( ESNUG 587 Item 3 ) ---------------------------------------------- [05/30/19]
Subject: SCOOP -- Anirudh is launching a new Spectre-X SPICE at this DAC'19
- from 263 engineers on their 2017 SPICE use and SPICE leaders
From: [ John Cooley of DeepChip.com ]
I love love LOVE scoops like this!
Remember how Anirudh declared Total War against ANSYS HFSS because Cadence
discovered a totally new *breakthrough* way of SOLVING MATRICES???
(Here's the CDNlive'19 slide where Anirudh first publically spoke about his
new "breakthrough algorithm" in his Clarity launch.):
THE NIGHTMARE FOR ANSYS: It was all in Anirudh's Slide 16. He's launching
a 3D solver with structure-aware meshing. He said "totally new breakthough
algorithm" quite a few times -- meaning it's super fast. Then "the physics
does not change" -- meaning no approximations were involved; it's accurate!
(click on pic to enlarge image)
|
Then Anirudh spoke about his new *compact* data architecture designed from
the beginning to be auto-partitioned to run on a sea of cheap smaller memory
machines; plus it dynamically manages latency/fault-tolerance involved with
massive distributed computing on the cloud -- meaning it's high capacity!
- from Anirudh goes total war on Ansys mothership at CDNlive'19!
Well, spies tell me that Anirudh's using this exact same solver technology
to give his elderly Spectre SPICE franchise a big dose of the BLUE PILL.
Yes, grandpappy Spectre is now a hot new young virile Spectre-X that's now:
- massively parallel (runs up to 128 CPUs)
- new compact data model (5X smaller than present day Spectre)
- super fast (up to 10X faster than present day Spectre)
- 100% backwards compatible to present day Spectre/APS/RF/XPS/AMS
This new Cadence Spectre-X is Full SPICE accurate -- that is it's NOT Fast
SPICE, it's Full SPICE.
Of course it's been fully integrated in to Tom Beckley's Virtuoso ADE, after
all it's a Cadence tool and Tom rules their analog/RF/AMS domains. What
makes this interesting (and bad news for Ravi Subramanian of MENT BDA AFS
and Aart de Geus, Czar of all the many SNPS SPICEs) is how Spectre-X
disrupts the SPICE ecosystem.
For example, I recently published some SPICE user benchmarks. (See DAC'18 #6)
In each benchmark, good olde Full SPICE Spectre was the good but really slow
golden reference SPICE simulator.
Here's our 3-way SPICE benchmark:
APS+HPA = Cadence Spectre APS plus "+postlayout=hpa" for best speed
AFS = Mentor BDA AFS
ALPS = Empyrean ALPS
(click on pic to enlarge image)
|
- from Empyrean ALPS in SPICE benchmarks get #6 "Best of 2018"
In this chart the present day Spectre SPICE was the baseline and
Mentor BDA AFS was 2x faster than present day Spectre
Empyrean ALPS was 2x to 5x faster than present day Spectre
and other benchmarks in that same link effectively said:
Synopsys HSPICE was 1x the speed of present day Spectre
But what happens if you replace present day Spectre with Anirudh's totally
new 10X faster Spectre-X??? That makes
Spectre-X is now 5x faster than present day Mentor BDA AFS
Spectre-X is now 2x to 5x faster than present day Empyrean ALPS
Spectre-X is now 10x faster than present day Synopsys HSPICE
Ask yourself: with this Spectre-X speed-up, how will it impact the future
of the overall SPICE market? Here's Amit Gupta's SPICE data from 2017:
So what's going to happen to Synopsys and Mentor SPICE with this 10x change?
(Sorry, Aart.) (Sorry, Ravi.)
---- ---- ---- ---- ---- ---- ---- ----
NEW SPECTRE-X (MOSTLY) LINEAR: When it came to the CPU scalability question,
I got this data:
1 CPU core: :############################################# 45 hours
4 CPU core: :############ 11.5 hours (3.9X scaling)
16 CPU core: :#### 3.7 hours (12x scaling)
64 CPU core: :## 1.9 hours (24x scaling)
128 CPU core: :# 1.5 hours (30x scaling)
CMOS image sensor design with 150K tranistors and 45M paracitic Rs & Cs
So things are roughly linear for Spectre-X at 64 CPUs. Beyond 64 CPUs it's
diminishing returns.
- John Cooley
DeepChip.com Holliston, MA
---- ---- ---- ---- ---- ---- ----
Related Articles
263 engineers surveyed on their Custom IC design types & nodes
263 engineers on their present day SPICE use and SPICE leaders
Marmaduke's AMS circuit designer's Austin DAC'16 Trip Report
What's REALLY up on new Virtuoso ADE and Custom Compiler
CDNS admits DeepChip rumor right on Virtuoso ADE 12.X spilt
Join
Index
Next->Item
|
