Well put, especially the last part about calling BS when it is needed. I
witnessed more than a dozen companies sink due to chasing a wrong idea that
many on the team knew was wrong and did not or could not speak up about.
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On ;
Behalf Of Scott McMorrow
Sent: Wednesday, October 4, 2017 3:30 PM
To: chris.cheng@xxxxxxx; si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: Keynote Speeches from EDICon Last month
Chris understands. Sometimes in betting your job you just have to move ahead
to another opportunity before the current one overtakes you.
Rambus was a great technology for streams that could be pipelined. It was
totally inappropriate for Servers or Laptops, and was marginal even in Desktops
at the time. Follow on chipsets using SDRAM were also hampered by the bus.
IIRC Intel took around a $1B loss in chipsets around this time, not to mention
several processors in the family that were aborted before making it to market.
The nice thing about being a consultant who had many clients was that I was not
beholden to Intel or any company, so I could tell the truth in meetings.
Perhaps, I would argue, if you are stifling yourself in meetings and
communication, because of culture, pushback from above, or the fear of being
fired, you might be in the wrong job.
Obviously these are often personal decisions that one must make, but from where
I'm sitting I want engineers to challenge me, to call BS, and to speak the
convictions of their own engineering and personal knowledge. Always.
Scott McMorrow, CTO Signal Integrity Group Samtec Office 401-284-1827 |
+1-800-726-8329 www.samtec.com
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On ;
Behalf Of Cheng, Chris
Sent: Wednesday, October 4, 2017 5:37 PM
To: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: Keynote Speeches from EDICon Last month
There are plenty of Nintando64 and PlayStation with RAMBUS to answer your how
practical it is question.
When Bill Gunning and I was working on the first GTL systems, the target
impedance was in the 30's. All the server backplanes I designed for that
workstation company was in that range during that time. Bill was just borrowing
some of his old tricks in designing ECL bus with lower impedance to minimize
heavy loading impact on propagation delay and impedance mismatch.
Again, I am not here to defend RAMBUS's patent troll practice or lack of
attention to memory latency. But to say the electrical design is fundamentally
flawed does not give justice to many things that came after it. ALL the AGTL
FSB from Intel, DDR1,2,3,4 are double date rate source synchronous buses.
And even for that memory latency issue has a solution. We had a server design
almost the same as RAMBUS architecture before RAMBUS even exist. You just have
to double your memory and FSB bus width. That's why I had to pull my hair out
to design those crazy 1000+ pins processor packages in that company. And you
see those papers from some members of Si-list describing how much struggle they
had to deal with Swiss cheesed power planes power delivery and signals routing
through anti-pad gaps. Is it more practical to do that vs. pushing bandwidth
per pin and dropping the bus impedance ?
It's one thing when you are a workstation/server company that ships a few
thousands of processors a year with high margin. You can bury your package and
huge PCB cost for a wide bus in your profit margin.
It's another thing when you are King of processors and ship 100mil+ units a
year. At a few cents per pin per package, if you half your bus width, a few $$
savings times a 100 million is a lot of money. More than enough to justify even
the performance degradation and engineering resources to make that happen.
I served both masters and they asked me to do drastically different goals. But
I fully understand where they came from. SI exist to make a particular system
architecture works. That is very different from just making everything as fast
as possible with the coolest technology.
And to get back to Scott's original point. If you think something is not good,
do something about it. ServerWorks did and for a while they cornered the Intel
workstation/server chipset market. Before Intel can react to correct itself,
they exited with an 1 billion M&A with Broadcom. Now that's a cool Unicorn
story.
Chris Cheng
Distinguished Technologist , Electrical
Hewlett-Packard Enterprise Company
+1 510 344 4439/ Tel
chris.cheng@xxxxxxx / Email
4209 Technology Dr
Fremont, CA 94538
USA
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On ;
Behalf Of Lee Ritchey
Sent: Wednesday, October 04, 2017 10:34 AM
To: dmarc-noreply@xxxxxxxxxxxxx; si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: Keynote Speeches from EDICon Last month
During the 80s, I ran a design company that was designing PCBs for a group of
engineers at Lawrence Livermore Labs for a super computer for the US Navy. The
team that eventually formed Rambus was on that project and wanted us to design
a transmission line whose impedance was 28 ohms, not impossible, but not very
practical.
The navy cancelled the project and that team went on to form Rambus.
We wondered at the time how practical the idea was. Time showed us.
