Stephen,
I would suggest to move the technical aspects of this discussion to the IBIS
ATM email reflector. I think that forum would be the best place to continue
with this topic. So I am adding that email address to the "To:" list, and I
would
ask you to remove the SI-List address if/when you reply. This would hopefully
give us a smooth transition to the ATM email reflector for this discussion.
For now, I just want to mention briefly that using encryption is going to be
a tricky thing, because currently we do not support encryption in any of the
IBIS and related specifications, and for that reason each EDA vendor would
end up using their own encryption solutions. This will make the models
vendor specific, and IBIS does not approve/support anything along those
lines. This would also make the model maker's life a lot more difficult,
because they would end up having to make a different model for every
single EDA tool with their respective encryption utilities.
There were discussions about looking into an industry-wide (common)
encryption solution which would work for all simulators, but we haven't
pursued that idea yet. If you think encryption is necessary for these
models to become a reality, we will need to look into this topic more
seriously.
However, it seems that when IBIS-AMI was introduced, people felt that
the binary nature of the AMI executable models were safe enough for IP
protection. This might be another reason to consider making changes to
the AMI specification to support DDR5 models and simulations. Extending
the AMI specification in those directions might be easier and faster than
introduce a completely new encryption syntax for IBIS.
Thanks,
Arpad
==========================================================
-----Original Message-----
From: slater@xxxxxxxxxxxx [mailto:slater@xxxxxxxxxxxx] ;
Sent: Wednesday, October 25, 2017 10:39 PM
To: Muranyi, Arpad <Arpad_Muranyi@xxxxxxxxxx>; si-list@xxxxxxxxxxxxx
Subject: RE: EQ for DDR5
Hi Arpad,
I do certainly believe in the founding principles of IBIS, and agree that it
benefits the entire community to pull together under standard modeling
constructs. I should clear up that I mean only that there is no need to
complicate the *IBIS-AMI* part of the standard at present. I do see a need for
an additional external EQ file (encrypted) to be shared with IBIS files. The
encryption is not for the purpose of making it proprietary to the EDA vendor,
but for the purpose of protecting IP. I see no issue to raise a BIRD with the
IBIS committee in the future.
Message me personally if you'd like to connect on this topic.
Best regards,
Stephen.
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On ;
Behalf Of Muranyi, Arpad
Sent: Wednesday, October 25, 2017 8:08 PM
To: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: EQ for DDR5
Stephen,
The technical part was OK and appreciated.
The offending part was your discussion of Keysight's solutions and its
capabilities, practically inviting people to turn to Keysight for support.
The worst of this was your conclusion that there is no need to add anything to
IBIS because you offer a (most likely proprietary) solution for these needs.
Mentor, and probably other vendors could also offer similar or equivalent
solutions and would love to encourage everyone to come to us because we are
eager to provide our expertise and support.
Obviously there is no need for a spec or standard if everyone would just use
"MY" product(s) to do all their work... But IBIS is there despite all of its
shortcomings because it promises portability, interoperability, tool
independence, etc... so that different IC and tool vendors and system designers
could all work together.
Arpad
=======================================================
-----Original Message-----
From: slater@xxxxxxxxxxxx [mailto:slater@xxxxxxxxxxxx]
Sent: Wednesday, October 25, 2017 9:38 PM
To: Muranyi, Arpad <Arpad_Muranyi@xxxxxxxxxx>; si-list@xxxxxxxxxxxxx
Subject: RE: EQ for DDR5
Arpad, I was careful to include technical solution details for a problem that
many of our mutual customers are asking. My aim was to further the discussion
on whether or not IBIS-AMI is suitable for DDR or not. It is a topic that many
SI-listers do have an interest in. My aim was to be educational, and to settle
any fears that our EDA community doesn't have working approaches to these
problems.
If all deem my opinion here to be unworthy of pushing forward the conversation,
then for that I am truly sorry.
Best Regards,
Stephen.
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On ;
Behalf Of Muranyi, Arpad
Sent: Wednesday, October 25, 2017 4:54 PM
To: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] Re: EQ for DDR5
Stephen,
This forum should not be used for blatant product advertisements!!!
You are not the only one who has proprietary solutions for this type of
simulations.
Arpad
======================================================
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On ;
Behalf Of slater@xxxxxxxxxxxx
Sent: Wednesday, October 25, 2017 6:19 PM
To: si-list@xxxxxxxxxxxxx
Subject: [SI-LIST] EQ for DDR5
Dear SI-Listers,
I wanted to pick up an earlier thread from Nathan, Nitin and Walter regarding
how best to simulate EQ in the upcoming DDR5 interfaces (and the highest speed
grade of DDR4 @3200MT/s). Indeed this has been a recurring topic of
conversation, between chip vendors, system designers and EDA vendors. There
have been multiple papers at DesignCon 2016 and 2017 presenting analyses of EQ
applied to DDR.
Here's a link to a DesignCon paper demonstrating the application of CTLE,
co-authored with a partner:
https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.xilinx.com%2Fpublications%2Fevents%2Fdesigncon%2F2017%2Fcharacterization-ddr4-receiver-sensitivity-impact-post-equalization-eye-paper.pdf&data=02%7C01%7C%7Cce2034fe09a34e8d0b1108d51c03c4ad%7C63545f2732324d74a44dcdd457063402%7C1%7C1%7C636445724876252401&sdata=1g2rgyHdxFGDv7FysP%2BsVSYPWskAZLwXa%2Byv3CQdmRE%3D&reserved=0
Walter's article pointed out that IBIS-AMI is designed for differential
signaling and is not presently suited to tackle single-ended DDR interfaces.
He also quite correctly pointed out that the eye in the single-ended DQ and CA
signals is asymmetric, meaning the rising edges and falling edges are very
different. To the casual reader it may seem that EDA vendors don't have a
working solution for this today. However this is not the case, as Keysight has
had a solution for DDR with EQ since 2014.
The technical solution is a DDR Bus Sim (a statistical channel simulation that
characterizes both rising and falling edges independently). Here the IBIS
model is characterized as part of the channel, and the thankfully the EQ that
is being proposed for DDR5 can be modeled very well statistically, such as CTLE
and DFE. These Rx settings are part of the normal simulation setup. The
ultimate aim is to present the equalized eye and the BER contours in order to
make measurements of margin to the Rx BER Mask.
If you're a HW engineer (not a chip designer), then my recommendation would be
to partner up with your DRAM vendor and Memory Controller vendor to best start
your simulation of these designs. Keysight is always there to provide support,
and can encrypt the EQ settings to protect the chip vendors' IP.
We don't see a need to add complexity to the IBIS-AMI standard at this point in
time. I hope this adds an alternative perspective to the discussion.
Best regards,
Stephen Slater
Product Manager for High Speed Digital Design Keysight EEsof EDA Division
------------------------------------------------------------------
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:
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.freelists.org%2Fwebpage%2Fsi-list&data=02%7C01%7C%7Cce2034fe09a34e8d0b1108d51c03c4ad%7C63545f2732324d74a44dcdd457063402%7C1%7C0%7C636445724876252401&sdata=l8dmjrw4Uir3w4VBvFT7sauINIz1Fb46ILdOfREuzkg%3D&reserved=0
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List forum is accessible at:
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Ftech.groups.yahoo.com%2Fgroup%2Fsi-list&data=02%7C01%7C%7Cce2034fe09a34e8d0b1108d51c03c4ad%7C63545f2732324d74a44dcdd457063402%7C1%7C0%7C636445724876252401&sdata=kUy1m3Sz6F6y8o%2BaIWK6efuc3d0tQ%2BmyhtogcNFV0cQ%3D&reserved=0
List archives are viewable at:
