( DAC 04 Item 18 ) --------------------------------------------- [ 02/09/05 ]
Subject: Magma Blast Create, Blast Plan, Blast Rail
GARY'S 2 PERCENT -- When I talk with Gary Smith of Dataquest about his
EDA marketshare numbers, he tells me he puts something like 98% of all
Magma revenue under what you'd call the P&R market. This is what the
other 2% is going towards... Blast Create, Blast Noise, Blast Rail,
and Blast Plan Pro.
I thought the Blast Create demo was good. It brings frontend designers
closer to getting post-layout timing closure earlier in design cycle.
- [ An Anon Engineer ]
All the Magma DAC demos were geared towards prospective customers with
little or no exposure to Magma toolset. Therefore, each session was
more like an overview of the specific functionality. All sessions ran
a small design during the demo, thus making it a "live demo". I find
this it to be pointless. A few other larger EDA vendors did the same
bullshit "live demo" sessions. Moreover, the presenters pulled custom-
prepared Tcl scripts from the pull-down menu that are not usually
shipped with the toolset. In most cases, the presenters failed to
mention that.
Toolset:
- Heavily Tcl based. You'll understand why once you've seen the
Magma user interface.
- Magma data-model structure is confusing, particularly if you're
coming from a standard LEF/DEF/.lib flow. (I guess it is a piece
of cake if you're from an Avanti or IBM background.)
- Blast Create, which I believe is their RTL synthesis tool, isn't
really useful in a sense that mapping is done to the Magma
Super_Cell level only. You can only write out a min_size library
gate level netlist. For full mapping, you need to move on to
Blast Fusion through placement and optimization. Therefore, I
don't see it as a player in the stand alone synthesis market.
(I don't think Magma is trying to do that, anyway).
- Blast Plan Pro was presented as a superior prototype tool, with
tight correlation to Blast Fusion. In my opinion there really
is no good tool for top-down design partition and time budgeting
in the market. (First Encounter is easy to use, though you cannot
rely on its internal extraction/timing engine and TrialRoute; the
best you get is partition pin placement/order which is sometime
more than enough.) Given the amount of time we spend on prototype
stage, the capabilities/promises of Blast Pan Pro look exciting.
- IR drop analysis and IR drop induced delay analysis and
optimization, as claimed, should be a strong selling point.
- Concurrent setup and hold analysis for both system and scan mode
is also a good Magma feature if it does work properly. (You need
to be careful with this in PKS).
All sessions started with the emphasis on the same timing engine,
extraction, etc. on a single database that Magma uses from RTL-to-GDS.
This is in fact the strongest selling point for Magma.
- [ An Anon Engineer ]
I didn't know about Magma or their tools until DAC this year when I saw
their Blast Create demo. Until then Synopsys was the only option for
me as far as ASIC synthesis was concerned. It is encouraging to see
that there is a possible prospect which can bring competition to the
synthesis arena.
I am Synopsys user, and I have no first hand experience with Magma.
However, I think it is worth investigating their synthesis approach
and claims. I liked their demo and I think their approach was novel.
I like their physical synthesis approach, concept of Super Cell, and
gain models. Most of all, I really like the fact their tool has the
end in sight (place & route) instead of blasting away at the logic.
I would like to evaluate them myself someday. Unfortunately, Magma
and their tools are still unknown by most ASIC designers and not
perceived as serious contender yet. Also, there are a lot of unknowns
still. For one, I don't know whether vendor synthesis library is
readily available. In addition, most companies carry tons of legacy
scripts and designs in Design Compiler. I believe Magma has to prove
itself beyond the reasonable doubt and provide substantial benefit
before it can be considered by ASIC design establishments.
- [ An Anon Engineer ]
Still getting by with DC
- [ An Anon Engineer ]
Magma's Blast Create, Blast Fusion, Blast Noise and Blast Rail demos
showed concurrent optimization of timing noise, power, SI, and
automatic block placement that might actually work. Its single
executable (which was internally developed instead of acquired from
multiple sources) looks very clean and easy to use.
I believe that Magma has demonstrated a substantial lead over Cadence
and Mentor in RTL-to-GDSII technology, especially at 130 nm and below.
Cadence and Mentor have their work cut out to compete with Magma, and
can only hope to do so with their current offerings by further reducing
their pricing with respect to Magma.
- Dan Risler of Cirrus Logic
Based on their DAC demo, the most interesting part of Blast Plan Pro
for me are the glassbox abstraction and integration with synthesis and
layout. This seems to be a strength of Magma, which is the only company
in EDA industry to have a common date base shared among synthesis,
prototyping, clock tree, and P&R tools. Auto partition and legalized
hard macro placement were also neat for new prototyping users.
- Charles Kuo of Progate Group Corp.
Blast Fusion is a great tool to use. Design kit & library perparation
requires some support. The Blast Plan Pro is not very efficient.
- [ An Anon Engineer ]
I do my own floorplanning fine. Magma's Blast Plan Pro tool seems to
be integrated into the rest of their suite so if I am using the Magma
tools I might as well use it.
- [ An Anon Engineer ]
Index
Next->Item
|
|