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On ;
Behalf Of heyfitch (Redacted sender "heyfitch" for DMARC)
Sent: Tuesday, October 3, 2017 2:34 PM
To: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: Keynote Speeches from EDICon Last month
Sent from my phone
On Oct 3, 2017, at 13:47, Lyndell Asbenson <Lyndell.Asbenson@xxxxxxxxxxx>
wrote:
Chris,
Regarding RAMBUS, it too had a lot of daemons. I have a wonderful plaque
celebrating our achievement in the Intel Chipset 820, and 840. The 840 was the
1st Intel Workstation utilizing high speed RAMBUS memory. What we were
attempting to do is figure out how to test the chipsets with RAMBUS interface
in very high volume manufacturing. This brought us fact to face with all kinds
of SI challenges we had never before seen. And had a very short time to figure
out. We know we had to test it at speed, yet there was no production tester
fast enough to do the job. This required a new technology in testing call flyby
channel multiplexing.
Regards,
Lyndell Lee Asbenson
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On ;
Behalf Of Cheng, Chris
Sent: Monday, October 2, 2017 11:07 PM
To: si-list@freelists.
Subject: [SI-LIST] Re: Keynote Speeches from EDICon Last month
Funny you mention Merced while Scott was talking about RAMBUS. 460GX chipset
was using PC100 SDRAM, not RAMBUS.
Merced had a lot of daemons, RAMBUS was not one of it.
And I sat through the decision process in the other processor company when Bill
Joy and Andy Bechtolsheim overrule the engineers recommendation to use RAMBUS.
They correctly realized that bandwidth per pin could not justified the degraded
performance due to increase in memory latency imposed by RAMBUS.
Speed and signal integrity was never a concern and I know both the engineer
responsible to make that happen in RAMBUS and the engineer in that workstation
company tasked to make sure it worked.
RAMBUS exist to address a simple problem at that time, highest bandwidth per
pin for cost sensitive application. And it had plenty of examples in
Nintando64 and PlayStation. And Intel always had an alternative as backup, it's
called ServerWorks.
Chris Cheng
Distinguished Technologist , Electrical
Hewlett-Packard Enterprise Company
+1 510 344 4439/ Tel
chris.cheng@xxxxxxx / Email
4209 Technology Dr
Fremont, CA 94538
USA
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On ;
Behalf Of Lee Ritchey
Sent: Monday, October 02, 2017 11:40 AM
To: Scott@xxxxxxxxxxxxx; si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: Keynote Speeches from EDICon Last month
Scott,
Almost makes me want to go out and become an engineer! Nice work!
Didn't know you lived through the Intel Merced disaster! Only time I ever
heard a CEO admit publicly that Intel had made that mistake. Craig Barret did
that.
I watched the Apollo landing because I had a radio on board. That design
was my first job out of college. Couldn't believe they gave such important
jobs to beginners! But, then, everyone on that project was a beginner at
designing those things!
Lee
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On ;
Behalf Of Scott McMorrow
Sent: Monday, October 2, 2017 10:02 AM
To: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Keynote Speeches from EDICon Last month
https://www.signalintegrityjournal.com/articles/567-edi-con-usa-keynote-talk
s---bet-your-job-5g-in-5-years-and-tm-value
Scott McMorrow, CTO Signal Integrity Group Samtec Office 401-284-1827 |
+1-800-726-8329 www.samtec.com<http://www.samtec.com/>
This email and any appended documents are only for the intended person/entity
and may contain information of Samtec, Inc., that is PRIVILEGED, PROPRIETARY,
CONFIDENTIAL, AND/OR PROTECTED BY LAW. If you are not the intended recipient
you are hereby notified that any dissemination, disclosure, use or copying of
this email or its contents is prohibited. If you received this message in
error, please notify Samtec immediately and delete the email, attachments and
all copies. The intended recipient should not disclose the content to third
parties or reproduce the content without Samtec's written consent.
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List forum is accessible at:
http://tech.groups.yahoo.com/group/si-list
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List forum is accessible at:
http://tech.groups.yahoo.com/group/si-list
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List forum is accessible at:
http://tech.groups.yahoo.com/group/si-list
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List forum is accessible at:
http://tech.groups.yahoo.com/group/si-list
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
.
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List forum is accessible at:
http://tech.groups.yahoo.com/group/si-list
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List forum is accessible at:
http://tech.groups.yahoo.com/group/si-list
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List forum is accessible at:
http://tech.groups.yahoo.com/group/si-list
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
This email and any appended documents are only for the intended person/entity
and may contain information of Samtec, Inc., that is PRIVILEGED, PROPRIETARY,
CONFIDENTIAL, AND/OR PROTECTED BY LAW. If you are not the intended recipient
you are hereby notified that any dissemination, disclosure, use or copying of
this email or its contents is prohibited. If you received this message in
error, please notify Samtec immediately and delete the email, attachments and
all copies. The intended recipient should not disclose the content to third
parties or reproduce the content without Samtec’s written consent.
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List forum is accessible at:
http://tech.groups.yahoo.com/group/si-list
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List forum is accessible at:
http://tech.groups.yahoo.com/group/si-list
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu