNUG 443 #6 ) Users Note that DC XG Mode also Found in PhysOpt
> For two of those designs, based on Linux64 platform, we found:
>
> Test Case 1 (~2 M gates + 107 RAMs)
>
> dc_shell dc_shell-xg
> -------- -----------
> WNS -0.45 -0.45
> TNS -0.83 -0.83
> Memory 2.3 G 1.8 G 22% smaller
> Compile Time 3.7 hrs 2.6 hrs 1.4x faster
>
> Test Case 2 (~1.5 M gates + 85 RAMs)
>
> WNS -3.01 -2.74 9% less
> TNS -1127.39 -429.30 62% less
> Memory 3.0 G 2.4 G 20% smaller
> RunTime 2.1 hrs 1.4 hrs ~1.5 faster
>
> No changes were needed in the DC scripts, since already tcl-based. Note
> that design constraints have been made particularly aggressive in order
> to stress DC (runtime).
>
> - Fabio Collina
> LSI Logic Agrate Brianza, Italy
From: Alan Hsu <alanhsu=user company=sis spot gone>
Hello John,
I have to confess that I am not a so-called "early-adoptor" type of user,
and I don't upgrade to a new tool just for the sake of it, unless there is
something really makes my life easier.
I have been using PhysOpt during the past 4 years and thought I'd pretty
much explored the potential of this tool. A couple of months ago, I taped
out a 1M instance design (flattened) using PhysOpt 2003.06. The result was
good, especially for CTS and routing. The only thing that bothered me was
it took 5 days for the whole run to finish.
The PhysOpt 2004.12 XG mode release claimed "2X capacity and 2X runtime".
I was leery about this claim and wondered what would this "XG mode" really
buy me? I decided to give it a try. My design had 2M instances flattened,
130 nm. I used an AMD 64 bit machine with 16G memory:
2004.12 No XG 2004.12 XG Mode Change
runtime 7 days 2 days 71% reduction
mem use 23 G 11 G 52% reduction
timing -0.3 ns -0.3 ns no change
area 360 mil 360 mil no change
The results were not bad... Actually, well, I was happily surprised. With
the same timing and area, the mem use was cut 52%, and the overall runtime
was shortened 71%. Moving from the non-XG to XG mode, it took me about
half an hour to modify the scripts. Overall, the effort required to move
to PHysOpt-XG were well worth it.
- Alan Hsu
Silicon Integrated Systems Corp. Hsinchu, Taiwan
---- ---- ---- ---- ---- ---- ----
From: Stanley Peng <gotsu=user company=cisco spot gone>
Hi John,
I am seeing ~44% less memory usage with PhysOpt 2004.12-SP2 XG compared to
non-XG for the same design. The non-XG footprint w/ db generated is 14 GB.
With XG it's 7.8 GB.
- Stanley Peng
Cisco Systems San Jose, CA
Index
Next->Item
|
|