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.freelists.org%2Farchives%2Fsi-list&data=02%7C01%7C%7Cce2034fe09a34e8d0b1108d51c03c4ad%7C63545f2732324d74a44dcdd457063402%7C1%7C0%7C636445724876252401&sdata=a6Qu%2B825n3VpPQ1B%2BrhVNkGTOfeQYS%2FpWOKHUTPbPSY%3D&reserved=0
Old (prior to June 6, 2001) list archives are viewable at:
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.qsl.net%2Fwb6tpu&data=02%7C01%7C%7Cce2034fe09a34e8d0b1108d51c03c4ad%7C63545f2732324d74a44dcdd457063402%7C1%7C0%7C636445724876252401&sdata=eOBuKUjhPYH96ebuDAsC5N78WEq4yaSuFPsXYOdg2Yc%3D&reserved=0
------------------------------------------------------------------
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:
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.freelists.org%2Fwebpage%2Fsi-list&data=02%7C01%7C%7Cce2034fe09a34e8d0b1108d51c03c4ad%7C63545f2732324d74a44dcdd457063402%7C1%7C0%7C636445724876252401&sdata=l8dmjrw4Uir3w4VBvFT7sauINIz1Fb46ILdOfREuzkg%3D&reserved=0
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List forum is accessible at:
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Ftech.groups.yahoo.com%2Fgroup%2Fsi-list&data=02%7C01%7C%7Cce2034fe09a34e8d0b1108d51c03c4ad%7C63545f2732324d74a44dcdd457063402%7C1%7C0%7C636445724876252401&sdata=kUy1m3Sz6F6y8o%2BaIWK6efuc3d0tQ%2BmyhtogcNFV0cQ%3D&reserved=0
List archives are viewable at:
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.freelists.org%2Farchives%2Fsi-list&data=02%7C01%7C%7Cce2034fe09a34e8d0b1108d51c03c4ad%7C63545f2732324d74a44dcdd457063402%7C1%7C0%7C636445724876252401&sdata=a6Qu%2B825n3VpPQ1B%2BrhVNkGTOfeQYS%2FpWOKHUTPbPSY%3D&reserved=0
Old (prior to June 6, 2001) list archives are viewable at:
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.qsl.net%2Fwb6tpu&data=02%7C01%7C%7Cce2034fe09a34e8d0b1108d51c03c4ad%7C63545f2732324d74a44dcdd457063402%7C1%7C0%7C636445724876252401&sdata=eOBuKUjhPYH96ebuDAsC5N78WEq4yaSuFPsXYOdg2Yc%3D&reserved=0
------------------------------------------------------------------
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:
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.freelists.org%2Fwebpage%2Fsi-list&data=02%7C01%7C%7C92ee42420196402b557408d51c1ed892%7C63545f2732324d74a44dcdd457063402%7C1%7C0%7C636445841169737159&sdata=%2Fqt1Xgmm6Nq2KAC4KXLXjanQ12IL%2F9wz4xZsTozBprw%3D&reserved=0
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List forum is accessible at:
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Ftech.groups.yahoo.com%2Fgroup%2Fsi-list&data=02%7C01%7C%7C92ee42420196402b557408d51c1ed892%7C63545f2732324d74a44dcdd457063402%7C1%7C0%7C636445841169737159&sdata=9LklRUuWi4YN%2FHf0YbYm3foAaKBBLLTwMEVZljdrFbc%3D&reserved=0
List archives are viewable at:
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.freelists.org%2Farchives%2Fsi-list&data=02%7C01%7C%7C92ee42420196402b557408d51c1ed892%7C63545f2732324d74a44dcdd457063402%7C1%7C0%7C636445841169737159&sdata=p9YTZhkrJ0k8EeFzHEaA2fnX3Cu4RPmsOXQhgJ1zzqI%3D&reserved=0
Old (prior to June 6, 2001) list archives are viewable at:
https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.qsl.net%2Fwb6tpu&data=02%7C01%7C%7C92ee42420196402b557408d51c1ed892%7C63545f2732324d74a44dcdd457063402%7C1%7C0%7C636445841169737159&sdata=U73JRwOrihZL2qK7eVayag7Mm8sCQ7cDGTB6bAdnf%2F0%3D&reserved=0
---------------------------------------------------------------------
IBIS Macro website : http://ibis.org/macromodel_wip/
IBIS Macro reflector: //www.freelists.org/list/ibis-macro
To unsubscribe send an email:
To: ibis-macro-request@xxxxxxxxxxxxx
Subject: unsubscribe